Systems and methods for testing and analyzing visual acuity and changes thereto

Abstract

Disclosed herein are systems and methods for testing and analyzing visual acuity and changes using an acuity model generated based on one or more acuity chart design parameters and candidate acuity parameters calculated using acuity test data of a subject, the visual acuity model including a chart-specific psychometric function determined using a family of multi-landolt-circles psychometric functions, and wherein the acuity model is configured to estimate a probability of obtaining the acuity test data of the subject.

Description

[0001] This application is a divisional application of Chinese patent application No. 201880054850.6, filed on June 22, 2018, entitled "System and method for testing and analyzing visual acuity and its changes" (the corresponding PCT application was filed on June 22, 2018, and has the application number PCT / US2018 / 039137).

[0002] Cross-references

[0003] This application claims the benefit of U.S. Provisional Application No. 62 / 524,414, filed June 23, 2017, which is incorporated herein by reference in its entirety. Background Technology

[0004] Visual acuity charts are commonly used to measure visual acuity due to their standardization, ease of use, and wide distribution. Visual acuity tests typically involve asking patients to identify visual patterns of optotypes—letters or objects—presented on test slides (or a series of slides). Each slide contains one or more optotypes arranged in one or more rows. The patient's ability to identify the optotype based on its size is recorded and scored to produce an estimate of visual acuity. The resulting estimate of visual acuity (e.g., the smallest optotype size the patient can reliably see) can be fundamentally dependent on the design of the test and scoring, which is based on predetermined rules and heuristics specific to each chart. Summary of the Invention

[0005] Two methods for visual acuity testing—graphical tests where patients identify multiple lines of optotypes presented simultaneously or computerized tests where one optotype is presented at a time—have become the dominant technologies. Computerized testing has recently gained favor because it offers the potential advantage of improved test resolution through flexible and accurate sampling of optotype size on a digital display, as well as the application of adaptive and personalized testing strategies that focus on the patient's specific visual acuity.

[0006] A significant drawback of existing sensitivity tests is the discrepancy in chart design between those used in clinical trials and those used in clinical practice. Consequently, sensitivity test results collected in clinical practice are not precisely correlated with those collected on the "gold standard" charts used in clinical trials. Because clinical trials require lengthy testing periods, the chart design and testing standards remain unsuitable for clinical testing by eye health professionals or general practitioners. Even with extended timeframes allowed in clinical studies, sensitivity test results using pediatric charts (e.g., HOTV, Lea design) still exhibit systematic bias when compared to results obtained from charts designed for adults.

[0007] Accurately testing visual acuity presents a challenge. Many charts currently used in clinical settings provide only qualitative, not quantitative, results for acuity testing (e.g., 20 / 20 or 20 / 40). Furthermore, the diverse chart designs circulating across clinical testing sites can make comparing and integrating acuity tests obtained from different charts difficult. Without precision and consistency among charts used for clinical vision testing, tracking how vision changes with disease or treatment remains challenging.

[0008] The systems and methods disclosed herein for testing and analyzing visual acuity address problems existing in visual acuity testing. Advantages of the systems and methods disclosed herein include the ability to develop personalized acuity charts based on subject-specific (interchangeable with target-invariant) information relating to his / her visual acuity. In some embodiments, another advantage of the systems and methods disclosed herein includes the ability to analyze visual acuity data collected using a wide range of acuity charts (including, but not limited to, legacy charts). In some embodiments, another advantage of the systems and methods disclosed herein includes the generation of accurate acuity parameters independent of the design of the acuity charts used to collect the data.

[0009] In some embodiments, this document discloses a scoring algorithm (which is interchangeably referred to herein as Algorithm S) capable of analyzing visual acuity data optionally collected on different acuity chart designs. In some cases, this document discloses an adaptive algorithm (which is interchangeably referred to herein as Algorithm A) that improves the quality of visual acuity data by generating or selecting personalized chart designs that precisely focus on the acuity of the test subjects. In some cases, this document discloses a combined algorithm that iteratively combines Algorithms A and S: evaluations of previously collected acuity data (from subjects or from a population) are used to personalize the chart design to collect acuity chart data using Algorithm A. This data, in turn, can be scored using Algorithm S, which provides a current estimate of the patient's visual acuity. This iterative process can provide highly accurate estimates of visual acuity and reduce testing time. In some embodiments, this document discloses a sensitivity model for analyzing visual acuity that predicts visual performance during testing using different acuity chart designs. In some cases, such a sensitivity model provides the basis for Algorithms S and A. In some implementations, the sensitivity models disclosed herein provide a robust description of sensitivity chart performance, summarized as predicting visual performance across different sensitivity charts using subject-specific model parameters and / or chart-specific model parameters. These subject-specific model parameters (which may be interchangeable herein as chart-invariant parameters) provide valuable sensitivity parameters that can be used to track visual health over time.

[0010] For example, in the sensitivity model disclosed herein, the distinction between chart-specific and chart-invariant indicators (which are interchangeable as parameters) provides flexibility and versatility for estimating visual acuity. These two sets of sensitivity indicators can be important and complementary. The set of chart-specific indicators can describe visual acuity within the scope of a specific sensitivity chart used during testing. An alternative set of chart-invariant indicators can describe visual acuity independently of the design features of the sensitivity chart. Therefore, these composite parameters can be flexible enough to deliver sensitivity indicators specific to or independent of the sensitivity chart used for testing. Chart-specific indicators are important for R&D applications that use the same chart over time. The chart-invariant indicator approach is important for evaluating and comparing data from different clinical practices and populations. Figure 3A As shown, the flowchart of an exemplary implementation of generating a chart-invariant sensitivity model may include operation 230 using chart-invariant sensitivity parameters, which may then be converted into chart-specific sensitivity parameters based on the design factors of the sensitivity chart.

[0011] The sensitivity model disclosed in this paper can also provide a foundation for novel representations of visual acuity data. Visualization tools can help patients and clinicians understand the measured changes in sensitivity.

[0012] Because the same chart can often be used to measure visual acuity in different eyes, at different distances, or in situations where repeated measurements of visual acuity are required (e.g., tracking the progression of an eye disease or the effectiveness of treatment), it is difficult to prevent contamination between conditions. Recommendations from the National Academy of Sciences-National Research Council (NAS-NRC) committee and ANSI (American National Standards Institute) standards stipulate that sensitivity charts should contain multiple rows of optotypes of logarithmic decreasing size. Among other design recommendations (5-10 optotypes per row, random sampling without replacement), the standards do not recommend single-letter representations for computerized adaptive sensitivity methods.

[0013] Unlike traditional adaptive sensitivity testing methods that optimize optotype selection for only one letter of a single size, the system and method disclosed herein advantageously achieve composite optimization of sensitivity charts, which simultaneously estimate chart-specific and chart-invariant sensitivity indices. Additional advantages associated with the method and system disclosed herein include, but are not limited to: precise personalized attention for each individual subject while adhering to chart design standards; rapid testing time; accuracy comparable to time-consuming single-letter sensitivity tests; and subject visual acuity indices that can be described independently of optotype size, test chart, or testing procedure. Given knowledge of sensitivity chart design, both sets of sensitivity parameters (whether chart-specific or chart-invariant) are interchangeable and translatable.

[0014] In some embodiments, the systems and methods disclosed herein include graph-based adaptive sensitivity tests, which offer the standardization advantages of graph-based tests, along with the flexibility and accuracy advantages of single-letter tests. The systems and methods disclosed herein can apply Bayesian adaptive algorithms to personalize and precisely focus sensitivity graphs to optimally test each subject's visual acuity while adhering to design constraints or recommendations from an adjudication committee, or more generally, while fully maintaining any desired design constraints. In some embodiments, alternatively using Bayesian adaptive algorithms and signal detection frameworks to develop optimization methods for design constraints in visual acuity testing provides one or more of the following: adherence to graph design standards; precise personalization through individualized adaptive testing; shorter testing times for graphs; higher accuracy for single-letter sensitivity tests; correction for guessing behavior across different optotype sets; and visual acuity estimation that remains invariant to practical problems in sensitivity testing, such as the design of visual acuity tests or the optotypes used.

[0015] In one aspect, this paper discloses a computer-implemented method for generating a visual acuity model that scores the visual acuity of a subject, the method comprising: a) obtaining one or more acuity chart design parameters; b) collecting the subject's acuity test data; c) selecting a first set of chart-specific acuity parameters from one or more sets of candidate acuity parameters; d) generating a acuity model, comprising: i) generating a single-optotype mental measurement function, the single-optotype mental measurement function including the first set of chart-specific acuity parameters; ii) calculating a family of multi-optotype mental measurement functions using the single-optotype mental measurement function; and iii) generating a chart-specific mental measurement function using the family of multi-optotype mental measurement functions, the chart-specific mental measurement function including a second set of chart-specific acuity parameters. The acuity model is configured to estimate the probability of obtaining the subject's acuity test data based on the one or more acuity chart design parameters. In some embodiments, the method further includes: after (c) and before (i), generating a sensitivity-based psychometric function comprising a first set of chart-invariant sensitivity parameters, wherein the first set of chart-invariant sensitivity parameters includes a sensitivity threshold and a sensitivity range, and wherein the chart-invariant sensitivity parameters are generated based on the first set of chart-specific sensitivity parameters. In some embodiments, the sensitivity-based psychometric function is independent of the one or more sensitivity chart design parameters. In some embodiments, the sensitivity-based psychometric function is generated based on the one or more sensitivity chart design parameters and one or more additional parameters of the subject, the one or more additional parameters being chart-invariant. In some embodiments, the sensitivity-based psychometric function is generated based on signal detection theory. In some embodiments, the sensitivity-based psychometric function is configured to describe the subject's visual acuity performance as a d' function of one or more optotype sizes, and is independent of the one or more chart design parameters. In some embodiments, the method further includes, before (i), converting the sensitivity-based psychometric function into a single optotype psychometric function. In some embodiments, the first set of chart-specific sensitivity parameters includes a sensitivity threshold and a sensitivity range. In some embodiments, the single-visual psychometric function is chart-specific. In some embodiments, the second set of chart-specific sensitivity parameters includes a sensitivity threshold and a sensitivity range. In some embodiments, the first or second set of chart-specific sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a change in the sensitivity threshold between the two test conditions, a change in the sensitivity range between the two test conditions, a change in the sensitivity slope between the two test conditions, or a combination thereof. In some embodiments, the first or second set of chart-specific sensitivity parameters includes one or more numerical values. In some embodiments, the first or second set of chart-specific sensitivity parameters includes at least one posterior probability density function for its parameters.In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, a first or second set of chart-specific sensitivity parameters includes a posterior probability density function of sensitivity thresholds from the at least two different test conditions. In some embodiments, the one or more sensitivity chart design parameters include: total number of optotypes, number of optotypes per row, optotype sampling method, optotype size, number of rows, and the subject's response or a combination thereof. In some embodiments, the single-optotype psychometric function is based on at least one optotype size. In some embodiments, each function in the family of multi-optotype psychometric functions is based on two or more different optotype sizes. In some embodiments, the single-optotype psychometric function includes the subject's sensitivity threshold and sensitivity range. In some embodiments, one or more functions in the family of multi-optotype psychometric functions are based on multiple optotypes per row in a sensitivity chart or sub-chart. In some embodiments, (iii) further includes using a weighted sum of the family of multi-optotype psychometric functions. In some embodiments, the weighting of the family of psychometric functions is based on the subject's sensitivity chart data. In some embodiments, the weighting is determined manually by the user or automatically by a computer program. In some embodiments, (b) further includes displaying the subject one or more sensitivity charts generated based on the one or more sensitivity chart design parameters. In some embodiments, each of the one or more sensitivity charts includes at least one optotype selected from a optotype library. In some embodiments, the contrast of the one or more sensitivity charts is approximately 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol. In some embodiments, one or more sets of candidate sensitivity parameters are generated based on the one or more sensitivity chart design parameters, the subject's sensitivity test data, or both.

[0016] On the other hand, this document discloses a computer-implemented method for scoring the visual acuity of a subject, the method comprising: a) obtaining one or more acuity chart design parameters; b) collecting acuity test data of the subject; c) generating one or more sets of candidate acuity parameters based on the one or more acuity chart design parameters, the subject's acuity test data, or both; d) generating a acuity model, including generating one or more chart-specific psychometric functions, each chart-specific psychometric function including a set of chart-specific acuity parameters from the one or more sets of the chart-specific acuity parameters; e) generating the probability of observing the acuity test data using the acuity model and the one or more sets of candidate acuity parameters; f) ranking the one or more sets of candidate acuity parameters of the subject based on the probability; and g) selecting a set from the one or more sets of candidate acuity parameters of the subject based on the ranking. In some embodiments, each set of candidate acuity parameters includes: a acuity threshold and a acuity range. In some embodiments, each set of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between the two test conditions, a sensitivity range change between the two test conditions, a sensitivity slope change between the two test conditions, or a combination thereof. In some embodiments, each set of candidate sensitivity parameters includes a posterior probability density function of the one or more sensitivity parameters. In some embodiments, each set of candidate sensitivity parameters includes a posterior probability density function of each of the one or more sensitivity parameters. In some embodiments, (b) includes summarizing sensitivity test data from the subject. In some embodiments, the method further includes, after (g), displaying sensitivity test results to the subject or the user, the results including a selected set of candidate sensitivity parameters. In some embodiments, the subject's sensitivity test data includes data only from a first test condition. In some embodiments, the subject's sensitivity test data includes data from both a first and a second test condition. In some embodiments, the selected set of candidate sensitivity parameters includes a first posterior probability density function of a sensitivity threshold or a sensitivity range. In some embodiments, b) includes selecting data having the first test condition. In some embodiments, the method further includes, after g), selecting data having the second test condition from the subject's sensitivity test data; and repeating c) to g). In some embodiments, the selected set of candidate sensitivity parameters includes a second posterior probability density function of the sensitivity threshold or the sensitivity range. In some embodiments, the method further includes obtaining a difference distribution using the first and second posterior probability density functions.In some embodiments, the method further includes calculating a change in a sensitivity threshold or sensitivity range between the first test condition and the second test condition based on the difference distribution. In some embodiments, the visual acuity is chart-specific or chart-invariant. In some embodiments, (e) is based on Bayes' theorem or maximum likelihood estimation. In some embodiments, the posterior probability density function is a Bayesian posterior probability density function. In some embodiments, the first or second posterior probability density function is a Bayesian posterior probability density function. In some embodiments, (d) includes: generating a single-viewpoint psychometric function, the single-viewpoint psychometric function including a first set of chart-specific sensitivity parameters; calculating a family of multi-viewpoint psychometric functions using the single-viewpoint psychometric function; and generating a chart-specific psychometric function using the family of multi-viewpoint psychometric functions, the chart-specific psychometric function including a second set of chart-specific sensitivity parameters. In some embodiments, the method further includes generating a sensitivity-based psychometric function including a first set of chart-invariant sensitivity parameters before generating the single-viewpoint psychometric function. In some embodiments, the sensitivity-based psychometric function is independent of the one or more sensitivity chart design parameters. In some embodiments, the first set of chart-invariant sensitivity parameters includes a sensitivity threshold and a sensitivity range, and the first set of chart-invariant sensitivity parameters is generated based on one or more sets of candidate sensitivity parameters. In some embodiments, the sensitivity-based psychometric function is generated based on the one or more sensitivity chart design parameters and one or more additional parameters of the subject, the one or more additional parameters being chart-invariant. In some embodiments, the sensitivity-based psychometric function is generated based on signal detection theory. In some embodiments, the sensitivity-based psychometric function is configured to describe the subject's visual acuity performance as a d' function of one or more optotype sizes, and is independent of the one or more chart design parameters. In some embodiments, the method further includes converting the sensitivity-based psychometric function into a single optotype psychometric function. In some embodiments, the chart-specific sensitivity parameter set includes a sensitivity threshold and a sensitivity range. In some embodiments, the single optotype psychometric function is chart-specific. In some embodiments, the chart-specific sensitivity parameter set includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold variation between the two test conditions, a sensitivity range variation between the two test conditions, a sensitivity slope variation between the two test conditions, or a combination thereof. In some embodiments, the chart-specific sensitivity parameter set includes one or more numerical values. In some embodiments, the chart-specific sensitivity parameter set includes at least one posterior probability density function for its parameters.In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, the chart-specific sensitivity parameter set includes posterior probability density functions of sensitivity thresholds from the at least two different test conditions. In some embodiments, the one or more sensitivity chart design parameters include: total number of optotypes, number of optotypes per row, optotype sampling method, optotype size, number of rows, and the subject's response or a combination thereof. In some embodiments, the single-opposite-scale psychometric function is based on at least one optotype size. In some embodiments, each function in the family of multi-opposite-scale psychometric functions is based on two or more different optotype sizes. In some embodiments, the single-opposite-scale psychometric function includes the subject's sensitivity threshold and sensitivity range. In some embodiments, one or more functions in the family of multi-opposite-scale psychometric functions are based on multiple optotypes per row in a sensitivity chart or sub-chart. In some embodiments, generating the chart-specific psychometric function further includes using a weighted sum of the family of multi-opposite-scale psychometric functions. In some embodiments, the weighting of the family of psychometric functions is based on the subject's sensitivity chart data. In some embodiments, the weighting is determined manually by a user or automatically by a computer program. In some embodiments, (b) further includes displaying to the subject one or more sensitivity charts generated based on the one or more sensitivity chart design parameters. In some embodiments, each of the one or more sensitivity charts includes at least one optotype, wherein the at least one optotype is selected from a optotype library. In some embodiments, the contrast of the one or more sensitivity charts is approximately 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol. In some embodiments, the method includes: h) generating a plurality of candidate sensitivity charts based on the one or more chart design parameters; and i) selecting one or more of the plurality of candidate sensitivity charts for the subject's sensitivity test based on the ranking of the plurality of candidate sensitivity charts, wherein the ranking is based on a selected set of the subject's candidate sensitivity parameters. In some implementations, the method further includes, after (b), collecting data from the subject using the selected one or more candidate sensitivity charts; and repeating steps c) to g) and h) to i) if the stopping criteria are not met, until the stopping criteria are met.

[0017] In another aspect, this document discloses a computer-implemented method for adaptively evaluating the visual acuity of a subject, the method comprising: a) obtaining one or more acuity chart design parameters; b) collecting acuity test data of the subject; c) generating one or more sets of candidate acuity parameters based on the one or more acuity chart design parameters and the subject's acuity test data, each of the one or more sets of candidate acuity parameters including a ranking obtained by analyzing the subject's acuity test data; d) generating a plurality of candidate acuity charts based on the one or more acuity chart design parameters; e) ranking the plurality of candidate acuity charts based on the ranking of each of the one or more sets of candidate acuity parameters; and f) selecting one or more acuity charts from the plurality of candidate acuity charts based on the ranking of the plurality of candidate acuity charts to be presented to the subject for acuity testing. In some embodiments, the method further includes, after f), presenting the one or more selected acuity charts to the subject. In some embodiments, the method further includes, after f), updating the acuity test data with new data collected from the subject using the one or more selected acuity charts. In some embodiments, the method further includes repeating: presenting the subject with the one or more selected sensitivity charts; updating the sensitivity test data with new data collected from the subject using the one or more selected sensitivity charts; and steps b) to f); until a stopping criterion is met. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes: a sensitivity threshold and a sensitivity range. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a change in the sensitivity threshold between the two test conditions, a change in the sensitivity range between the two test conditions, a change in the sensitivity slope between the two test conditions, or a combination thereof. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes a posterior probability density function for the one or more sensitivity parameters. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes a posterior probability density function for each of the one or more sensitivity parameters. In some embodiments, (b) includes summarizing the sensitivity test data from the subject. In some embodiments, the subject's sensitivity test data includes data only from a first test condition. In some implementations, the sensitivity test data of the subject includes data from a first test condition and a second test condition.In some embodiments, each of one or more sets of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a change in the sensitivity threshold between two test conditions, a change in the sensitivity range between the two test conditions, a change in the sensitivity slope between the two test conditions, or a combination thereof. In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, each of one or more sets of candidate sensitivity parameters includes a posterior probability density function of the sensitivity threshold or sensitivity range. In some embodiments, the one or more sensitivity chart design parameters include: the total number of optotypes, the number of optotypes per row, the optotype sampling method, the optotype size, the number of rows, and the subject's response, or a combination thereof. In some embodiments, each of the candidate sensitivity charts includes at least one optotype, wherein the at least one optotype is selected from a optotype library. In some embodiments, the contrast of one or more candidate sensitivity charts is approximately 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol.

[0018] In another aspect, this document discloses a computer-implemented system for generating a visual acuity model that scores the visual acuity of a subject, the system comprising: a digital processing device including an operating system and a memory configured to execute executable instructions; and a computer program including instructions executable by the digital processing device to create a scoring application including software modules, the software modules being: a) obtaining one or more acuity chart design parameters; b) collecting the subject's acuity test data; c) selecting a first set of chart-specific acuity parameters from one or more sets of candidate acuity parameters; and d) generating a acuity model comprising: i. generating a single-opposite-target psychometric function, the single-opposite-target psychometric function including the first set of chart-specific acuity parameters; ii. calculating a family of multi-opposite-target psychometric functions using the single-opposite-target psychometric function; and iii. generating a chart-specific psychometric function using the family of multi-opposite-target psychometric functions, the chart-specific psychometric function including a second set of chart-specific acuity parameters. The acuity model is configured to estimate the probability of obtaining the subject's acuity test data based on the one or more acuity chart design parameters. In some embodiments, the system further includes: after (c) and before (i), generating a sensitivity-based psychometric function comprising a first set of chart-invariant sensitivity parameters, wherein the first set of chart-invariant sensitivity parameters includes a sensitivity threshold and a sensitivity range, and wherein the first set of chart-invariant sensitivity parameters is generated based on a first set of chart-specific sensitivity parameters. In some embodiments, the sensitivity-based psychometric function is independent of one or more sensitivity chart design parameters. In some embodiments, the sensitivity-based psychometric function is generated based on the one or more sensitivity chart design parameters and one or more additional parameters of the subject, the one or more additional parameters being chart-invariant. In some embodiments, the sensitivity-based psychometric function is generated based on signal detection theory. In some embodiments, the sensitivity-based psychometric function is configured to describe the subject's visual acuity performance as a d' function of one or more optotype sizes, and is independent of the one or more chart design parameters. In some embodiments, the system further includes, before (i), converting the sensitivity-based psychometric function into a single optotype psychometric function. In some embodiments, the first set of chart-specific sensitivity parameters includes a sensitivity threshold and a sensitivity range. In some implementations, the single-target psychometric function is chart-specific. In some implementations, the second set of chart-specific sensitivity parameters includes sensitivity thresholds and sensitivity ranges.In some embodiments, the first or second set of chart-specific sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a change in the sensitivity threshold between the two test conditions, a change in the sensitivity range between the two test conditions, a change in the sensitivity slope between the two test conditions, or a combination thereof. In some embodiments, the first or second set of chart-specific sensitivity parameters includes one or more numerical values. In some embodiments, the first or second set of chart-specific sensitivity parameters includes at least one posterior probability density function for its parameters. In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, the first or second set of chart-specific sensitivity parameters includes a posterior probability density function for sensitivity thresholds from the at least two different test conditions. In some embodiments, the one or more sensitivity chart design parameters include: the total number of optotypes, the number of optotypes per row, the optotype sampling method, the optotype size, the number of rows, and the subject's response, or a combination thereof. In some embodiments, the single-opotype psychometric function is based on at least one optotype size. In some embodiments, each function in the family of multi-opotype psychometric functions is based on two or more different optotype sizes. In some embodiments, the single-opposite psychometric function includes the subject's sensitivity threshold and sensitivity range. In some embodiments, one or more functions in the family of multi-opposite psychometric functions are based on multiple optotypes per row in a sensitivity chart or sub-chart. In some embodiments, (iii) further includes using a weighted sum of the family of multi-opposite psychometric functions. In some embodiments, the weighting of the family of psychometric functions is based on the subject's sensitivity chart data. In some embodiments, the weighting is determined manually by a user or automatically by a computer program. In some embodiments, (b) further includes using one or more sensitivity charts generated based on the design parameters of the one or more sensitivity charts. In some embodiments, each of the one or more sensitivity charts includes at least one optotype, wherein the at least one optotype is selected from a library of optotypes. In some embodiments, the contrast of the one or more sensitivity charts is approximately 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol. In some implementations, one or more sets of the candidate sensitivity parameters are generated based on the one or more sensitivity chart design parameters, the subject's sensitivity test data, or both.

[0019] In another aspect, this document discloses a computer system for generating a visual acuity model that scores the visual acuity of a subject, the system comprising: a digital processing device including an operating system and a memory configured to execute executable instructions; and a computer program including instructions executable by the digital processing device to create a scoring application including software modules, the software modules being: a) obtaining one or more acuity chart design parameters; b) collecting acuity test data of the subject; c) generating one or more sets of candidate acuity parameters based on the one or more acuity chart design parameters, the subject's acuity test data, or both; d) generating a acuity model, including generating one or more chart-specific psychometric functions, each chart-specific psychometric function including a set of chart-specific acuity parameters from the one or more sets of the chart-specific acuity parameters; e) generating the probability of observing the acuity test data using the acuity model and the one or more sets of candidate acuity parameters; f) ranking the one or more sets of candidate acuity parameters of the subject based on the probability; and g) selecting a set from the one or more sets of candidate acuity parameters of the subject based on the ranking. In some embodiments, each set of candidate sensitivity parameters includes a sensitivity threshold and a sensitivity range. In some embodiments, each set of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between the two test conditions, a sensitivity range change between the two test conditions, a sensitivity slope change between the two test conditions, or a combination thereof. In some embodiments, each set of candidate sensitivity parameters includes a posterior probability density function of the one or more sensitivity parameters. In some embodiments, each set of candidate sensitivity parameters includes a posterior probability density function of each of the one or more sensitivity parameters. In some embodiments, (b) includes summarizing sensitivity test data from the subject. In some embodiments, the system further includes, after (g), displaying sensitivity test results to the subject or the user, the results including a selected set of candidate sensitivity parameters. In some embodiments, the subject's sensitivity test data includes data only from the first test condition. In some embodiments, the subject's sensitivity test data includes data from both the first and second test conditions. In some embodiments, the selected set of candidate sensitivity parameters includes a first posterior probability density function of a sensitivity threshold or a sensitivity range. In some embodiments, b) includes selecting data having the first test condition. In some embodiments, the system includes, after g), selecting data having the second test condition from the subject's sensitivity test data; and repeating c) to g). In some embodiments, the selected set of candidate sensitivity parameters includes a second posterior probability density function of the sensitivity threshold or the sensitivity range.In some embodiments, the system further includes obtaining a difference distribution using the first posterior probability density function and the second posterior probability density function. In some embodiments, the system further includes calculating a change in a sensitivity threshold or sensitivity range between the first test condition and the second test condition based on the difference distribution. In some embodiments, the visual acuity is chart-specific or chart-invariant. In some embodiments, (e) is based on Bayes' theorem or maximum likelihood estimation. In some embodiments, the posterior probability density function of the method is a Bayesian posterior probability density function. In some embodiments, the first or second posterior probability density function of the method is a Bayesian posterior probability density function. In some embodiments, (d) includes: generating a single-opposite psychometric function, the single-opposite psychometric function including a first set of chart-specific sensitivity parameters; calculating a family of multi-opposite psychometric functions using the single-opposite psychometric function; and generating a chart-specific psychometric function using the family of multi-opposite psychometric functions, the chart-specific psychometric function including a second set of chart-specific sensitivity parameters. In some embodiments, the system further includes generating a sensitivity-based psychometric function comprising a first set of chart-invariant sensitivity parameters before generating the single-target psychometric function. In some embodiments, the sensitivity-based psychometric function is independent of one or more sensitivity chart design parameters. In some embodiments, the first set of chart-invariant sensitivity parameters includes a sensitivity threshold and a sensitivity range. In some embodiments, the sensitivity-based psychometric function is generated based on the one or more sensitivity chart design parameters and based on one or more additional parameters of the subject, the one or more additional parameters being chart-invariant. In some embodiments, the sensitivity-based psychometric function is generated based on signal detection theory. In some embodiments, the sensitivity-based psychometric function is configured to describe the subject's visual acuity performance as a d' function of one or more target sizes and is independent of the one or more chart design parameters. In some embodiments, the system further includes converting the sensitivity-based psychometric function into a single-target psychometric function. In some embodiments, the chart-specific sensitivity parameter set includes a sensitivity threshold and a sensitivity range. In some embodiments, the single-target psychometric function is chart-specific. In some embodiments, the chart-specific sensitivity parameter set includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold variation between the two test conditions, a sensitivity range variation between the two test conditions, a sensitivity slope variation between the two test conditions, or a combination thereof. In some embodiments, the chart-specific sensitivity parameter set includes one or more numerical values. In some embodiments, the chart-specific sensitivity parameter set includes at least one posterior probability density function for its parameters.In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, the chart-specific sensitivity parameter set includes a posterior probability density function for sensitivity thresholds under the at least two different test conditions. In some embodiments, the one or more sensitivity chart design parameters include: total number of optotypes, number of optotypes per row, optotype sampling method, optotype size, number of rows, and the subject's response or a combination thereof. In some embodiments, the single-opposite-scale psychometric function is based on at least one optotype size. In some embodiments, each function in the family of multi-opposite-scale psychometric functions is based on two or more different optotype sizes. In some embodiments, the single-opposite-scale psychometric function includes the subject's sensitivity threshold and sensitivity range. In some embodiments, one or more functions in the family of multi-opposite-scale psychometric functions are based on multiple optotypes per row in a sensitivity chart or sub-chart. In some embodiments, generating a chart-specific psychometric function further includes using a weighted sum of the family of multi-opposite-scale psychometric functions. In some embodiments, the weighted sum of the family of psychometric functions is based on the subject's sensitivity chart data. In some embodiments, the weighting is determined manually by the user or automatically by a computer program. In some embodiments, (b) further includes displaying the subject one or more sensitivity charts generated based on the one or more sensitivity chart design parameters. In some embodiments, each of the one or more sensitivity charts includes at least one optotype, wherein the at least one optotype is selected from a optotype library. In some embodiments, the contrast of the one or more sensitivity charts is approximately 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol. In some embodiments, the system further includes: h) generating a plurality of candidate sensitivity charts based on the one or more chart design parameters; and i) selecting one or more of the plurality of candidate sensitivity charts for sensitivity testing of the subject based on a ranking of the plurality of candidate sensitivity charts, wherein the ranking is based on a selected set of the subject's candidate sensitivity parameters. In some implementations, the system further includes, after (b), collecting data from the subject using the selected one or more candidate sensitivity charts; and repeating steps c) to g) and h) to i) if the stopping criteria are not met, until the stopping criteria are met.

[0020] On the other hand, this document discloses a computer-implemented system for adaptively evaluating the visual acuity of a subject, the system comprising: a digital processing device including an operating system and memory configured to execute executable instructions; and a computer program including instructions executable by the digital processing device to create a scoring application including software modules, the software modules being: a) obtaining one or more acuity chart design parameters; b) collecting acuity test data of the subject; c) generating one or more sets of candidate acuity parameters based on the one or more acuity chart design parameters and the subject's acuity test data, each of the one or more sets of candidate acuity parameters including a ranking obtained by analyzing the subject's acuity test data; d) generating a plurality of candidate acuity charts based on the one or more acuity chart design parameters; e) ranking the plurality of candidate acuity charts based on the ranking of each of the one or more sets of candidate acuity parameters; and f) selecting one or more acuity charts from the plurality of candidate acuity charts based on the ranking of the plurality of candidate acuity charts to be presented to the subject for acuity testing. In some embodiments, the system further includes presenting the one or more selected acuity charts to the subject after f). In some embodiments, the system further includes, after f), updating the sensitivity test data with new data collected from the subject using the one or more selected sensitivity charts. In some embodiments, the system further includes repeating: presenting the one or more selected sensitivity charts to the subject; updating the sensitivity test data with new data collected from the subject using the one or more selected sensitivity charts; and steps b) through f); until a stopping criterion is met. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes: a sensitivity threshold and a sensitivity range. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a change in the sensitivity threshold between the two test conditions, a change in the sensitivity range between the two test conditions, a change in the sensitivity slope between the two test conditions, or a combination thereof. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes a posterior probability density function for the one or more sensitivity parameters. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes a posterior probability density function for each of the one or more sensitivity parameters. In some embodiments, (b) includes summarizing sensitivity test data from the subject. In some embodiments, the subject's sensitivity test data includes data from only the first test condition. In some embodiments, the subject's sensitivity test data includes data from both the first and second test conditions.In some embodiments, each of one or more sets of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a change in the sensitivity threshold between the two test conditions, a change in the sensitivity range between the two test conditions, a change in the sensitivity slope between the two test conditions, or a combination thereof. In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, each of one or more sets of candidate sensitivity parameters includes a posterior probability density function of the sensitivity threshold or sensitivity range. In some embodiments, the one or more sensitivity chart design parameters include: the total number of optotypes, the number of optotypes per row, the optotype sampling method, the optotype size, the number of rows, and the subject's response, or a combination thereof. In some embodiments, each of the candidate sensitivity charts includes at least one optotype selected from a optotype library. In some embodiments, the contrast of one or more candidate sensitivity charts is approximately 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol.

[0021] Consistent with other disclosed implementations, a non-transitory computer-readable storage medium may store program instructions that are executed by a processor to perform any of the methods described herein.

[0022] Other aspects and advantages of this disclosure will become readily apparent to those skilled in the art in the following detailed description, in which only illustrative embodiments of the disclosure are shown and described. It will be appreciated that this disclosure is capable of other and different embodiments, and that modifications can be made to several details therein in various obvious respects, all without departing from this disclosure. Therefore, the drawings and description should be regarded, in fact, as illustrative rather than restrictive.

[0023] Incorporation

[0024] All publications, patents, and patent applications mentioned in this specification are incorporated herein by reference to the extent that each individual publication, patent, or patent application is specifically and individually identified as incorporated by reference. Where any publication, patent, or patent application incorporated by reference contradicts the disclosure contained in this specification, this specification is intended to supersede and / or give precedence to any such contradictory material. Attached Figure Description

[0025] A better understanding of the features and advantages of the invention will be obtained by referring to the following detailed description and accompanying drawings (also referred to herein as "the drawings" and "Figures") illustrating illustrative embodiments utilizing the principles of the invention, in which:

[0026] Figure 1 A non-restrictive example of a flowchart of the scoring algorithm S is shown;

[0027] Figures 2A-2B A non-limiting example of a library for generating visual acuity charts and subcharts with different chart designs is shown; in this case, the ETDRS chart (Early Treatment of Diabetic Retinopathy Study) is used. Figure 2A ) and Snellen charts ( Figure 2B );

[0028] Figure 3A A non-limiting example of a flowchart for generating a graph-invariant sensitivity model is shown;

[0029] Figure 3B A non-limiting example of a flowchart for generating a chart-specific sensitivity model is shown;

[0030] Figures 4A-4B A non-limiting example of a single-opposite psychometric function based on sensitivity (d') is shown, which is a psychometric function (in signal detection theory units of d') that is chart-invariant and describes visual sensitivity independent of decision-level factors affected by visual sensitivity chart design parameters;

[0031] Figure 5 A non-restrictive example of a set of chart-specific psychometric functions is shown, which is based on chart design factors that may include the number of response alternatives that determine the guessing rate (from two alternative forced choices (2AFC) to ten alternative forced choices (10AFC)), reflecting the transformation from chart-invariant single-opposite psychometric functions (in d' units) to chart-specific single-opposite psychometric functions;

[0032] Figures 6A-6B This illustrates a single-target psychometric function based on ten guesses and substitutions, according to the size of the single target. Figure 6A ) and different aggregated multi-opsight psychometric functions generated by the sensitivity model for identifying five optotypes presented on a slide. Figure 6B Non-restrictive examples.

[0033] Figure 6C It shows from Figure 6B A non-restricted example of a chart-specific psychometric function generated by the weighted summation of a family of multi-viewpoint psychometric functions;

[0034] Figures 7A-7B A non-limiting example of a chart-specific psychometric function defined by a chart-specific sensitivity threshold and a chart-specific sensitivity range based on the size of the optotypes is shown for identifying five optotypes presented on a slide.

[0035] Figures 8A-8B The single-target psychometric function is shown in terms of cumulative probability density function. Figure 8A ) and different aggregated multi-opinion psychometric functions ( Figure 8B Non-restrictive examples;

[0036] Figure 9 A non-limiting example of the flowchart of the adaptive algorithm A disclosed herein is shown;

[0037] Figure 10 A non-limiting example of an extended library that generates candidate visual acuity charts by determining a sampling pattern that varies the size of the optotype, based on chart design principles;

[0038] Figures 11A-11F A non-limiting example of a sensitivity chart using candidate visual sensitivity charts with different anchor points is shown, based on complete standard constraints;

[0039] Figures 12A-12C A non-limiting example of a specific area or sub-chart of a visual acuity chart showing the full standard constraints is shown;

[0040] Figures 13A-13C A non-limiting example of collecting responses from subjects in three different modes is shown, in which the subject reads down from the first column ( Figure 13A ), Read the chart line by line downwards ( Figure 13B ) or read the single letter displayed in the center of the screen. Figure 13C );

[0041] Figures 14A-14B A non-limiting example comparing the computational results of single-row optimization and compound multi-row subgraph optimization is shown;

[0042] Figure 15 A non-limiting example of a personalized sensitivity chart for selecting sequential test slides (n+1 in a series) based on the subject’s response to the nth test slide is shown.

[0043] Figure 16 A non-limiting example of a flowchart using a combination of algorithms A and S as disclosed herein is shown.

[0044] Figure 17 Another non-limiting example of a flowchart using a combination of algorithms A and S as disclosed herein is shown.

[0045] Figure 18 A non-limiting example of using Bayesian posterior probability to estimate a subject's sensitivity threshold is shown, in which three slides are used. The illustration shows that the sensitivity estimate converges as the number of test slides increases.

[0046] Figure 19 Non-restrictive examples of variability variations with different numbers of test slides are shown in an adaptive sensitivity test with standard constraints having 1 to 5 slides and four different runs of each test.

[0047] Figure 20 A non-restrictive example of the verification of the combinatorial algorithm is shown.

[0048] Figure 21 A non-restrictive example of the verification of the combinatorial algorithm is shown.

[0049] Figures 22A-22D A non-limiting example showing the results of a visual acuity test is presented.

[0050] Figure 23 Another non-limiting example of showing the results of a visual acuity test to a subject is shown.

[0051] Figure 24 Non-limiting examples of digital processing devices as disclosed herein are shown.

[0052] Figure 25 A non-limiting schematic diagram of a web / mobile application delivery system is shown; in this case, the system provides a browser-based and / or locally mobile user interface;

[0053] Figure 26 A non-limiting schematic diagram of a cloud-based web / mobile application delivery system is shown; in this case, the system includes elastically load-balanced, auto-scaling web server and application server resources, as well as synchronously replicated databases;

[0054] Figures 27A-27B This paper presents a non-limiting example of estimating visual acuity variations between different test conditions using the Bayesian posterior probability density function (pdf); and

[0055] Figures 28A-28B A non-restricted example of the difference distribution of the Bayesian posterior probability density function under different test conditions is shown, which provides an indicator of the change in sensitivity threshold. Detailed Implementation

[0056] The methods, systems, and apparatus disclosed herein relate to visual acuity assessment using visual acuity testing.

[0057] The systems and methods disclosed herein offer a wide range of advantages compared to existing sensitivity testing and analysis methods or systems. Unlike existing adaptive sensitivity testing methods that optimize optotype selection for only one letter of a single size, the current systems and methods can optimize for the overall optotypes presented on each test slide. The systems and methods disclosed herein enable the simultaneous estimation of both chart-specific and chart-invariant sensitivity parameters or indices. Additional advantages associated with the methods and systems disclosed herein may include: adherence to chart design standards while accurately focusing on each individual subject, rapid testing time, accuracy comparable to time-consuming single-letter sensitivity tests, and visual sensitivity independent of optotype size, test chart, or testing procedure. The sets of chart-specific and chart-invariant sensitivity parameters described herein are interchangeable and mutually translatable.

[0058] In some embodiments, this document discloses a computer-implemented method for generating a sensitivity model that scores the visual acuity of a subject, the method comprising: a) obtaining one or more sensitivity chart design parameters; b) collecting sensitivity test data of the subject; c) selecting a first set of chart-specific sensitivity parameters from one or more sets of candidate sensitivity parameters; d) generating a sensitivity model, comprising: i) generating a single-opposite-target psychometric function, the single-opposite-target psychometric function including the first set of chart-specific sensitivity parameters; ii) calculating a family of multi-opposite-target psychometric functions using the single-opposite-target psychometric function; and iii) generating a chart-specific psychometric function using the family of multi-opposite-target psychometric functions, the chart-specific psychometric function including a second set of chart-specific sensitivity parameters. The sensitivity model is configured to estimate the probability of obtaining the subject's sensitivity test data based on the one or more sensitivity chart design parameters. In some embodiments, the method further includes: after (c) and before (i), generating a sensitivity-based psychometric function comprising a first set of chart-invariant sensitivity parameters, wherein the first set of chart-invariant sensitivity parameters includes a sensitivity threshold and a sensitivity range, wherein the first set of chart-invariant sensitivity parameters is generated based on a first set of chart-specific sensitivity parameters. In some embodiments, the sensitivity-based psychometric function is independent of one or more sensitivity chart design parameters. In some embodiments, the sensitivity-based psychometric function is generated based on the one or more sensitivity chart design parameters and one or more additional parameters of the subject, the one or more additional parameters being chart-invariant. In some embodiments, the sensitivity-based psychometric function is generated based on signal detection theory. In some embodiments, the sensitivity-based psychometric function is configured to describe the subject's visual acuity performance as a d' function of one or more optotype sizes, and is independent of the one or more chart design parameters. In some embodiments, the method further includes, before (i), converting the sensitivity-based psychometric function into a single optotype psychometric function. In some embodiments, the first set of chart-specific sensitivity parameters includes a sensitivity threshold and a sensitivity range. In some embodiments, the single-viewpoint psychometric function is chart-specific. In some embodiments, the second set of chart-specific sensitivity parameters includes a sensitivity threshold and a sensitivity range. In some embodiments, the first or second set of chart-specific sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a change in the sensitivity threshold between the two test conditions, a change in the sensitivity range between the two test conditions, a change in the sensitivity slope between the two test conditions, or a combination thereof. In some embodiments, the first or second set of chart-specific sensitivity parameters includes one or more numerical values.In some embodiments, the first or second set of chart-specific sensitivity parameters includes at least one posterior probability density function for its parameters. In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, the first or second set of chart-specific sensitivity parameters includes a posterior probability density function of sensitivity thresholds from the at least two different test conditions. In some embodiments, the one or more sensitivity chart design parameters include: total number of optotypes, number of optotypes per row, optotype sampling method, optotype size, number of rows, and the subject's response or a combination thereof. In some embodiments, the single-optotype psychometric function is based on at least one optotype size. In some embodiments, each function in the family of multi-optotype psychometric functions is based on two or more different optotype sizes. In some embodiments, the single-optotype psychometric function includes the subject's sensitivity threshold and sensitivity range. In some embodiments, one or more functions in the family of multi-optotype psychometric functions are based on multiple optotypes per row in a sensitivity chart or sub-chart. In some embodiments, (iii) further includes using a weighted sum of the family of multi-optotype psychometric functions. In some embodiments, the weighting of the family of psychometric functions is based on the subject's sensitivity chart data. In some embodiments, the weighting is determined manually by the user or automatically by a computer program. In some embodiments, (b) further includes displaying the subject one or more sensitivity charts generated based on the one or more sensitivity chart design parameters. In some embodiments, each of the one or more sensitivity charts includes at least one optotype. In some embodiments, the contrast of the one or more sensitivity charts is approximately 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol. In some embodiments, one or more sets of candidate sensitivity parameters are generated based on the one or more sensitivity chart design parameters, the subject's sensitivity test data, or both.

[0059] In some embodiments, this document discloses a computer-implemented method for scoring the visual acuity of a subject, the method comprising: a) obtaining one or more acuity chart design parameters; b) collecting acuity test data of the subject; c) generating one or more sets of candidate acuity parameters based on the one or more acuity chart design parameters, the subject's acuity test data, or both; d) generating a acuity model, including generating one or more chart-specific psychometric functions, each chart-specific psychometric function including a set of chart-specific acuity parameters from the one or more sets of the chart-specific acuity parameters; e) generating the probability of observing the acuity test data using the acuity model and the one or more sets of candidate acuity parameters; f) ranking the one or more sets of candidate acuity parameters of the subject based on the probability; and g) selecting a set from the one or more sets of candidate acuity parameters of the subject based on the ranking. In some embodiments, each set of candidate acuity parameters includes: a acuity threshold and a acuity range. In some embodiments, each set of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between the two test conditions, a sensitivity range change between the two test conditions, a sensitivity slope change between the two test conditions, or a combination thereof. In some embodiments, each set of candidate sensitivity parameters includes a posterior probability density function of the one or more sensitivity parameters. In some embodiments, each set of candidate sensitivity parameters includes a posterior probability density function of each of the one or more sensitivity parameters. In some embodiments, (b) includes summarizing sensitivity test data from the subject. In some embodiments, the method further includes, after (g), displaying sensitivity test results to the subject or the user, the test results including a selected set of candidate sensitivity parameters. In some embodiments, the subject's sensitivity test data includes data only from a first test condition. In some embodiments, the subject's sensitivity test data includes data from both a first and a second test condition. In some embodiments, the selected set of candidate sensitivity parameters includes a first posterior probability density function of a sensitivity threshold or a sensitivity range. In some embodiments, b) includes selecting data having the first test condition. In some embodiments, the method further includes, after g), selecting data having the second test condition from the subject's sensitivity test data; and repeating c) to g). In some embodiments, the selected set of candidate sensitivity parameters includes a second posterior probability density function of the sensitivity threshold or the sensitivity range. In some embodiments, the method further includes obtaining a difference distribution using the first and second posterior probability density functions.In some embodiments, the method further includes calculating a change in a sensitivity threshold or sensitivity range between the first test condition and the second test condition based on the difference distribution. In some embodiments, the visual acuity is chart-specific or chart-invariant. In some embodiments, (e) is based on Bayes' theorem or maximum likelihood estimation. In some embodiments, the posterior probability density function is a Bayesian posterior probability density function. In some embodiments, the first or second posterior probability density function is a Bayesian posterior probability density function. In some embodiments, (d) includes: generating a single-opposite psychometric function, the single-opposite psychometric function including a first set of chart-specific sensitivity parameters; calculating a family of multi-opposite psychometric functions using the single-opposite psychometric function; and generating a chart-specific psychometric function using the family of multi-opposite psychometric functions, the specific psychometric function including a second set of chart-specific sensitivity parameters. In some embodiments, the method further includes generating a sensitivity-based psychometric function including a first set of chart-invariant sensitivity parameters before generating the single-opposite psychometric function. In some embodiments, the sensitivity-based psychometric function is independent of one or more sensitivity chart design parameters. In some embodiments, the first set of chart-invariant sensitivity parameters includes a sensitivity threshold and a sensitivity range, and the first set of chart-invariant sensitivity parameters is generated based on one or more sets of candidate sensitivity parameters. In some embodiments, the sensitivity-based psychometric function is generated based on the one or more sensitivity chart design parameters and one or more additional parameters of the subject, the one or more additional parameters being chart-invariant. In some embodiments, the sensitivity-based psychometric function is generated based on signal detection theory. In some embodiments, the sensitivity-based psychometric function is configured to describe the subject's visual acuity performance as a d' function of one or more optotype sizes, and is independent of the one or more chart design parameters. In some embodiments, the method further includes converting the sensitivity-based psychometric function into a single optotype psychometric function. In some embodiments, the chart-specific sensitivity parameter set includes a sensitivity threshold and a sensitivity range. In some embodiments, the single optotype psychometric function is chart-specific. In some embodiments, the chart-specific sensitivity parameter set includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold variation between the two test conditions, a sensitivity range variation between the two test conditions, a sensitivity slope variation between the two test conditions, or a combination thereof. In some embodiments, the chart-specific sensitivity parameter set includes one or more numerical values. In some embodiments, the chart-specific sensitivity parameter set includes at least one posterior probability density function for its parameters.In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, the chart-specific sensitivity parameter set includes posterior probability density functions of sensitivity thresholds from the at least two different test conditions. In some embodiments, the one or more sensitivity chart design parameters include: total number of optotypes, number of optotypes per row, optotype sampling method, optotype size, number of rows, and the subject's response or a combination thereof. In some embodiments, the single-opposite-scale psychometric function is based on at least one optotype size. In some embodiments, each function in the family of multi-opposite-scale psychometric functions is based on two or more different optotype sizes. In some embodiments, the single-opposite-scale psychometric function includes the subject's sensitivity threshold and sensitivity range. In some embodiments, one or more functions in the family of multi-opposite-scale psychometric functions are based on multiple optotypes per row in a sensitivity chart or sub-chart. In some embodiments, generating the chart-specific psychometric function further includes using a weighted sum of the family of multi-opposite-scale psychometric functions. In some embodiments, the weighting of the family of psychometric functions is based on the subject's sensitivity chart data. In some embodiments, the weighting is determined manually by the user or automatically by a computer program. In some embodiments, (b) further includes displaying the subject one or more sensitivity charts generated based on the one or more sensitivity chart design parameters. In some embodiments, each of the one or more sensitivity charts includes at least one optotype. In some embodiments, the contrast of the one or more sensitivity charts is about 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol. In some embodiments, the method includes: h) generating a plurality of candidate sensitivity charts based on the one or more chart design parameters; and i) selecting one or more of the plurality of candidate sensitivity charts for the subject's sensitivity test based on the ranking of the plurality of candidate sensitivity charts, wherein the ranking is based on a selected set of the subject's candidate sensitivity parameters. In some embodiments, the method further includes, after (b); collecting data from the subject using the selected one or more candidate sensitivity charts; and repeating steps c) to g) and h) to i) if the stopping criteria are not met, until the stopping criteria are met.

[0060] In some embodiments, this document discloses a computer-implemented method for adaptively evaluating the visual acuity of a subject, the method comprising: a) obtaining one or more acuity chart design parameters; b) collecting acuity test data of the subject; c) generating one or more sets of candidate acuity parameters based on the one or more acuity chart design parameters and the subject's acuity test data, each of the one or more sets of candidate acuity parameters including a ranking obtained by analyzing the subject's acuity test data; d) generating a plurality of candidate acuity charts based on the one or more acuity chart design parameters; e) ranking the plurality of candidate acuity charts based on the ranking of each of the one or more sets of candidate acuity parameters; and f) selecting one or more acuity charts from the plurality of candidate acuity charts based on the ranking of the plurality of candidate acuity charts to be presented to the subject for acuity testing. In some embodiments, the method further includes, after f), presenting the one or more selected acuity charts to the subject. In some embodiments, the method further includes, after f), updating the acuity test data with new data collected from the subject using the one or more selected acuity charts. In some embodiments, the method further includes repeating: presenting the subject with the one or more selected sensitivity charts; updating the sensitivity test data with new data collected from the subject using the one or more selected sensitivity charts, and steps b) to f); until a stopping criterion is met. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes: a sensitivity threshold and a sensitivity range. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a change in the sensitivity threshold between the two test conditions, a change in the sensitivity range between the two test conditions, a change in the sensitivity slope between the two test conditions, or a combination thereof. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes a posterior probability density function for the one or more sensitivity parameters. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes a posterior probability density function for each of the one or more sensitivity parameters. In some embodiments, (b) includes summarizing the sensitivity test data from the subject. In some embodiments, the subject's sensitivity test data includes data only from a first test condition. In some implementations, the sensitivity test data of the subject includes data from a first test condition and a second test condition.In some embodiments, each of one or more sets of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a change in the sensitivity threshold between the two test conditions, a change in the sensitivity range between the two test conditions, a change in the sensitivity slope between the two test conditions, or a combination thereof. In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, each of one or more sets of candidate sensitivity parameters includes a posterior probability density function of the sensitivity threshold or sensitivity range. In some embodiments, the one or more sensitivity chart design parameters include: the total number of optotypes, the number of optotypes per row, the optotype sampling method, the optotype size, the number of rows, and the subject's response, or a combination thereof. In some embodiments, each of the candidate sensitivity charts includes at least one optotype. In some embodiments, the contrast of one or more candidate sensitivity charts is approximately 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol.

[0061] In some embodiments, this document discloses a computer-implemented system for generating a sensitivity model that scores the visual acuity of a subject. The system includes: a digital processing device including an operating system and memory configured to execute executable instructions; and a computer program including instructions executable by the digital processing device to create a scoring application including software modules that: a) obtain one or more sensitivity chart design parameters; b) collect sensitivity test data of the subject; c) select a first set of chart-specific sensitivity parameters from one or more sets of candidate sensitivity parameters; and d) generate a sensitivity model comprising: i. generating a single-opposite mental measurement function including the first set of chart-specific sensitivity parameters; ii. calculating a family of multi-opposite mental measurement functions using the single-opposite mental measurement function; and iii. generating a chart-specific mental measurement function using the family of multi-opposite mental measurement functions including a second set of chart-specific sensitivity parameters. The sensitivity model is configured to estimate the probability of obtaining the subject's sensitivity test data based on the one or more sensitivity chart design parameters. In some embodiments, the system further includes: after (c) and before (i), generating a sensitivity-based psychometric function comprising a first set of chart-invariant sensitivity parameters, wherein the first set of chart-invariant sensitivity parameters includes a sensitivity threshold and a sensitivity range, and wherein the first set of chart-invariant sensitivity parameters is generated based on a first set of chart-specific sensitivity parameters. In some embodiments, the sensitivity-based psychometric function is independent of one or more sensitivity chart design parameters. In some embodiments, the sensitivity-based psychometric function is generated based on the one or more sensitivity chart design parameters and one or more additional parameters of the subject, the one or more additional parameters being chart-invariant. In some embodiments, the sensitivity-based psychometric function is generated based on signal detection theory. In some embodiments, the sensitivity-based psychometric function is configured to describe the subject's visual acuity performance as a d' function of one or more optotype sizes, and is independent of the one or more chart design parameters. In some embodiments, the system further includes, before (i), converting the sensitivity-based psychometric function into a single optotype psychometric function. In some embodiments, the first set of chart-specific sensitivity parameters includes a sensitivity threshold and a sensitivity range. In some implementations, the single-target psychometric function is chart-specific. In some implementations, the second set of chart-specific sensitivity parameters includes sensitivity thresholds and sensitivity ranges.In some embodiments, the first or second set of chart-specific sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a change in the sensitivity threshold between the two test conditions, a change in the sensitivity range between the two test conditions, a change in the sensitivity slope between the two test conditions, or a combination thereof. In some embodiments, the first or second set of chart-specific sensitivity parameters includes one or more numerical values. In some embodiments, the first or second set of chart-specific sensitivity parameters includes at least one posterior probability density function for its parameters. In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, the first or second set of chart-specific sensitivity parameters includes a posterior probability density function of the sensitivity thresholds from the at least two different test conditions. In some embodiments, the one or more sensitivity chart design parameters include: the total number of optotypes, the number of optotypes per row, the optotype sampling method, the optotype size, the number of rows, and the subject's response, or a combination thereof. In some embodiments, the single-opotype psychometric function is based on at least one optotype size. In some embodiments, each function in the family of multi-opotype psychometric functions is based on two or more different optotype sizes. In some embodiments, the single-opposite psychometric function includes the subject's sensitivity threshold and sensitivity range. In some embodiments, one or more functions in the family of multi-opposite psychometric functions are based on multiple optotypes per row in a sensitivity chart or sub-chart. In some embodiments, (iii) further includes using a weighted sum of the family of multi-opposite psychometric functions. In some embodiments, the weighting of the family of psychometric functions is based on the subject's sensitivity chart data. In some embodiments, the weighting is determined manually by a user or automatically by a computer program. In some embodiments, (b) further includes using one or more sensitivity charts generated based on the design parameters of the one or more sensitivity charts. In some embodiments, each of the one or more sensitivity charts includes at least one optotype. In some embodiments, the contrast of the one or more sensitivity charts is approximately 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol. In some embodiments, one or more sets of candidate sensitivity parameters are generated based on the one or more sensitivity chart design parameters, the subject's sensitivity test data, or both.

[0062] In some embodiments, this document discloses a computer system for generating a sensitivity model that scores the visual acuity of a subject, the system comprising: a digital processing device including an operating system and a memory configured to execute executable instructions; and a computer program including instructions executable by the digital processing device to create a scoring application including software modules, the software modules being: a) obtaining one or more sensitivity chart design parameters; b) collecting sensitivity test data of the subject; c) generating one or more sets of candidate sensitivity parameters based on the one or more sensitivity chart design parameters, the subject's sensitivity test data, or both; d) generating a sensitivity model, including generating one or more chart-specific psychometric functions, each chart-specific psychometric function including a set of chart-specific sensitivity parameters from the one or more sets of chart-specific sensitivity parameters; e) generating the probability of observing the sensitivity test data using the sensitivity model and the one or more sets of candidate sensitivity parameters; f) ranking the one or more sets of candidate sensitivity parameters of the subject based on the probability; and g) selecting a set from the one or more sets of candidate sensitivity parameters of the subject based on the ranking. In some embodiments, each set of candidate sensitivity parameters includes a sensitivity threshold and a sensitivity range. In some embodiments, each set of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between the two test conditions, a sensitivity range change between the two test conditions, a sensitivity slope change between the two test conditions, or a combination thereof. In some embodiments, each set of candidate sensitivity parameters includes a posterior probability density function of the one or more sensitivity parameters. In some embodiments, each set of candidate sensitivity parameters includes a posterior probability density function of each of the one or more sensitivity parameters. In some embodiments, (b) includes summarizing sensitivity test data from the subject. In some embodiments, the system further includes, after (g), displaying sensitivity test results to the subject or the user using a digital display, the results including a selected set of candidate sensitivity parameters. In some embodiments, the subject's sensitivity test data includes data only from the first test condition. In some embodiments, the subject's sensitivity test data includes data from both the first and second test conditions. In some embodiments, the selected set of candidate sensitivity parameters includes a first posterior probability density function of a sensitivity threshold or a sensitivity range. In some embodiments, b) includes selecting data having the first test condition. In some embodiments, the system includes, after g), selecting data having the second test condition from the subject's sensitivity test data; and repeating c) to g). In some embodiments, the selected set of candidate sensitivity parameters includes a second posterior probability density function of the sensitivity threshold or the sensitivity range.In some embodiments, the system further includes obtaining a difference distribution using the first posterior probability density function and the second posterior probability density function. In some embodiments, the system further includes calculating a change in a sensitivity threshold or sensitivity range between the first test condition and the second test condition based on the difference distribution. In some embodiments, the visual acuity is chart-specific or chart-invariant. In some embodiments, (e) is based on Bayes' theorem or maximum likelihood estimation. In some embodiments, the posterior probability density function of the method is a Bayesian posterior probability density function. In some embodiments, the first or second posterior probability density function of the method is a Bayesian posterior probability density function. In some embodiments, (d) includes: generating a single-opposite psychometric function, the single-opposite psychometric function including a first set of chart-specific sensitivity parameters; calculating a family of multi-opposite psychometric functions using the single-opposite psychometric function; and generating a chart-specific psychometric function using the family of multi-opposite psychometric functions, the chart-specific psychometric function including a second set of chart-specific sensitivity parameters. In some embodiments, the system further includes generating a sensitivity-based psychometric function comprising a first set of chart-invariant sensitivity parameters before generating the single-target psychometric function. In some embodiments, the sensitivity-based psychometric function is independent of one or more sensitivity chart design parameters. In some embodiments, the first set of chart-invariant sensitivity parameters includes a sensitivity threshold and a sensitivity range. In some embodiments, the sensitivity-based psychometric function is generated based on the one or more sensitivity chart design parameters and based on one or more additional parameters of the subject, the one or more additional parameters being chart-invariant. In some embodiments, the sensitivity-based psychometric function is generated based on signal detection theory. In some embodiments, the sensitivity-based psychometric function is configured to describe the subject's visual acuity performance as a d' function of one or more target sizes and is independent of the one or more chart design parameters. In some embodiments, the system further includes converting the sensitivity-based psychometric function into a single-target psychometric function before (i). In some embodiments, the chart-specific sensitivity parameter set includes a sensitivity threshold and a sensitivity range. In some embodiments, the single-target psychometric function is chart-specific. In some embodiments, the chart-specific sensitivity parameter set includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold variation between the two test conditions, a sensitivity range variation between the two test conditions, a sensitivity slope variation between the two test conditions, or a combination thereof. In some embodiments, the chart-specific sensitivity parameter set includes one or more numerical values. In some embodiments, the chart-specific sensitivity parameter set includes at least one posterior probability density function for its parameters.In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, the chart-specific sensitivity parameter set includes a posterior probability density function for sensitivity thresholds under the at least two different test conditions. In some embodiments, the one or more sensitivity chart design parameters include: total number of optotypes, number of optotypes per row, optotype sampling method, optotype size, number of rows, and the subject's response or a combination thereof. In some embodiments, the single-opposite-scale psychometric function is based on at least one optotype size. In some embodiments, each function in the family of multi-opposite-scale psychometric functions is based on two or more different optotype sizes. In some embodiments, the single-opposite-scale psychometric function includes the subject's sensitivity threshold and sensitivity range. In some embodiments, one or more functions in the family of multi-opposite-scale psychometric functions are based on multiple optotypes per row in a sensitivity chart or sub-chart. In some embodiments, generating a chart-specific psychometric function further includes using a weighted sum of the family of multi-opposite-scale psychometric functions. In some embodiments, the weighted sum of the family of psychometric functions is based on the subject's sensitivity chart data. In some embodiments, the weighting is determined manually by the user or automatically by a computer program. In some embodiments, (b) further includes displaying one or more sensitivity charts generated based on the one or more sensitivity chart design parameters. In some embodiments, each of the one or more sensitivity charts includes at least one optotype. In some embodiments, the contrast of the one or more sensitivity charts is approximately 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol. In some embodiments, the system further includes: h) generating a plurality of candidate sensitivity charts based on one or more chart design parameters; and i) selecting one or more of the plurality of candidate sensitivity charts for sensitivity testing of the subject based on a ranking of the plurality of candidate sensitivity charts, wherein the ranking is based on a selected set of the candidate sensitivity parameters of the subject. In some embodiments, the system further includes, after (b); collecting data from the subject using the selected one or more candidate sensitivity charts; and repeating steps c) to g) and h) to i) if the stopping criteria are not met until the stopping criteria are met.

[0063] In some embodiments, this document discloses a computer-implemented system for adaptively evaluating the visual acuity of a subject, the system comprising: a digital processing device including an operating system and memory configured to execute executable instructions; and a computer program including instructions executable by the digital processing device to create a scoring application including software modules, the software modules being: a) obtaining one or more acuity chart design parameters; b) collecting acuity test data of the subject; c) generating one or more sets of candidate acuity parameters based on the one or more acuity chart design parameters and the subject's acuity test data, each of the one or more sets of candidate acuity parameters including a ranking obtained by analyzing the subject's acuity test data; d) generating a plurality of candidate acuity charts based on the one or more acuity chart design parameters; e) ranking the plurality of candidate acuity charts based on the ranking of each of the one or more sets of candidate acuity parameters; and f) selecting one or more acuity charts from the plurality of candidate acuity charts based on the ranking of the plurality of candidate acuity charts to be presented to the subject for acuity testing. In some embodiments, the system further includes, after f), presenting the one or more selected sensitivity charts to the subject. In some embodiments, the system further includes, after f), updating the sensitivity test data with new data collected from the subject using the one or more selected sensitivity charts. In some embodiments, the system further includes repeating: presenting the one or more selected sensitivity charts to the subject; updating the sensitivity test data with new data collected from the subject using the one or more selected sensitivity charts; and steps b) to f); until a stopping criterion is met. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes: a sensitivity threshold and a sensitivity range. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between the two test conditions, a sensitivity range change between the two test conditions, a sensitivity slope change between the two test conditions, or a combination thereof. In some embodiments, each of the one or more sets of candidate sensitivity parameters includes a posterior probability density function for each of the one or more sensitivity parameters. In some embodiments, each of one or more sets of candidate sensitivity parameters includes a posterior probability density function for each of the one or more sensitivity parameters. In some embodiments, (b) includes summarizing sensitivity test data from the subject. In some embodiments, the subject's sensitivity test data includes data from only a first test condition. In some embodiments, the subject's sensitivity test data includes data from both a first and a second test condition.In some embodiments, each of one or more sets of candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a change in the sensitivity threshold between the two test conditions, a change in the sensitivity range between the two test conditions, a change in the sensitivity slope between the two test conditions, or a combination thereof. In some embodiments, the subject's sensitivity test data includes data from at least two different test conditions. In some embodiments, each of one or more sets of candidate sensitivity parameters includes a posterior probability density function of the sensitivity threshold or sensitivity range. In some embodiments, the one or more sensitivity chart design parameters include: the total number of optotypes, the number of optotypes per row, the optotype sampling method, the optotype size, the number of rows, and the subject's response, or a combination thereof. In some embodiments, each of the candidate sensitivity charts includes at least one optotype. In some embodiments, the contrast of one or more candidate sensitivity charts is approximately 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart. In some embodiments, the at least one optotype is a letter, number, or symbol.

[0064] Target

[0065] In some implementations, optotypes are standardized symbols used to test visual acuity or visual acuity. In some implementations, visual acuity charts or visual acuity tests follow one or more standards selected from the following: American National Standards Institute, National Academy of Sciences, National Research Council, American Academy of Ophthalmology, National Eye Institute, American Academy of Optometry, and Food and Drug Administration.

[0066] In some embodiments, the target is a letter, graphic, number, photograph, or geometric symbol of any shape. In some embodiments, the target has a predetermined size. In some embodiments, the minimum size of the target generates a visual acuity of -0.3 logMAR. In some embodiments, the minimum size of the target generates visual acuity of -0.25, -0.35, -0.4, -0.45, and -0.5 logMAR. In some embodiments, the maximum size of the target generates visual acuity of 1.2, 1.25, 1.3, 1.5, and 1.55 logMAR.

[0067] Visual acuity charts and subcharts

[0068] In some embodiments, a visual acuity chart (or interchangeably herein referred to as a “chart,” “acuity chart,” or “visual chart”) is used to measure the spatial resolution of a subject’s visual system. In some embodiments, a visual subchart is part of a full visual chart. In some embodiments, a visual chart is a visual acuity chart, a visual test chart, or a visual acuity test chart, etc. In some embodiments, a visual chart or subchart includes one or more optotypes. In some embodiments, visual charts are categorized into different types based on at least one chart parameter or visual acuity test parameter. These parameters may include one or more selected from: optotype type, number of optotypes, number of rows, number of columns, optotype size, difference in optotype size between adjacent rows or columns, number of test charts in a visual test, or optotype sampling method, etc. In some embodiments, a visual chart is an ETDRS (Early Treatment Diabetic Retinopathy Study) logMAR (logarithm of minimum resolvable angles on the x and y axes) chart, a Snellen chart, a flipped E chart, a Landolt C chart, a Lea test, or a Jaeger chart.

[0069] In some embodiments, a visual chart or subchart includes one or more icons. In some embodiments, at least one icon forms a visual test chart or subchart. In some embodiments, at least one row of icons, each row including at least one icon, forms a visual test chart or subchart. In some embodiments, at least one column of icons, each column including at least one icon, forms a visual test chart or subchart. In some embodiments, the different icons in each row have the same icon size. In some embodiments, the icon size monotonically decreases from the first row to the last row of the visual chart or visual subchart. In some embodiments, the number of icons in each row of a visual chart or subchart is the same. In some embodiments, each row has 5 icons. In some embodiments, the number of icons in each row of a visual chart or subchart ranges from 1 to 5 icons. In some embodiments, the chart or subchart has 3 rows of icons. In some embodiments, the chart or subchart has 1 to 5 rows of icons. In some embodiments, the full visual chart has 14 rows of icons.

[0070] In some implementations, the target size for each row is specified as the base-10 logarithm of the decimal visual acuity. As an example, a decimal visual acuity of 1 is 0.00 in LogMAR sensitivity, and a decimal visual acuity of 0.1 is 1.0 in LogMAR sensitivity.

[0071] Sensitivity Chart Design

[0072] In some embodiments, the methods, systems, or apparatuses disclosed herein include sensitivity charts and subcharts. In some embodiments, the sensitivity chart or subchart is constrained by one or more chart-specific design features and / or target design features.

[0073] In some embodiments, in the systems and methods disclosed herein, design features of the target are considered in addition to design features used to present the target's sensitivity chart (or sub-chart). In further embodiments, the target's parameters (which may be interchangeable with features) and / or the charts presenting the target's parameters include, but are not limited to:

[0074] i. The number of possible target sizes (e.g., 11 target sizes for the Snellen chart and 14 target sizes for the ETDRS chart);

[0075] ii. Possible target size (e.g., Figures 11A-11F )

[0076] ii. Definition of sight target:

[0077] 1. Letters

[0078] 2. Numbers

[0079] 3. Pattern

[0080] 4. Landolt C

[0081] 5. Flip E;

[0082] iii. Size of the pool used for sampling targets

[0083] 1.Sloan, Bailey-Lovie is 10;

[0084] 2.Landolt C, Tumbling E, HOTV, Lea and Hands are 4;

[0085] 3. The traditional Snellen score is 9;

[0086] 4. The British Snellen standard BS 4274-1:2003 specifies 12;

[0087] 5. The English alphabet has 26 letters;

[0088] iv. Methods used for sampling the target pool (e.g., are there alternatives);

[0089] v. The number of rows presented in a chart, subchart, or series of subcharts;

[0090] vi. The number of views displayed per line;

[0091] vii. Definition of sensitivity metrics (e.g., sensitivity threshold, sensitivity range, and / or sensitivity slope);

[0092] Current techniques in clinical care have focused on the nominal goal of sensitivity testing: to determine that patients cannot reliably identify the row of optotypes according to some predetermined criteria: for example, they cannot identify the size of at least two out of five optotypes. Current testing ignores several design factors that influence visual performance across different chart designs. For example, it is relatively easy to correctly guess the smaller pool of optotypes even without visual information. Specifically, when testing sensitivity using Landolt C optotypes (which identify one of four potential directions), it is easier to guess correctly (25% chance) compared to sensitivity charts with lower guessing rates (e.g., 10% for the Sloan set). Currently, sensitivity results are unacceptable or incorrect for guessing rates across different charts. This can confuse the contribution of chart-specific design factors with patient-specific factors to performance.

[0093] In some implementations, the disclosure herein includes a comprehensive consideration of one or more of the characteristics of a sensitivity chart design, including chart-specific and subject-specific parameters (equivalent to chart-invariant parameters) that form the basis of the sensitivity model or predictive model herein. By using a sensitivity chart design to precisely define the contribution to performance and considering the different contributions of each chart to performance, the model can achieve a measurement of subject sensitivity performance independent of the chart used to collect sensitivity data.

[0094] In some implementations, visual acuity testing conforms to design principles based on committee standards and / or historical and cultural preferences. (See reference...) Figure 2A In a particular implementation, the visual acuity chart design recommended by the ETDRS committee comprises fourteen rows of optotypes, with five optotypes per row, and the size of each row decreasing logarithmically. A .10 decimal logarithmic unit series provides the size to decrease by approximately 25% with each successive row of optotypes. Each row contains five letters sampled from a 10-letter Sloan combination without substitution; these letters are... Figure 2A The diagram shows visual acuity charts that meet the ETDRS standard.

[0095] In some implementations, Snellen charts are used to display more letters at smaller viewpoint sizes and fewer letters at larger viewpoint sizes. Figure 2BThe standard constrained chart remains the dominant chart used in clinical practice. In some implementations, the ETDRS, Snellen, and similar standard constrained chart designs involve a static range and resolution for sampling target size. Due to the non-adaptive deterministic testing procedure, the same range of targets is presented to all subjects. Furthermore, the invariant sampling resolution of the test results in insensitivity to a measured visual acuity value corresponding to the target size, which is the median of the target sizes presented on fourteen established target sizes.

[0096] In some implementations, visual acuity charts may conform to design principles defined by standards committees and / or historical preferences. For example, the ETDRS chart is a recommended visual acuity chart for use in clinical trials and clinical research. Figure 2A The ETDRS chart is designed with 14 rows of views, each row representing 5 views sampled from the 10-letter Sloan set without replacement, and the view sizes between rows progress in a constant logarithmic order. A 10-decimal-logarithmic series corresponds to an increase in size of approximately 25% from each row to the row above it.

[0097] Design of Snellen sensitivity charts ( Figure 2B It consists of twelve rows, each displaying a different number of optotypes, and the rows vary in size. Although its design features are considered inferior to those of the ETDRS chart, the Snellen chart remains the dominant sensitivity chart design used by eye care professionals and general practitioners in clinical practice.

[0098] In some implementations, to improve the sampling resolution of target size in visual acuity testing while strictly adhering to visual acuity design standards and principles, a method using an extended library for generating visual acuity charts is employed. (See also...) Figure 10In a particular embodiment, the topmost pattern 1001 represents the sampling scheme represented by the ETDRS standard chart, which presents fourteen standard sizes of targets with a sampling resolution of .10 log units between consecutive sizes, ranging from -.3 to 1.0 logMAR. Anchor point 1002 of .35 represents the average target size of the ETDRS standard chart. In some embodiments, changing the anchor point 1002 of the standard constraint chart is used to generate a larger complementary set of sensitivity charts, which improves the overall resolution used for sampling target sizes while maintaining compliance with the sensitivity chart standard. In this particular embodiment, a set of ten sensitivity charts is generated by changing the anchor point of the standard chart in the range of .3-.4 logMAR, resulting in 1003 with an anchor point resolution of .01 logMAR. In some embodiments, each test slide uses the same sampling resolution as the ETDRS standard. However, when the set of charts constrained by design is considered collectively, these charts provide a finer resolution for the target size represented by the bottommost pattern, which is ten times finer than the ETDRS standard resolution. In this implementation, a set of 10 charts was developed to evaluate visual acuity with a resolution of 0.01 log units across a wide range of 141 different optotype sizes.

[0099] In some implementations, for standard ETDRS charts that follow a log-linear, row-by-row progression design principle for target size, the "anchor" is the target size corresponding to the average (or median) logMAR size of the fourteen target sizes encompassing the entire chart. In the case of ETDRS charts, the anchor corresponds to the target size between rows 20 / 40 and 20 / 50 (using Snellen notation).

[0100] Reference Figures 11A-11F In some implementations, exemplary subsets of visual acuity charts for design constraints are shown. In these implementations, the ETDRS design principle with three different anchor points (i.e., .30logMAR, .35logMAR, and .40logMAR) is used. Figures 11A-11C ) and the Snellen design principle with three different anchor points (i.e., .30logMAR, .35logMAR, and .40logMAR). Figures 11D-11F Generate a sensitivity chart. In these implementations, each row of the sensitivity chart may optionally have a fixed number or a variable number of targets.

[0101] Reference Figures 12A-12C In some specific implementations, test slides are focused on specific areas of an adaptive test chart with full standard constraints. The full design-constrained visual acuity chart is subsampled by presenting only the constraint set of the targets in the full visual acuity chart, which in turn is a subsample of a larger library of design-constrained charts. Rather than presenting as... Figure 2A The entire 14 rows of the ETDRS chart shown are less valuable than presenting a sub-sample of the visual acuity tests from a single sub-set (e.g., 3 rows of optotypes) that is presented on the full chart. This is achieved by presenting only a subsampled region of the full chart. Figure 12A ), presenting sub-sampling regions with higher contrast ( Figure 12B ), relative to the low contrast of the rest of the chart, or only presenting unblurred subsampled areas ( Figure 12C And blur other parts of the image to distinguish the target letters in the concentrated visual acuity test.

[0102] In existing technologies, sensitivity chart testing analysis is based on heuristics (row assignment or letter-by-letter), which produces inaccurate results and is difficult to reconcile across different charts.

[0103] In some implementations, the systems or methods disclosed herein provide a general statistical and computational framework for accurate testing of visual acuity and its variations using charts of any design, such as ETDRS, Snellen, or others.

[0104] In some implementations of visual acuity testing, the subject responds after the test slides are presented to them. In some implementations, the response from the subject may include the identification of one or more visual targets and / or the identification of one or more features of the visual targets. See also... Figures 13A-13C In some implementations, the subject's response is read down from the first column until instructed to stop. Figure 13A In some implementations, the subject reads at least one line in response until instructed to stop. Figure 13B In some implementations, the subject reads a single letter presented (). Figure 13C In some implementations, the target is solid black or gray. In some implementations, a gray target indicates a lower priority than a solid black target.

[0105] In some implementations, to improve the accuracy of visual acuity testing, a library of visual acuity charts has been created that displays high precision for sampled target sizes. Individually, these charts have the same sampling precision for sampled target sizes as existing acuity charts, and their appearance conforms to the design principles recommended by visual acuity chart standards committees (e.g., ANSI, ISO, NSC / NAS). However, compared to existing acuity charts, this expanded library generally displays a finer resolution for sampling target sizes (see [link to relevant documentation]). Figure 2A and Figure 15This library or similar libraries may contain a wide range of design features for viewpoint slides, including but not limited to: (1) slides with many rows that match the full ETDRS and Snellen charts; (2) slides that subsample the entire ETDRS chart and present only 1-5 rows at a time; and (3) slides with a single viewpoint. Similarly, the library may also include high-precision families of charts inspired by the Snellen design.

[0106] In some implementations, targets are randomly sampled from a target library. In some implementations, targets are adaptively sampled from a target library. In some implementations, adaptive sampling of targets is based on prior information about the subject's visual acuity. In some implementations, adaptive sampling is based on estimated visual sensitivity parameters from previous test runs, namely visual acuity threshold and / or visual acuity range and / or visual acuity slope. In other implementations, adaptive sampling is based on the results of previous tests, the patient's medical history, and the subject's response in previous tests or test runs. In some implementations, adaptive sampling optimizes target size for rapid and accurate estimation of sensitivity parameters. For example, if, on a first test chart, subject A correctly identifies all three rows with five targets per row, and subject B correctly identifies three rows with five, four, and four targets per row, respectively, then on a second test chart, targets of average size smaller than those presented to subject B are presented to subject A.

[0107] In some implementations, the visual acuity chart and / or subcharts provide a resolution that is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 times higher than the standard resolution of existing acuity charts (e.g., the standard ETDRS resolution).

[0108] In some implementations, the Snellen chart remains the primary sensitivity test in clinical practice, while the ETDRS chart serves as the standard for ophthalmic spatial clinical research and clinical trials. The ETDRS chart can present optotypes in fourteen rows (see [link to ETDRS chart]). Figure 2AThe minimum optotype size is logMAR = -3, the maximum optotype size is 1.0, and intermediate sizes are separated by a constant logarithmic linear series (e.g., .10 log units). These design principles can be specified by a standards committee. One drawback of these existing chart designs is the inflexibility of static, predetermined, and non-adaptive testing. Testing all patients in the same way—using a limited number of optotype sizes available on a single chart—can make it difficult to accurately estimate the broad visual acuity of a population (from healthy to visually impaired). To improve the flexibility and accuracy of visual acuity testing, and to better adapt testing to different levels of healthy and visually impaired individuals, current technology has moved towards computerized sensitivity testing using digital displays. Currently, adaptive testing algorithms focus on changing the size of a single-letter optotype based on correct and incorrect responses. These existing algorithms use simple heuristics limited to single-letter results rather than more complex multi-letter response results.

[0109] The design characteristics of different sensitivity tests can be specified by external or internal criteria or design principles. External criteria may reflect recommendations issued by ANSI or ISO committees, while internal criteria may reflect company or cultural preferences. External criteria can be considered directives, and internal criteria preferences. Both types of criteria limit the emergence and implementation of sensitivity tests. Currently, single-letter sensitivity tests do not conform to the design principles and criteria recommended by ANSI and ISO committees. As an example, consider a standard ETDRS chart with an anchor of 0.35 logMAR, and consider a set of sensitivity charts generated from ten other anchors (ranging from .3 to .4 logMAR and having a .01 log unit resolution). (The resolution of anchor sampling may become increasingly finer, but the advantage of finer resolution will ultimately be limited by the print resolution of paper charts or the pixel resolution of digital displays). These 10 charts (in addition to the ETDRS standard) are generated, each containing 14 optotype sizes, providing precise coverage of 141 different optotype sizes ranging from -0.4 to 1.1 logMAR. Since the chart design standard is a logarithmic series of 0.1 log units between rows on the same chart, the complete set of optotype sizes represented by these complementary charts can represent a subset of optotype sizes linearly interpolated in logarithmic space.

[0110] In some implementations, sensitivity charts that conform to the design constraints of ETDRS charts but have different anchor points are developed, resulting in a library of novel charts with larger viewpoint sizes compared to standard ETDRS charts (see [link]). Figure 2AA novel chart, defined as conforming to ETDRS chart design standards, may look largely the same as the original chart, but sampling for different viewpoint sizes can produce improved range and resolution. In some implementations, while maintaining standard constraint dimensional progression between adjacent rows, different anchor points result in slightly different maximum and minimum sizes. Figures 11A-11C Similarly, a novel library of charts with larger viewpoint sizes relative to standard Snellen charts can be generated.

[0111] Using sets of anchors (e.g., anchors not constrained by existing visual acuity charts) can generate alternative sets of standard constrained charts that provide more flexible sampling resolution for optotype sizes. Subcharts with fewer rows can further extend the complete 14-row visual acuity chart library. These visual acuity chart libraries can be generated via chart subsampling (e.g., displaying only 1-5 rows at a time). Spatial subsampling of the full chart can focus on a subset of the chart, e.g., rows 1-5 of optotypes, rather than all 14 chart rows. In some implementations, the top and / or bottom of the chart represent relatively uninformative areas that can be avoided during testing. Testing on these rows may result in correct reports of all optotypes or no optotypes. Better sampling of the full chart may involve optotype sizes with different possible intermediate results (e.g., 2, 3, or 4 optotypes out of 5 correct results). Complementary chart sets can represent stimulus sets that can be sampled once or repeatedly via deterministic, random, and / or optimized sampling. Sampling of standard ETDRS visual acuity charts in... Figures 11A-11C The order of magnitudes between rows is constant in logarithmic space. The anchor points for these magnitudes can be chosen as the average value in logarithmic space. In this implementation, seven sizes are greater than the selected anchor points and seven sizes are less than the selected anchor points.

[0112] In some implementations, the chart libraries disclosed herein exhibit a design feature that provides combinations of viewpoint sizes that significantly expand upon the finite set of fourteen sizes used by the ETDRS standard charts. Different libraries may be generated specifically for each chart and its design principles (e.g., ETDRS, Snellen, HOTV, Lea Symbol, etc.).

[0113] Despite emerging use of ETDRS in clinical research and trials, the Snellen chart has remained the dominant sensitivity test in clinical practice for 150 years. For practical purposes, the methods used to calculate sensitivity indices rely on psychological scoring heuristics, which are explicitly related to letters and design principles easily understood by the testing practitioner. Two main methods for estimating sensitivity are row-by-row allocation or letter-by-letter allocation. For example, in the Snellen sensitivity test, the row-by-row allocation method calculates the final sensitivity estimate as the optotype size corresponding to the last row of three out of five optotypes correctly reported by the patient. For comparison with row-by-row allocation, the ETDRS chart uses letter-by-letter scoring, where each correctly reported letter is recorded in .02 logMAR units.

[0114] Unlike existing techniques that only consider a single psychometric function that defines the probability of correctly identifying a single optotype according to its size (sometimes referred to as the "observation frequency curve"), the method and system in this paper involve analyzing the response to estimate sensitivity by considering a family of psychometric functions that define the perfectly complementary probabilities of correct combinations of responses to observing multiple optotypes according to optotype size (see Figure 4-). Figure 9 In the example of an ETDRS chart presenting five optotypes per row, there are six complementary psychometric functions describing the probability of correctly reporting 0, 1, 2, 3, 4, or 5 optotypes out of the total five optotypes presented. Figure 6B , Figure 8B Using existing techniques, measuring six psychometric functions only doubles the problem of measuring a single psychometric function. Practical limitations in clinical testing time also make it impossible to estimate sensitivity using existing techniques with these six empirically observed chart-specific psychometric functions.

[0115] Sensitivity parameters

[0116] Visual acuity can be the sharpness, clarity, or resolution of vision. In some embodiments, visual acuity is represented by at least one psychological test function. In some embodiments, visual acuity is represented by at least one implicit or explicit psychological test function. In some embodiments, the psychological test function is chart-specific or chart-invariant. In some embodiments, visual acuity is represented by at least one sensitivity (d') psychological measurement function. In signal detection theory, the sensitivity (d') parameter may represent behavioral performance independent of task-based factors influencing response bias or decision criteria. In some embodiments, the sensitivity parameter in a sensitivity test determines visual performance independent of specific design features used for the visual acuity test. In some embodiments, the psychological test function is uniquely defined by at least two sensitivity parameters. In some embodiments, the two visual acuity parameters form a two-dimensional parameter space. In some embodiments, visual acuity is represented by at least one single-optotype psychological test function and / or a multi-optotype psychological test function.

[0117] In some embodiments, at least one sensitivity parameter is used to determine or measure visual acuity. The sensitivity parameters disclosed herein are interchangeable with sensitivity model parameters or sensitivity indices. In some embodiments, the sensitivity parameter is a sensitivity threshold. In some embodiments, the sensitivity parameter is a sensitivity range. In some embodiments, the sensitivity parameter is a sensitivity slope. In some embodiments, two sensitivity parameters are used to determine or measure visual acuity. In some embodiments, the two sensitivity parameters are a sensitivity threshold and a sensitivity range. In some embodiments, the two sensitivity parameters are a sensitivity threshold and a sensitivity slope.

[0118] In some embodiments, the sensitivity parameters described herein include a probability density function, cumulative probability density function, or probability distribution function of a sensitivity threshold, sensitivity range, sensitivity slope, sensitivity threshold, sensitivity range, sensitivity slope, or a combination thereof. In some embodiments, the sensitivity parameters described herein include a probability density function, cumulative probability density function, or probability distribution function of the change in sensitivity threshold, sensitivity range, sensitivity slope, sensitivity threshold, sensitivity range, sensitivity slope, or a combination thereof between two different test conditions. In some embodiments, the probability distribution function and probability density function described herein are interchangeable.

[0119] For example, a sensitivity threshold may include a probability density function (pdf) for different target sizes, which includes the probability that only the subject can detect different target sizes. In some implementations, the sensitivity threshold includes a range of different target sizes that include only the different target sizes that can be detected by the subject. In some implementations, the sensitivity threshold includes a target size value and a size variation, where the target size is the smallest size that the subject can detect.

[0120] In some implementations, the sensitivity parameters herein include a sensitivity (d') threshold, which is configured to provide a visual performance standard similar to the sensitivity threshold. The sensitivity threshold determined by a chart-invariant psychological testing function can be a sensitivity (d')-based threshold. In some implementations, the sensitivity threshold can be defined within a chart-invariant region represented by a chart-invariant psychological testing function (e.g., a sensitivity (d')-based psychological testing function). Alternatively, the sensitivity threshold can be defined by a chart-specific psychological testing function, such as an empirical psychological testing function that depends on factors related to the design of the sensitivity chart. A directional shift may exist between sensitivity thresholds defined by chart-specific psychological functions and chart-invariant psychological functions.

[0121] Reference Figures 4A-4B and / or Figures 7A-7B In certain embodiments, thresholds 401 and 701 are shown using different coordinates, and sensitivity thresholds 401 and 701 can be expressed in different units. For example... Figure 7A As shown, in some embodiments, the sensitivity threshold 701 represents the optotype size corresponding to a pre-specified visual performance standard: for example, the size corresponding to the probability of correctly identifying an optotype of a certain size, or the size corresponding to 3 out of 5 optotypes correctly reported in a row of an ETDRS chart. In the same embodiment, a second key metric used to describe the performance of the sensitivity chart is the sensitivity slope 402 ( Figure 4B ) or sensitivity slope 702 ( Figure 7B This describes the rate at which a chart-specific psychometric function changes around thresholds 401 and 701. In some implementations, the slope represents the rate at which sensitivity changes with target size. In some implementations, the sensitivity or sensitivity slopes 402 and 702 are inversely proportional to the sensitivity range, which represents the width of the dynamic range of the sensitivity function. In this case, the sensitivity or sensitivity ranges 403 and 703 determine the logarithmic difference between two target sizes corresponding to the expected performance levels of correctly identified 1.5 and 4.5 letters. .40 logMAR Target Size Difference ( Figure 4BThe sensitivity range in the chart corresponds to the sensitivity range of four rows on the ETDRS chart. In this case, within the four rows on the ETDRS chart, a subject reading the ETDRS chart will range from their best average performance (correctly identifying 5 out of 5 in the higher rows) to their lowest average performance (correctly identifying 1 or 2 out of 5 in the lower rows). Algorithm S (scoring algorithm) can produce visual acuity thresholds and range estimates as a key assessment of visual acuity. In some implementations, the sensitivity threshold can provide a primary indicator or parameter, but identifying changes in sensitivity range / slope can provide an auxiliary assessment or parameter when characterizing visual acuity loss in an eye disease. The function of the sensitivity model can include reducing the rich and complex data collected through visual acuity testing to a simpler visual acuity prediction with two parameters.

[0122] In some embodiments, the sensitivity parameters described herein include a sensitivity threshold, sensitivity range, sensitivity slope, sensitivity threshold, sensitivity range, sensitivity slope, or a combination thereof under one or more test conditions. In some embodiments, the sensitivity parameters described herein include a sensitivity threshold, sensitivity range, sensitivity slope, sensitivity threshold, sensitivity range, sensitivity slope, or a combination thereof between two test conditions. In some embodiments, the sensitivity parameters described herein are numerical values ​​expressed in units reflecting target size, such as 20 / 20, 20 / 40, 3 / 5, 1 / 5, or 0.3 logMAR. In some embodiments, the sensitivity parameters described herein correspond to a predetermined level of visual performance, which can be defined using graph-invariant terminology (sensitivity) or graph-specific terminology, thereby generating sensitivity-based sensitivity parameters (e.g., sensitivity threshold and sensitivity range) and sensitivity parameters (e.g., sensitivity threshold and sensitivity range), respectively. (See also...) Figures 4A-4B In some embodiments, the chart-invariant sensitivity threshold can be a target size corresponding to a predetermined sensitivity level d' = 1 or d' = 2, for example, 0.3 logMAR. In the same embodiment, the chart-invariant sensitivity range can be the difference between two target sizes corresponding to low sensitivity (e.g., -0.3 logMAR at d' = 0.5) and high sensitivity (e.g., 1 logMAR at d' = 4). See also... Figures 7A-7B As shown, it presents a chart-specific psychometric function, where the chart-specific sensitivity threshold can be the optotype size corresponding to the number of correctly reported pre-specified letters (expected to be 5), such as 3 out of 5, 3 / 5, 2, 4, or 3.5 out of 5, for example, 0.3 logMAR. In the same embodiment, the sensitivity range can be the difference between a smaller number of correctly reported letters (e.g., -0.3 logMAR corresponds to 1 out of 5) and a larger number of correctly reported letters (e.g., 0.9 logMAR corresponds to 4 out of 5).

[0123] In some implementations, the systems and methods described herein are used to determine, for example, the change in sensitivity between two different test conditions, e.g., via analysis of a Bayesian posterior probability density function (pdf). For the case of contrastive sensitivity, a Bayesian pdf of the sensitivity difference between the two test conditions can be generated from a single pdf of the test conditions (Hou et al. 2016, which are incorporated herein by reference in their entirety):

[0124]

[0125] Where 'a' represents the sensitivity measured under the first test condition, and Δa represents the sensitivity difference between the two conditions; p 差异 p1(.) is the probability function of sensitivity difference, and p2(.) and p1(.) are the Bayesian PDFs defined for the sensitivity parameter under two conditions.

[0126] Reference Figures 27A-27B In a specific implementation, Bayesian posterior probability density functions for sensitivity thresholds under different test conditions 1, 2, 3, and 4 are generated. In this implementation, the probability of sensitivity change between two test conditions can be calculated by the relative overlap of the Bayesian posterior probability density functions measured under different test conditions. Figure 27A As shown, the low variability of the sensitivity threshold estimate (e.g., the width of the probability density function) of the Bayesian posterior probability density function provides confidence in the change in visual sensitivity between test 1 and test 2, which is reflected in the minimum overlap of the sensitivity threshold posteriors collected under different conditions 2710a. (Refer to...) Figure 27B In some implementations, the average difference in sensitivity between test conditions 2810 is the same, but there is a wider range of variability represented by the Bayesian posterior (pdf), increased deviation, and increased overlap 2710b, such as the width of the probability density function in the sensitivity threshold estimation.

[0127] The calculation of sensitivity changes or differences via Bayesian signal detection analysis may include, for example, generating a difference distribution among sensitivity values ​​obtained under different conditions 2810 before or after medical intervention (e.g., ...). Figures 28A-28B The Bayesian PDF in Test 1 (shown in 2811) – the Bayesian PDF in Test 4. After collecting the posterior under each test condition, the generation of the difference distribution 2811 provides, as shown in [example 2811]. Figure 28B The equivalent index of the sensitivity threshold change shown.

[0128] In some implementations, the generation of the difference distribution between the Bayesian posteriors of two test conditions, such as test 1 and test 2, provides an estimate that the change between the conditions is greater than a threshold standard level of sensitivity change. For example, the threshold standard level could be greater than one row, two rows, or three rows of a given sensitivity chart.

[0129] In some implementations, in addition to testing high contrast visual acuity with black targets on a white background, low contrast visual acuity can also be tested with grayscale targets that show a smaller brightness difference from the grayscale background. For example, a design using a Sloane low-contrast chart can be employed. In some implementations, the visual acuity chart used to test low contrast visual acuity may appear similar to an ETDRS, but different charts calculated relative to a standard background for the visual acuity chart can present targets with contrast ratios of approximately 25.0%, 2.5%, and 1.5%. In some implementations, the low-contrast visual acuity chart has one or more contrast ratios of approximately no more than 60%, 50%, 40%, 30%, 25%, 24%, 20%, or even lower than the contrast ratios of standard ETDRS and / or Sloane sensitivity charts. In some implementations, the low-contrast visual acuity chart has one or more contrasts that are about 25%, 20%, 18%, 15%, 12%, 10%, 8%, 5%, 2.5%, 2%, 1.5% or even lower than the contrast of the standard ETDRS and / or Sloane acuity chart.

[0130] In some implementations, when using a more complex sensitivity model, such as one with two sensitivity thresholds and two sensitivity ranges, the difference distribution can be estimated directly instead.

[0131] In some implementations, the sensitivity model change reflecting sensitivity variation or difference can be directly determined using algorithms S and A. In some implementations, a two-dimensional sensitivity model, determined using the two sensitivity parameters described herein, can be refined into a model with higher dimensions, such as three, four, five, six, or even more dimensions. For example, a two-dimensional sensitivity model can be used to extend to a four-dimensional model in two ways: (a) by determining the sensitivity threshold and sensitivity range under two test conditions; then algorithm A can select which stimulus is the optimal stimulus to be tested on a trial-by-trial basis and which condition is the optimal condition to be tested on a trial-by-trial basis; (b) by determining the sensitivity threshold and range under one condition, and the difference between the conditions within the threshold and range as a second set of parameters, namely Δthreshold and Δrange. For example, the sensitivity model can assume no difference in the expected sensitivity range between the two conditions. In other words, the sensitivity difference between the two test conditions can be characterized as a lateral shift in a graph-specific sensitivity function while the sensitivity range parameter remains constant.

[0132] In some implementations, the results from sensitivity model analysis are evaluated to determine the sensitivity threshold, sensitivity slope, and / or the probability of changes in the sensitivity range.

[0133] In some embodiments, the sensitivity model described herein is configured to directly estimate sensitivity changes. Such sensitivity changes may occur between two test conditions or within two or more test conditions. In some embodiments, the different test conditions may include any differences in the subject. For example, before or after a medical intervention. As another example, testing may be performed on a diseased eye and a normal eye, or on the left eye and the right eye. In some embodiments, the sensitivity model described herein may estimate a sensitivity threshold, a sensitivity range, and changes in the sensitivity threshold and sensitivity range. In some embodiments, the sensitivity parameters include a sensitivity threshold and sensitivity range from a first test condition and a second condition different from the first test condition, and additional parameters that determine the differences between these parameters (e.g., a Δ sensitivity threshold between the first and second test conditions). In some embodiments, the first condition is the left eye and the second condition is the right eye. In some embodiments, the first condition is before treatment and the second condition is after treatment. In some embodiments, the first condition is measured at low brightness and the second condition is measured at high brightness. In some embodiments, the first condition is measured at low contrast and the second condition is measured at high contrast.

[0134] Summary of sensitivity chart data

[0135] In some embodiments, the systems and methods described herein include sensitivity chart data obtained by collecting responses from test subjects to presented sensitivity charts or subcharts. Sensitivity chart data is interchangeable herein as sensitivity test data and / or responses. In some embodiments, the responses of test subjects may be summarized in, for example, a data table. For a single optotype, the response r may be the correctness of a single response to that optotype. For rows of multiple optotypes, the combined response r may summarize how many optotypes m were correctly identified out of all n optotypes presented in each row. Because each row can correctly report 0 to n (n in total) optotypes, there can be n+1 potential outcomes for the combined response.

[0136] like Figure 1 As shown in operations 100 and 400, in some cases, sensitivity chart design characteristics can be determined and chart design parameters can be summarized. In some implementations, the data tables in this document (e.g., Table 1) collectively define the parameters defining the chart design as well as the parameters of the subjects' responses that are important for the application of the sensitivity model.

[0137] Table 1 shows exemplary summary sensitivity chart data collected from test subjects.

[0138] Table 1.

[0139]

[0140] In Table 1, each row can define a single sensitivity chart condition, including the size of a single optotype or a row of optotypes, and the corresponding single or compound response of the presented optotype. In the specific case shown in Table 1, sensitivity chart data collected from test subjects is summarized. This summary helps identify design features that contribute to visual performance observed during sensitivity chart testing, in order to apply Algorithm S to scoring the sensitivity chart data. In this summary table, each row entry represents reading a single optotype or a row of optotypes from the visual acuity chart or subchart. Each row of optotypes can be considered a combined stimulus, which, in addition to the key feature of optotype size, can also be described by: the size of the subset of optotypes sampled on each row, the size of the superset of optotypes sampled from it, and whether sampling from the superset was done with or without substitution. For each chart design parameter in the summary table, the response summary represents the number of correctly identified optotypes, whether for a single optotype or a row of optotypes. In some implementations, the viewpoint parameters, as shown in Table 1, may significantly affect the chart-specific psychometric function, and therefore are required to generate the chart-specific psychometric function.

[0141] In some implementations, sensitivity test data may include test data from 1, 2, 3, 4, 5 or even more test conditions.

[0142] Candidate parameters for sensitivity model

[0143] In some implementations, the systems and methods described herein can be used to generate one or more sets of candidate parameters. A set of candidate sensitivity model parameters may include a combination of two or more sensitivity parameters, such as sensitivity thresholds and sensitivity ranges. Various combinations of sensitivity parameters can form multiple sets of candidate parameters. Such sets of candidate parameters can be configured to describe summarized sensitivity chart data, as shown in Table 2 below. Figure 1 In operation 300, candidate parameters such as those shown in Table 2 can be determined at least in part by sensitivity chart data or sensitivity test data collected from test subjects in operation 400. Alternatively or in combination, such candidate parameters can be determined at least in part by the sensitivity chart design parameters determined in operation 100. In some embodiments, operation 300 can be independent of the design parameters in operation 100. In some embodiments, operation 300 occurs prior to operation 200. And each set of candidate parameters from operation 300 is used in operation 200 when determining the general sensitivity model, for example... Figure 3A and / or Figure 3B In some implementations, operation 300 may be performed in parallel or serially with operation 400 and / or operation 100.

[0144] Table 2 shows the candidate sets of sensitivity model parameters. These sets of sensitivity threshold and sensitivity range parameters include sensitivity model parameters that can potentially describe the sensitivity graph data collected from test subjects. Each entry row reflects a single combination (total of q1) of candidate sensitivity threshold parameters and q2 candidate sensitivity range / slope parameters. In this implementation, the table has a combination number of q1 multiplied by q2.

[0145] Table 2.

[0146]

[0147] In some implementations, the observed sensitivity chart data is analyzed to infer the sensitivity model parameters that best predict sensitivity from the test subjects. In some cases, the set of candidate sensitivity model parameters is a list of parameter combinations for sensitivity thresholds and sensitivity ranges. Statistical inference can describe the probability of each parameter combination describing the observer. This is equivalent to estimating the probability of class membership. Alternatively, the set of candidate parameters comprises a two-dimensional space of sensitivity parameters—sensitivity thresholds and sensitivity ranges—providing exhaustive parameter combinations to describe the subject's sensitivity performance. In this method, marginal probabilities can be calculated through both dimensions of the sensitivity parameters.

[0148] The set of candidate sensitivity parameters can be implemented in a sensitivity model used for visual acuity analysis. For example, the set of candidate sensitivity parameters can be a one-dimensional array of sensitivity threshold-range pairs. As another example, the set of candidate sensitivity parameters can be two-dimensional and can be used to generate a sensitivity model that comprehensively describes the combination of sensitivity thresholds and sensitivity ranges. As yet another example, the set of candidate sensitivity parameters can be four-dimensional: the sensitivity thresholds and ranges under baseline conditions (e.g., left eye or pre-treatment conditions) and the variations in sensitivity thresholds and ranges relative to those conditions.

[0149] In some implementations, each set of sensitivity parameters may include 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 or any other integer number of sensitivity parameters.

[0150] In some implementations, the candidate sensitivity parameters can have 1, 2, 3, 4, 5…1000…10000 or any other integer set.

[0151] In some implementations, each set of candidate sensitivity parameters includes the same number of parameters. In some implementations, each set of candidate sensitivity parameters includes one or more of the following: sensitivity threshold, sensitivity range, sensitivity slope under one or more test conditions, and changes in sensitivity threshold, range, or slope between two test conditions.

[0152] In some embodiments, each set of candidate sensitivity parameters includes one or more of the following: a sensitivity threshold, sensitivity range, or sensitivity slope under one or more test conditions, and a change in the sensitivity threshold, range, or slope between two test conditions. In some embodiments, candidate sensitivity parameters herein include a sensitivity threshold, sensitivity range, sensitivity slope, sensitivity threshold, sensitivity range, sensitivity slope, or any combination thereof under one or more test conditions. In some embodiments, candidate sensitivity parameters herein include a change in the sensitivity threshold, sensitivity range, sensitivity slope, sensitivity threshold, sensitivity range, sensitivity slope, or any combination thereof between two test conditions. In some embodiments, candidate sensitivity parameters herein are numerical values ​​in units reflecting target size, such as 20 / 20, 20 / 40, 3 / 5, 1 / 5, or 0.3logMA. In some implementations, the candidate sensitivity parameters described herein correspond to predetermined levels of visual performance and can be defined in chart-invariant terms (sensitivity) or chart-specific terms, thereby generating sensitivity-based sensitivity parameters (e.g., sensitivity threshold and range) and sensitivity parameters (e.g., sensitivity threshold and range), respectively. In some implementations, candidate sensitivity parameters defined in chart-invariant terms (sensitivity) or chart-specific terms can be converted to each other. In other words, a set of chart-invariant sensitivity parameters can be converted to a set of chart-specific sensitivity parameters using the same target information, for example, target size is related to sensitivity performance, such as 0.3 logMAR, but the sensitivity or sensitivity terminology is, for example, at d' = 2 / 5 or 3 / 5, respectively. In some implementations, candidate sensitivity parameters are interchangeable with candidate sensitivity model parameters or candidate parameters.

[0153] Sensitivity Model

[0154] In some embodiments, the methods, systems, or apparatuses disclosed herein include one or more sensitivity models (which are interchangeable herein as "statistical models" or "predictive models"). Such terms are equivalent and interchangeable herein. In some embodiments, the sensitivity model is used to predict sensitivity chart data of a subject (e.g., the probability of observing a particular response from the subject), the sensitivity chart data indicating the subject's visual acuity performance. In some embodiments, such predictions using a sensitivity model are based on the design features of the sensitivity chart used to collect sensitivity chart data from test subjects. In some embodiments, such predictions using a sensitivity model are invariant to specific sensitivity chart design features.

[0155] Reference Figure 1In certain implementations, operations 200, 300, and / or 350 may be used to determine the sensitivity model. In some implementations, the sensitivity model herein includes model parameters. Such model parameters may include: (a) fixed sensitivity chart design parameters, at least partially determined by the sensitivity chart design, which are at least partially determined by operation 100; and (b) a set of candidate sensitivity parameters describing the patient's visual acuity, which may be at least partially determined in operation 300. In operation 300, a set of candidate sensitivity parameters to be used to populate the sensitivity model is estimated. In operation 200, each set of the estimated set of parameters in operation 300 may be used to generate operation 200 and... Figure 3A and / or Figure 3B The chart-specific functions, in some cases, are combined with sensitivity chart design parameters. Then, in operation 350, the chart-specific functions are used to populate the sensitivity model to be used in operation 500. In some implementations, the functions of the sensitivity model are used to output predictions of sensitivity chart data or test data (e.g., response probabilities) using the sensitivity chart design parameters (in operation 100), candidate sensitivity parameters (in operation 300), and collected sensitivity chart data (in operation 400). In some implementations, operation 350 may use at least a portion of a plurality of candidate parameters and their corresponding chart-specific psychometric functions to generate predictions of sensitivity chart data or test data (e.g., response probabilities). In some implementations, operation 350 may use all of a plurality of candidate parameters and their corresponding chart-specific psychometric functions to generate predictions of sensitivity chart data or test data (e.g., response probabilities).

[0156] A sensitivity model may include one or more sensitivity parameters. Non-limiting examples of sensitivity parameters include: sensitivity threshold, sensitivity slope, sensitivity range, sensitivity threshold, sensitivity range, variation of the sensitivity threshold between two test conditions, and variation of the sensitivity range between two test conditions. Given a sensitivity model, the sensitivity model parameters can be used to predict observed visual performance during tests designed with any sensitivity chart. These model parameters can provide valuable sensitivity metrics to characterize the visual acuity of the test subjects.

[0157] In some implementations, the sensitivity model includes a central component—a chart-specific psychometric function configured to allow determination of the expected number of correctly identified targets on a row of a sensitivity chart, based on the size of the targets on that row.

[0158] Figures 7A-7B An exemplary implementation of a chart-specific psychometric function 706 is shown, which represents the expected number of correctly identified targets in each row of targets, based on the number of targets presented in that row. Algorithm S can apply a sensitivity model to sensitivity chart data to estimate the characteristics of the chart-specific psychometric function.

[0159] The system and method disclosed herein can estimate chart-specific psychometric functions 606, 706, chart-invariant functions 404, or chart-specific single-optograph psychometric functions 604, 804, and / or multi-optograph psychometric functions 605, 805, and / or their sensitivity parameters, such as chart-specific sensitivity thresholds 701, sensitivity ranges 703, and slopes 702, or chart-invariant sensitivity thresholds 401, sensitivity ranges 403, and sensitivity slopes 402, to assess visual acuity.

[0160] For example, chart-specific psychometric functions 604 and 804 may depend primarily on the specific design features of the sensitivity chart used to collect the chart-specific sensitivity data. Chart-specific psychometric functions can also predict sensitivity chart data collected on many different sensitivity charts by incorporating features of the sensitivity chart design. These features include, but are not limited to:

[0161] The number of rows displayed in the chart or subchart;

[0162] The number of targets displayed in each row;

[0163] The number of views that can be sampled in each row;

[0164] The number of viewpoint sizes that can be sampled per row;

[0165] Sampling of the target pool is performed with or without a substitute, or a combination thereof.

[0166] As a non-limiting example, for a standard logMAR sensitivity chart (ETDRS or Bailey-Lovie) that presents five optotypes in each row, the chart-specific psychometric function will describe the number of optotypes expected to be correctly identified in each row, based on the optotype size.

[0167] As another example, for logMAR sensitivity charts, two important sensitivity metrics summarizing the chart-specific psychometric function are: the chart-specific sensitivity threshold—the standard number of optotype sizes (in logMAR) at which a subject can correctly identify each row of optotypes—and the chart-specific sensitivity range—this parameter relates to the width of the dynamic range of the psychometric function and is inversely related to the slope of the function. In some implementations, the sensitivity range describes the logMAR difference between two optotype sizes: one above and one below the sensitivity threshold. In some implementations, the range parameter describes how quickly the sensitivity chart-specific psychometric function revolves around the sensitivity threshold. A smaller range indicates that the observer's vision transitions rapidly from incorrect optotype recognition to correct optotype discrimination as the optotype size increases. In some implementations, the sensitivity range is a valuable metric because a hallmark of eye disease is increased variability in visual acuity, corresponding to an increase or change in the sensitivity range.

[0168] For example, for a logMAR sensitivity chart, the sensitivity threshold is the size of the correctly identified optotypes that are 3 out of 5 optotypes in each row, and the sensitivity range corresponds to the distance between the chart lines when the expected number of correctly identified optotypes in each row is 1.5 and 4.5.

[0169] In some implementations, other features of the psychometric function are defined by the design characteristics of the sensitivity chart. These features may also include the lower asymptote—defined by the ratio of the total number of possible targets guessed, and the upper asymptote—defined by the number of targets presented in each row.

[0170] In some implementations, in order to construct a chart-specific psychometric function 706, 606 that describes the expected number of correctly identified targets per row based on target size, by... Figure 3B Operation 220 or Figure 3A Operations 230 and 330 in the diagram generate a chart-specific single-viewpoint psychometric function. This can be based on sensitivity parameters (e.g., sensitivity chart design parameters or features) that are independent of the sensitivity chart and are based on task-based features. Figures 4A-4B The chart-invariant single optotype psychometric function 404, defined in the document, generates a chart-specific single optotype psychometric function. In some implementations, the single optotype psychometric function shows the probability of correctly identifying a single optotype based on optotype size, and in some cases, uses a cumulative distribution function, such as a Gaussian cumulative probability distribution function, to determine this probability using two sensitivity model parameters—for example, a sensitivity threshold and a sensitivity range. Then, based on the principles of signal detection theory (Green & Swett, 1961; McMillan & Creelman, 2004; Lesmes et al., 2015, which are incorporated herein by reference in their entirety) and the characteristics of the optotype set (…),… Figure 5 This transforms a sensitivity-based chart-invariant single-opposite psychometric function into a single-opposite chart-specific psychometric function. For example, the transformed single-opposite chart-specific psychometric function depends on characteristics of the sensitivity chart design, such as the number of potential optotypes. Based on the number of optotype substitutions, the same sensitivity threshold (corresponding to an optotype with d' = 1) can represent different levels of correct performance for the signal optotype. As mentioned in signal detection theory, it is easier to guess correctly when there are fewer guessed substitutions. Therefore, for the same optotype size corresponding to a given sensitivity threshold (d' = 1), the number of guessed substitutions in the optotype set can lead to different levels of the percentage of correct performance used to identify a single optotype: at the sensitivity threshold d' = 1, the range is 70-80% for two substitutions and 30-40% for ten substitutions. Figure 5 ).

[0171] In some implementations, combinations of operations such as 300, 200, and 350 can generate a sensitivity model that scores the sensitivity chart data collected and summarized in operation 400. This visual acuity model may include, for example, a large number of predicted probabilities in a table for the observed sensitivity chart data, which may be predicted probabilities for different responses to optotype recordings presented to the patient for a given set of candidate sensitivity model parameters describing the patient's vision.

[0172] In one implementation, the two-parameter description of visual acuity can use a sigmoid function to define visual performance according to target size (e.g., see...). Figure 4A and Figure 6A One parameter, the sensitivity threshold, describes the lateral position of the sigmoid function on the horizontal axis of the target size. The other parameter, the sensitivity range, describes how quickly the sigmoid function changes on the horizontal axis of the target size near the threshold. A smaller range indicates a rapid change in the sensitivity function from low visual performance represented by the lower asymptote to high visual performance represented by the upper asymptote. Both the sensitivity threshold and sensitivity range parameters can be defined in units of logMAR (target size).

[0173] Reference Figure 7A In some implementations, sensitivity thresholds and sensitivity ranges are defined using terminology specific to the ETDRS sensitivity chart design. For a range of target sizes, each row of the ETDRS chart can display five letters. Figure 7A and Figure 7B In this context, the sensitivity function thus represents the upper asymptotes of the five letters that are correctly reported. Figure 7A In this study, the sensitivity threshold of 701 is defined as the target size at which the patient is expected to correctly report three out of five presented targets. Figure 7B In this context, sensitivity range 703 defines the range of visual behavior on an ETDRS chart that ranges from correctly reporting 1.5 letters out of 5 to correctly reporting 4.5 letters out of 5. In these implementations, because the sensitivity threshold and sensitivity range are defined by three performance levels—correctly reporting 1.5, 3.0, and 4.5 letters out of 5—which are defined by ETDRS features, the sensitivity threshold and range can be viewed as chart-specific sensitivity model parameters.

[0174] In some implementations... Figure 4A and Figure 4BThe sensitivity threshold 401 and sensitivity range 403 presented are defined in sensitivity units (d') independent of the sensitivity chart design. The sensitivity threshold can be defined as the target size (in logMAR) corresponding to a visual performance d' = 1. The sensitivity range can be defined as the difference (in logMAR) between the target sizes corresponding to a lower level of visual performance (d' = .25) and a higher level of visual performance (d' = 4). Because these sensitivity models are defined in units of sensitivity, the sensitivity threshold and range can be considered as parameters of the chart-invariant sensitivity model.

[0175] To understand how Operation 350 generates a predictive model for sensitivity chart data based on chart-invariant parameters, consider a pair of general sensitivity model parameters: a sensitivity threshold and a sensitivity range (e.g., this parameter pair includes the first row of Table 2). Based on this chosen pair of sensitivity model parameters, the probability of observing different responses during sensitivity chart testing can be obtained based on the optotype size. In addition to these sensitivity model parameters, consider one factor in sensitivity chart design: the range of optotype sizes that might be tested in the sensitivity chart. For example, the ETDRS chart uses optotype sizes in the range of -3 to 1.0 logMAR with a resolution of 10 logMAR.

[0176] Figure 4A An exemplary implementation of a graph-invariant psychometric function is shown, which can be generated by continuous sampling of optotype sizes similar to the ETDRS test. Signal detection theory describes how to transform task-independent sensitivity parameters into a measure of task-based behavior (e.g., the probability of correctly identifying an optotype). Using signal detection theory (Green & Swets, 1961; McMillan & Creelman, 2004; Jacobs & Fine, 2002; Lesmes et al., 2015, which are incorporated herein by reference in their entirety), the probability of correctly identifying a single optotype is derived from the d' value based on a Gaussian distribution assuming internal representations of the signal and noise for an optotype task with different numbers of guesses. For the equations describing this transformation, see Appendix A, Equation 7 of Jacobs & Fine, 2002, or Equation 1 of Hacker & Ratcliff, 1979. Alternatively, McMillan & Creelman (2004) presented Table A5.7 as a lookup table to convert between the corresponding sensitivity level (d') and probability of correctness for m-AFC substitution tasks in the range of m=2 to m=1000.

[0177] As an example, Figure 5This indicates that for a general stimulus corresponding to a sensitivity level of d'=1, the probability of correct identification is between 70% and 80% when only two guessed alternatives (2AFC) exist (in any given trial, one alternative is correct and the other is incorrect, independent of the observer's visual proficiency). As the number of guessed alternatives increases, the probability of an incorrect response also increases. Therefore, the probability of a correct response decreases with increasing number of guessed alternatives (see [link to relevant documentation]). Figure 5 (Relative stratification of function 504). Compared to 2AFC, for a stimulus corresponding to a sensitivity level of d'=1, the probability of correctly reporting the stimulus is approximately 30% when presented with 10 guessed alternatives.

[0178] Based on this transformation from sensitivity (d') to correct probability 330 ( Figure 5 ),generate Figure 4A The sensitivity model parameter pair of the S-shaped function can generate a monotonically increasing function that describes the probability of correctly reporting a target based on its size. From Figure 5 The transformation used may depend on the number of other optotypes (guessed alternatives) in the optotype sampling pool. In the case of ETDRS using a Sloan set of 10 optotypes, a psychometric function 604 for single optotype presentation can be generated using a 10AFC function. Figure 6A ).

[0179] Just as the number of optotypes in the sampling pool can be a sensitivity chart design factor (operation 330) used to convert chart-invariant psychometric functions into single-opposite psychometric functions, another sensitivity chart design factor, namely the number of optotypes presented in each row, can be used to generate a family of multi-opposite psychometric functions through serial multiplication. The weighted sum of this family of psychometric functions, in turn, provides the chart-specific psychometric function. Figure 6C .

[0180] In some implementations, given a single pair of sensitivity model parameters, such as a sensitivity threshold and a sensitivity range, six psychometric functions can be generated that describe the probability of correctly identifying 0, 1, 2, 3, 4, or 5 targets from 5 targets, based on the target size. In operation 500, given sensitivity chart data collected during the test, these probabilities can be used to evaluate which sensitivity model parameters best describe the test patient.

[0181] In some implementations, operation 300 may include multiple sets of candidate parameters that could be used for the sensitivity model. The purpose of the scoring algorithm S may be to evaluate these sets of candidate sensitivity model parameters and to evaluate which candidate parameters best describe the sensitivity chart data summarized in operation 40.

[0182] In some implementations, during operation 300, the set of candidate parameters can be defined by one or more of the following: a one-dimensional array / list of sensitivity model pairs, such as sensitivity threshold and sensitivity range; a one-dimensional array of sensitivity model pairs, including different values ​​of the sensitivity threshold and fixed values ​​of the sensitivity range; a two-dimensional space defined by an exhaustive combination of the values ​​of two parameters, such as the sensitivity threshold parameter and the sensitivity range parameter; or a high-dimensional space (higher than two dimensions) describing the sensitivity model parameters and their variations under different conditions for the sensitivity threshold and sensitivity range parameters.

[0183] For each of these sets of sensitivity model parameters, a sensitivity model probability can be generated based on the sensitivity threshold and range. In operation 500, these probabilities obtained using each set of these candidate sensitivity parameters can be used to discover sensitivity model parameters that describe the sensitivity graph data.

[0184] For example, the number of potential targets can vary depending on the chart design:

[0185] Sloan and Bailey-Lovie each have 10 visual targets;

[0186] Landolt C, Tumbling E, HOTV, Lea, and Hands are four;

[0187] The traditional Snellen number is 9;

[0188] British Snellen standard BS 4274-1:2003 specifies 12;

[0189] The English alphabet has 26 letters;

[0190] Then, in some implementations, the single-opposite psychometric function P(s) is used to calculate a family of multi-opposite psychometric functions P that is specific to the design of the sensitivity chart (e.g., the number of optotypes presented in each row determines the number of psychometric functions). m (s). Such a family of multi-opinion psychometric functions P m(s), m = 1, 2, 3, ... illustrates a systematic pattern of performance in correctly identifying multiple optotypes presented on multiple rows of a visual acuity chart, thereby determining the combined probability of correctly identifying multiple optotypes presented on a row of the acuity chart based on optotype size. In some implementations, the family of n+1 psychometric functions includes a function defined for each of n+1 composite responses, which may be used to correctly identify m optotypes from n optotypes per row (where m ranges from 0 to n) based on optotype size. In some implementations, the number of psychometric functions in the family of multi-opposite psychometric functions depends on the number of optotypes presented on each row. For example, for an ETDRS chart design with 5 optotypes per row, there are 6 multi-opposite psychometric functions. As another example, 4 multi-opposite psychometric functions include a family of chart designs presenting three optotypes per row, and 2 multi-opposite psychometric functions include a family of e-ETDRS tests and other computer chart designs with single-opposite presentation.

[0191] In some implementations, a family of multi-optotype psychometric functions determines the probability of a composite response to multiple optotypes based on optotype size. Given optotype size *s*, such a probability for the entire optotype sequence can be calculated by continuously multiplying the probability P(s) of correctly identifying a single optotype by the probability 1-P(s) of incorrectly identifying a single optotype. Specifically, in a particular implementation, for a row of *n* optotypes, the probability of correctly identifying *m* optotypes from *n* optotypes is calculated based on optotype size *s* using the following equation:

[0192] P m (s)=P(s) m ×(1-P(s)) n-m ,

[0193] Where m is in the range of 0 to n, and P(s) is a single-optotype psychometric function that describes the probability of correctly identifying a single optotype based on optotype size. In some implementations, the family of multi-optotype psychometric functions includes n psychometric functions, i.e., P m (s), where m is in the range of 0 to n when there are n rows of views. Figure 6A A non-restrictive example of a single optotype psychometric function 604 determined by an optotype set with ten guessed alternatives is shown, and Figure 6B A non-restrictive example of six multi-opsight psychometric functions 605 is shown, determined by a chart design that presents five optotypes per row.

[0194] Reference Figures 7A-7BIn certain implementations, scoring sensitivity chart data using algorithm S, depending on the optotype size, may require predicting a sensitivity model for the composite response to multiple optotypes. The probability of different composite responses to multiple optotypes can be predicted by determining the sensitivity model parameter pairs (e.g., threshold and range / slope) of the single optotype psychometric function 704, which in turn can determine the family of optotype psychometric functions 705 for multiple optotypes through a series multiplication of the single optotype psychometric P(s) and its complement 1-P(s). For example, the multi-optotype psychometric function Pr('m=5') for 5 optotypes is calculated as (Pr('m=1')^ 5 (1-Pr('m=1')). In this case, for example, 'm=5' means correctly identifying exactly 5 optotypes from 5 optotypes. In this implementation, the single optotype psychometric function is obtained by presenting a single optotype to the subject in each test run. Figure 7A ), generating multi-target psychometric functions by presenting multiple targets to subjects in each test run ( Figure 7B In this case, the sensitivity chart is designed to present five optotypes per row, and each of the multi-opposite psychometric functions defines the probability of correctly identifying a certain number of optotypes presented in a row of the sensitivity chart, based on the optotype size. In this particular case, only two of these multi-opposite psychometric functions are monotonic. The probability of correctly identifying no optotypes, P(m=0), decreases as the optotype size increases, while the probability of correctly identifying five optotypes increases as the optotype size increases. The pattern of intermediate composite responses is marked by alternating peaks, which represent different states of optotype size, likely indicating an increase in the number of correctly identified optotypes. In some implementations, a chart-specific psychometric function describing the expected number of correctly identified optotypes is generated by a weighted sum of these multi-opposite psychometric functions. Figures 6A-6B ).

[0195] In some implementations, computer models are used to predict a subject's performance in a visual acuity test based on implicit sensitivity parameters. (See also...) Figures 6A-6B In this particular implementation, in order to more clearly illustrate the concept of multi-target psychometric functions, such as Figure 6A The single-target psychometric function 604 and the multi-target psychometric function 605 shown are presented ( Figures 6A-6B Let be the cumulative probability distribution function. In this case, 'm = 5' indicates that exactly 5 optotypes are correctly identified from 5 optotypes. In this case, the cumulative multi-opotype psychometric function describes the probability of correctly identifying at least a given number of optotypes according to the optotype size. In some implementations, the relative stratification of these functions indicates a systematic pattern of performance, showing that the probability of correctly identifying an increasing number of optotypes increases with the size of the optotypes.

[0196] In some implementations, scoring sensitivity chart data using algorithm S, based on optotype size, requires predicting a sensitivity model that forecasts composite responses to multiple optotypes. The probabilities of different composite responses to multiple optotypes can be predicted by describing the sensitivity model parameter pairs (e.g., thresholds and ranges) of a single optotype psychometric function, which in turn can generate a family of multi-optotype psychometric functions through serial multiplication of the single optotype psychometric P(s) and its complement 1-P(s). This system and method advantageously utilize multi-optotype psychometric functions, and their application in visual acuity testing is novel.

[0197] For an example sensitivity chart design with five optotypes per row, each of the multi-optotype psychometric functions can determine the probability of correctly identifying the number of optotypes presented in each row (which is less than or equal to five) based on the optotype size. For example... Figure 6B As shown, among these psychometric functions of the composite response, only two of the multi-optotype psychometric functions are monotonic. The probability of correctly identifying no optotype, P(m=0), may decrease monotonically with increasing optotype size, while the probability of correctly identifying all five optotypes, P(m=5), may increase with increasing optotype size. Other intermediate composite responses exhibit non-monotonicity: the progression of staggered peaks provides a quantitative method for intuition, i.e., increasing optotype size increases the probability of identifying more and more correct optotypes. The staggered peaks of the multi-optotype function support the idea that the different and continuous states of optotype size represent an increased probability of correctly identifying an increasing number of optotypes with increasing optotype size. The model with multi-optotype psychometric functions presented in this paper can be a rich and complex representation of sensitivity chart performance. Existing methods cannot estimate this family of psychometric functions in a process feasible for sensitivity chart testing. A chart-specific psychometric function describing the expected number of correctly identified optotypes can be generated by the weighted sum of these multiple optotype psychometric functions.

[0198] Reference Figures 8A-8B The single-target psychometric function is shown. Figure 8A 801 in the middle) and multi-ops psychometric functions as cumulative probability functions (e.g., Figure 8B An exemplary implementation of (802) is described. In this particular implementation, the accumulated multi-optotype psychometric functions describe the probability of correctly identifying at least a given number of optotypes based on optotype size. The relative stratification of these functions indicates a systematic performance pattern, showing that the probability of correctly identifying an increasing number of optotypes increases with increasing optotype size.

[0199] In some implementations, the family of multi-opsight psychometric functions depends on a simplified assumption of statistical independence of multi-opsight responses. This assumption relies primarily on sampling opts using substitutions, which is valid for a variety of sensitivity charts. Non-limiting examples of such sensitivity charts include, at a minimum, sensitivity charts using Landolt C, Tumbling E, HOTV, and Lea optsights.

[0200] In some implementations, for sensitivity chart designs that sample optotypes without replacement, such as ETDRS, additional factors are needed to calculate the family of multi-optotype psychometric functions. This additional factor can correct for the gradual decrease in available optotypes that occurs when optotypes are sampled without replacement, particularly as the response sequence progresses to the end of the previous row of the chart.

[0201] In some implementations, a family of multi-opsight psychometric functions is used to generate a chart-specific psychometric function, which determines the expected number of correctly identified optotypes per row on a sensitivity chart based on optotype size. In some cases, the chart-specific psychometric function represents a weighted sum of the family of multi-opsight psychometric functions. For example... Figure 6C As shown, the chart-specific psychometric function 606E(s) can be specifically calculated as a weighted sum, weighted by the number of correctly identified optotypes for each multi-opposite psychometric function 605:

[0202]

[0203] In some implementations, the weighting of each psychometric function in the family of multi-target psychometric functions 605, 805 can be determined manually by the user or automatically by a computer program. In some implementations, the weighting can be based on information about the test subjects, such as population statistics, previous sensitivity test results, and / or medical history.

[0204] Similar to single-optotype psychometric functions 604, 804 and / or multi-optotype psychometric functions 605, 805, such chart-specific functions 606, 706 can be determined by two chart-specific sensitivity model parameters 701, 703 that can provide important sensitivity indicators for characterizing visual acuity:

[0205] a. Chart-specific thresholds, which define the expected number of correctly identified optotypes and optotype sizes that meet predefined standards of visual performance; and

[0206] b. A chart-specific range that defines the difference (in logMAR) between the target sizes for two levels of sensitivity performance, respectively: one above a threshold performance level,

[0207] And a performance level below the threshold.

[0208] As a non-limiting example, in the case of a standard ETDRS chart, the chart-specific psychometric functions 606 and 706 are defined by the following two chart-specific sensitivity parameters:

[0209] a. Sensitivity threshold, which defines the target size that provides the expected value for correctly identifying 3 targets out of 5 targets; and

[0210] b. Sensitivity range, which defines the number of rows of optotypes (0.10 log MAR per row) that provide the expected values ​​for correctly identifying 1.5 and 4.5 optotypes from 5 optotypes, respectively.

[0211] In the case of Snellen charts, chart-specific psychometric functions can be defined using the same parameters as ETDRS charts (sensitivity thresholds and ranges defined in sensitivity units). However, two issues need to be considered: the number of optotypes differs in each row of the Snellen chart; and only one row in the Snellen chart has the same five optotypes as the ETDRS chart. Therefore, the conversion between chart-invariant sensitivity-based functions and chart-specific psychometric functions must be performed independently for each row of the Snellen chart, whether it has more or fewer than the five optotypes used by the ETDRS chart.

[0212] In sensitivity models, sensitivity indices—sensitivity thresholds and ranges—can be derived from, for example... Figures 6A-6B The single-viewpoint chart shown contains specific psychometric functions, such as... Figures 4A-4B and / or Figure 5 The single-viewpoint chart shown is an invariant psychometric function, or as... Figure 6C and / or Figures 7A-7B The parameters of the sensitivity model are defined at the levels of the chart-specific psychometric function E(s). Depending on the optotype size, these functions can provide visual acuity metrics typically associated with visual acuity tests. Summary metrics can also be defined at the levels of multiple optotype functions, but these are naturally more difficult to interpret.

[0213] In some implementations, another important formula for the sensitivity model is defined by the graph-invariant model parameters, which define an alternative formula for the single-target psychometric function, which can:

[0214] a. Supplement the initial formula for the sensitivity model of the chart-specific indicators that generate visual acuity;

[0215] b. Follow the same calculation steps as chart-specific indicators: from single-viewpoint to multi-viewpoint to chart-specific psychometric functions, but include additional preparation steps prior to single-viewpoint psychometric functions.

[0216] In some implementations, additional preparation steps involve applying signal detection theory to determine sensitivity threshold and range parameters under a sensitivity scale (d'), and the corresponding transformation from a single-target sensitivity psychometric function to a single-target psychometric function, based on task-based graph features including guessing the number of alternatives.

[0217] The conversion from sensitivity-based psychometric functions to chart-specific single-opposite psychometric functions can be based on signal detection theory incorporating design features of sensitivity chart designs. Given the same level of sensitivity threshold (corresponding to an opposite with d'=1), the correct recognition performance of a single opposite can vary from 30-40% accuracy in a discrimination task with ten alternatives to 70-80% accuracy in a discrimination task with two alternatives.

[0218] In some implementations, the graph-invariant model proposes and helps to address the issue that some reduction in visual performance may be due to design features of sensitivity tests that are not strictly correlated with visual factors.

[0219] In some implementations, the graph-invariant model provides an illustration of probabilistic behavior, where sensitivity performance observed in sensitivity tests reflects the dual, independent contributions of sensory and decision-making processes. In some implementations, the graph-invariant model proposes that the graph-specific or empirical threshold of the psychometric function depends on the number of guessed alternatives. For example, observing a 60% correct recognition rate when presented to subjects with a sensitivity graph sampling a set of targets with ten alternatives is not equivalent to observing a 60% correct recognition rate when sampling a set of four alternatives.

[0220] In some implementations, the graph-invariant model proposes that the threshold for correctly identifying three optotypes from five optotypes depends on the total number of optotypes sampled. For statistical reasons, independent of the sensitivity of the detection system, it is easier to correctly identify three letters from five when there are four alternatives for each optotype instead of ten.

[0221] In some implementations, the graph-invariant model makes an important distinction between factors dependent on the sensitivity test and factors dependent on the observer / subject and their performance during the sensitivity test. Sensitivity to visual performance and other detection level factors are related to sensitivity test factors and guessing behavior.

[0222] In some implementations, more alternatives to the visual identifier provide more opportunities for incorrect guessing of the visual identifier and reduce recognition performance.

[0223] In some implementations, chart-specific thresholds or traditional sensitivity thresholds are confused / contaminated by guessing behavior. In some implementations, variations in recognition performance reflect factors of the sensitivity chart design. For example, a sensitivity chart design with fewer alternatives is more likely to produce a correct guess. Increasing the number of alternatives may reduce recognition performance based on factors unrelated to visual acuity. To illustrate the confusion caused by guessing behavior in recognition performance, signal detection theory proposes equations for the transition between chart-specific behavior (defined by the percentage of correctness) and chart-invariant behavior (defined in units of sensitivity: d').

[0224] In some implementations, the chart-invariant model uses a chart-invariant psychometric function to define a sensitivity index that describes sensitivity (d') based on optotype size. This function, independent of the specific sensitivity chart design used to collect data from test subjects, and such a model, can be defined by: a chart-invariant sensitivity threshold—corresponding to an optotype size (in logMAR) of d' = 2.0; and a chart-invariant sensitivity range—corresponding to the difference (in logMAR) between optotype sizes with d' values ​​of 1.0 and 3.0.

[0225] The effectiveness of sensitivity-based methods can be supported by the following observations: visual acuity measured with HOTV (using only four substitutions) is superior to that measured with the Sloan set.

[0226] Algorithm S can simultaneously estimate both chart-specific and chart-invariant sensitivity indices. The advantage of visual sensitivity indices, chart-specific indices, or chart-invariant indices is that the two sets of sensitivity parameters are interchangeable and mutually translatable.

[0227] The sensitivity model presented in this paper provides novel predictions of sensitivity chart performance and represents a data generation mechanism for inference of sensitivity index models. The model can predict a rich set of relationships between target size and performance observed during sensitivity chart testing, using only sensitivity threshold and sensitivity range parameters.

[0228] Sensitivity models can be used to analyze sensitivity chart performance by evaluating how the probability of correctly identifying multiple numbers of targets changes with the size of the targets.

[0229] In some implementations, sensitivity models provide a precise scaling analysis of sensitivity chart performance, offering a good account of probabilistic behavior observed near sensitivity thresholds. Sensitivity models may be more robust to stochastic factors that can influence termination criteria currently practiced in sensitivity testing. Sensitivity models may be more robust to probabilistic behavior occurring near thresholds: for example, due to stochastic effects, when targets are presented to a subject at sizes near a sensitivity threshold of 501, such as a chart-specific sensitivity threshold, it is almost always the same to observe 2, 3, or 4 correctly identified targets (e.g., ...). Figure 5Sensitivity models can be general and flexible: for sensitivity chart designs with varying numbers of optotypes per row, such as Snellen charts, the predictive model generates different multi-optotype psychometric functions for different rows. Sensitivity models can ultimately be applied to sensitivity chart data to infer a sensitivity model describing the visual acuity of the test subjects. Sensitivity models can be defined to remove bias from Snellen sensitivity charts. Although used in many clinical settings, Snellen has been superseded in ophthalmological trials because it does not score in the same way as the gold standard. Sensitivity models can explain the confusion between Snellen, ETDRS, HOTV, and any sensitivity model design.

[0230] Although the sensitivity model can be concisely defined with only two sensitivity parameters, it is also flexible enough to generate chart-specific sensitivity metrics that describe performance based on a particular sensitivity chart, as well as chart-invariant sensitivity metrics that describe performance independent of the sensitivity chart used to collect the data.

[0231] In some implementations, for an algorithm S that scores sensitivity chart data, each row entry in the sensitivity chart data table represents a single optotype or a single row of optotypes read from the visual sensitivity chart or subchart. Each row of optotypes can be considered as a composite stimulus determined by at least one of the following varying conditions: the number of optotypes sampled per row, the number of potential optotypes sampled from there (i.e., the number of optotypes in the library), and sampling from the optotype set with or without replacement. The response entry in the summary table is the number of correctly identified optotypes, whether for a single optotype or a single row of optotypes. In this case, the response is a ratio calculated by dividing the number of correctly identified optotypes by the total number of optotypes in the same row / column.

[0232] In some implementations, visual acuity parameters are calculated using at least one sensitivity model. In some implementations, visual acuity parameters are calculated using Bayesian adaptive estimation. In some cases, visual acuity parameters are calculated using a two-dimensional Bayesian probability density function over two sensitivity parameters. In some implementations, visual acuity parameters are calculated using Bayes' rule. In some implementations, visual acuity parameters are calculated using at least one probability function. In a further implementation, a given sensitivity parameter d is used... 1,x and d 2,y Obtain the prior probability p(r) of the response r z |d 1,x ,d 2,y Calculate the visual acuity parameters, where x and y are integers between 1 and the maximum value m of parameter d1 and the maximum value n of parameter d2, respectively, and where z is an integer between 1 and the maximum value i of the response. In some implementations, the probability p(d) of having different sensitivity parameters is... 1,x ,d 2,y) and / or the probability p(r) of obtaining different responses z In some implementations, the posterior probability function p(d) 1,x ,d 2,y| r z ) represents visual acuity, where p(d) 1,x ,d 2,y| r z )=[p(d 1,x ,d 2,y) p(r z |d 1,x ,d 2,y )] / ∑p(d 1,x ,d 2,y) p(r z |d 1,x ,d 2,y ), where ∑p(d 1,x ,d 2,y) p(r z |d 1,x ,d 2,y ) represents the sum, where x takes any integer value between 1 and m, y takes any integer value between 1 and n, and z takes any integer value between 1 and i.

[0233] In some implementations, in order to update or generate test slides in the next test run, the entropy p(d) of the pdf is calculated in the space of the sensitivity parameters. 1,x ,d 2,y| r z In some implementations, entropy H(r) z )=∑p(d 1,x ,d 2,y| r z )log(p(d 1,x ,d 2,y| r z )), where ∑ represents the sum, where x takes any integer value between 1 and m, y takes any integer value between 1 and n, and z takes any integer value between 1 and i.

[0234] In some implementations, at least one principle related to sensitivity chart design is used as input to the statistical model. In some implementations, at least one principle is selected from: the number of rows presented on each test chart or subchart, the number of targets presented on each row of the test chart or subchart, the total number of targets in the target library; and sampling from the target library with or without replacement.

[0235] The sensitivity model described herein can be incorporated into at least a portion of a fully defined model of sensitivity chart design, such as five optotypes presented in each row. In some implementations, the primary sensitivity metric provided by the sensitivity model described herein is a chart-specific sensitivity threshold, such as 701, which represents the optotype size (or equivalent to a row on the chart) when a subject correctly identifies three optotypes from five optotypes in each row (or some other specific performance criterion that defines the threshold).

[0236] Predict sensitivity chart data given candidate parameters.

[0237] In some implementations, a sensitivity model is applied to generate a series of tables containing predictions of the probability of observing collected sensitivity chart data, given the parameters of a candidate sensitivity model. Figure 1 As shown in operation 500.

[0238] For each sensitivity chart condition or row in Table 1, Table 4 shows a table of candidate parameter combinations, where the first entry is a candidate parameter and subsequent entries are predicted probabilities of the observed composite response. For each sensitivity chart condition depicted in the tables shown in Table 1, the probability of observing that response is calculated for the complete set of candidate parameters. In some implementations, for each row in Table 2, a table of the complete set of candidate parameters as shown in Table 4 can be generated.

[0239] Table 3 presents the set of candidate sensitivity parameters for a single test condition (i.e., the “data entry” of response r1 in Table 2) and their predictions for sensitivity graph data collected from test subjects.

[0240] Table 3

[0241]

[0242]

[0243] Each table represents a candidate set of sensitivity parameters and a probability estimate of the observed response under that chart or subchart condition. In this case, q0 equals q1 multiplied by q2. Optionally, multiple candidate sensitivity model entries exist for a single data entry. In some implementations, each full sensitivity chart dataset includes multiple data tables. In some implementations, analysis of the full sensitivity chart dataset involves generating multiple data tables, and in each table, one row describes each chart condition, corresponding to... Figure 17 The operation is described in the flowchart. In this case, the probability of the observed response is calculated. In other words, given the sensitivity parameters α1 and β1, and the stimulus parameters described in S1, the probability of generating a single response is calculated.

[0244] Evaluate and rank the candidate parameters of the sensitivity model.

[0245] In some implementations, after generating the full series table (e.g., Table 3), a new summary table can be calculated, such as... Figure 1 The operation is shown in 600.

[0246] In a series of tables representing the predicted probabilities of candidate parameters and observed responses, each row can summarize the presentation and response of a single or single row of optotypes. The position of an entry in each table across the entire series can correspond to a single pair of candidate sensitivity model parameters. The same entry position in each table can correspond to the same combination of candidate parameters across the series. For the full set of candidate parameters, each table in the series corresponds to the predicted probability of observing a single or composite response to the optotype size condition described in that row of that table.

[0247] In some implementations, the sensitivity chart data presented in the full series list is summarized by merging the series tables into a summary table. In some implementations, serial multiplication on the full series table of predicted probability values ​​at corresponding entry positions provides a summary probability of observed full sensitivity chart data for those specific candidate parameters. A new summary table is computed by serial multiplication of all corresponding values ​​across all rows summarized in the summary table, multiplying corresponding entries and generating a new table, as shown in Table 4. The series tables can be merged into a summary index. In some implementations, serial multiplication provides a likelihood estimate of candidate sensitivity model parameters given observations of the full sensitivity chart dataset. In some implementations, the resulting summary table represents an evaluation of candidate parameter estimates based on observed chart data. In some cases, the resulting summary table represents a likelihood calculation of candidate parameters as a function of the data. In some cases, the resulting summary table represents a likelihood calculation of candidate sensitivity parameters based on sensitivity data, which can be combined with prior probability functions on candidate sensitivity parameters, such as those calculated using Bayesian rules. In some implementations, the prior probability function can be defined by information about the subject, such as the subject's age, the state of eye disease, neuropathological conditions affecting vision, or previous visual evaluations.

[0248] Table 4 shows the aggregated probabilities of multiple responses from test subjects based on different candidate sensitivity model parameters and sensitivity models. This table can be used for retrospective scoring of the data by first predicting sensitivity chart data and then ranking the sensitivity model parameters for their predictive quality and / or goodness of fit given a composite response. The table can also be proactively applied to improve the collection of sensitivity chart data.

[0249] As shown in Table 4, for a single test condition (data entry 1), the set of candidate sensitivity parameters collected from test subjects and their predictions for the sensitivity chart data are presented. Analysis of the full sensitivity chart dataset may involve generating a series of these tables, one for each test condition, described in the row summarized in Table 1. Each table can represent a candidate set of sensitivity parameters and a probability estimate of their observed responses to that chart or sub-chart condition.

[0250] Optionally, given a sensitivity threshold and a sensitivity range / slope, the probabilities corresponding to tests with responses from r1 to rk are calculated and presented in a ranking table. In this case, given a set of sensitivity parameters, the ranking table optionally ranks the aggregate probabilities of the responses, for example, p((r1,r2,...,rk); α1,β1). In this embodiment, the set of sensitivity parameters with the highest ranking is based on the estimated sensitivity of a given response in the test.

[0251] Table 4.

[0252]

[0253] Estimate the sensitivity model parameters describing the test subjects

[0254] Table 4 shows all candidate sets of the sensitivity parameters in Table 3, and the probability of observing a composite response as calculated on the sensitivity chart data collected in Table 2 for that candidate sensitivity parameter.

[0255] Given a table of likelihood values, i.e., the probabilities of response aggregation, as shown in Table 4, a sequential ordering of the likelihood values ​​(e.g., from highest to lowest) can provide an equivalent of maximum likelihood analysis. Figure 1 As shown in operation 700, selecting the highest ranking value in the classification table and the candidate sensitivity model parameters of the response can produce a maximum likelihood estimate of the sensitivity model parameters. Alternatively, if prior information such as population statistics or the results of previous vision tests are used for weighted serial multiplication, the classification ranking can correspond to a weighted likelihood or Bayesian parameter estimate.

[0256] In some implementations, prior information may include, but is not limited to, age, genetic risk, results of previous vision tests, results of previous disease diagnoses, or any other physiological or medical information about the subject. If the table is normalized to a sum of 1, the prior information table reflects Bayesian inference. Through normalization, the table represents a Bayesian probability distribution that can be defined by its mean, median, or mode. In addition to these statistics and indicators of central tendency (e.g., mean, median, mode), there are statistical measures of the diffusion, deviation, or variability of the Bayesian posterior probability function defined on the candidate parameters of the sensitivity chart model. In some implementations, the evaluation and ranking of sensitivity model parameters includes the standard deviation or percentiles of the sensitivity model parameters. In some implementations, the evaluation and ranking of sensitivity model parameters includes statistical inference (maximum likelihood, Bayesian, or any other applicable method) that, given observed sensitivity data and predicted probabilities generated by the sensitivity performance model, produces sensitivity indices via sensitivity parameter estimation. In some implementations, the evaluation and ranking of sensitivity model parameters involve statistical inference of sensitivity threshold and range parameters that best describe the sensitivity chart data collected from test subjects.

[0257] To generate chart-specific or chart-invariant visual acuity parameters, the systems and methods described herein may include one or more sub-steps for analysis and inference:

[0258] 1. Generate a table of candidates describing sensitivity thresholds and / or range parameters, such as Table 2.

[0259] 2. Given these candidate parameter combinations and the defined features of the sensitivity chart design, a predictive model can be constructed to score the behavior of the sensitivity chart.

[0260] 3. Given a response to a single or multiple optotypes determined by a sensitivity chart design (e.g., sensitivity test data or sensitivity chart data), statistical inference can be performed by providing maximum likelihood or Bayesian analysis that estimates the sensitivity parameters.

[0261] 4. Consider the number of rows of optotypes presented during the sensitivity chart test, and consider the optotype size and response observed by the test subjects.

[0262] 5. Start by considering the parameter space, which includes sensitivity chart design parameters for a range of single viewpoint sizes.

[0263] 6. Based on signal detection theory, define the parameters in the function based on sensitivity (d').

[0264] 7. The scoring model in step 2 is used to generate predicted probabilities of parameter estimation that can be accomplished through maximum likelihood or Bayesian estimation.

[0265] 8. A model that predicts the probability of patterns in observational sensitivity chart data.

[0266] 9. The general model used for analysis includes three sets of values ​​that can be organized in a table:

[0267] 10. To establish the analytical framework, three tables are important:

[0268] a) Data collected from subjects, such as Table 1

[0269] b) Candidate sensitivity model parameters, for example, Table 2

[0270] c) Response table of probabilities defined by candidate sensitivity parameters: For the probability table, for each stimulus-response combination collected during the sensitivity test—the size of the target presented in the rows of the graph and the number of correctly identified targets, as shown in Table 4.

[0271] 11. Based on the target size, generate candidate psychometric functions with predicted probabilities of response results.

[0272] 12. For each target size presented during the test, calculate the predicted probability of observing the response for at least a portion of the candidate sensitivity parameters.

[0273] 13. For each viewpoint size presented in each row of the chart, different tables present the predicted response probability.

[0274] 14. The methods and systems used for scoring can produce two sensitivity indicators: the threshold and range of the graph-invariant psychometric function.

[0275] 15. In addition to the confidence level of the estimate, Bayesian inference can also provide an estimate of the sensitivity index, or the maximum likelihood can provide an estimate of the sensitivity index.

[0276] Identify candidate graph sets for potential collection of sensitivity graph data from test subjects.

[0277] In some implementations, the sensitivity chart set can be evaluated based on its ability to provide underlying conditions for collecting sensitivity chart data, such as... Figure 10As shown. In some implementations, the analysis estimates the probability of observing prospective sensitivity chart data, and estimates how the results analysis determines the test conditions under which sensitivity chart data should be collected. In some implementations, due to the generality of the sensitivity model, the candidate sensitivity chart set can include any sensitivity chart design. In other cases, to comply with the standards committee's recommendations, the most common approach to defining a candidate chart set may be to define a sensitivity chart that follows the same design principles as existing charts (e.g., logMAR charts (ETDRS, Bailey-Lovie), Snellen charts, Sloan low-contrast charts, and / or sub-charts of logMAR or Snellen (single-row or single-lookahead)).

[0278] In some implementations, the principles of sensitivity chart design include the nature of the visual targets presented on the sensitivity chart:

[0279] i. The size of the target set to be sampled;

[0280] ii. The number of targets displayed in each row;

[0281] iii. The number of view rows presented in each chart;

[0282] iv. Target sampling method; or a combination thereof.

[0283] Although candidate chart sets follow the same general design principles as previous techniques used for standard sensitivity charts, candidate charts can be constructed with finer resolutions and a wider range of sampled target sizes. Figures 11A-11F (and / or Figure 12). Each individual chart can show the same accuracy for sampled target size as prior art, but collectively, the candidate chart set can show higher accuracy for sampled target size than any prior art in the field. This extended library of sensitivity charts can improve accuracy and expand flexibility for sampled target size.

[0284] To improve the sampling resolution of optotype size in visual acuity testing while strictly adhering to visual acuity design standards and principles, this paper discloses a method for providing an extended library of visual acuity charts, such as... Figures 11A-11F As shown. Figures 11A-11C An exemplary sampling scheme represented by the ETDRS standard chart is shown, presenting fourteen standard target sizes ranging from -0.3 to 1.0 logMAR, with a sampling resolution of 0.10 log units between consecutive sizes. In these implementations, an anchor point of 0.35 logMAR represents the average target size of the ETDRS standard chart. By changing the anchor points of the standard constraint charts to produce a larger, complementary set of sensitivity charts, the sub-scale of sampled target sizes can be significantly improved in chart collection while maintaining individual charts' compliance with sensitivity chart standards.

[0285] In some implementations, ten sensitivity chart sets are generated by varying the anchor point of the standard chart within the .3–.4 logMAR range with an anchor point resolution of .01 logMAR. Each chart uses the same sampling resolution as the standard. However, when considered in the context of the entire set of design-constrained charts, these charts provide a finer resolution for the target size. Taken together, this set of 10 charts exhibits a higher sampling resolution for target size (.01 log units) and across a wide range of 141 different target sizes.

[0286] As an example, the ETDRS chart design includes fourteen rows of icons, five icons per row, with the size of each descending row decreasing logarithmically. A .10 decimal logarithmic series provides approximately a 25% size reduction for each consecutive row of icons. These design elements can be kept in a library of charts with different icon sizes. Figures 11A-11C It follows the constraints of these design elements. The ETDRS design principle, with three different anchor points—.30, .35, and .40 logMAR—is used. Figures 11A-11C ) and the Snellen design principle with three different anchor points ( Figures 11D-11F This generates a sample subset of the design-constrained visual acuity chart library.

[0287] The drawbacks of these chart designs may be related to the static range and resolution of the sampled target sizes. Due to the non-adaptive deterministic testing procedure, the same range of targets is presented to all patients. Furthermore, the invariant sampling resolution of the test results leads to insensitivity to measurement sensitivity values ​​relative to the measurement sensitivity values ​​presented on fourteen predetermined target sizes, corresponding to intermediate target sizes.

[0288] like Figures 12A-12C As shown, subcharts can be optionally used for visual acuity testing, representing a concentrated test of a specific area of ​​the chart. Subcharts can be subsampled from the full design-constrained sensitivity chart by presenting only a restricted set of optotypes from the full chart, which in turn is a subsample of a larger library of design-constrained charts. Presenting a set of subcharts allows visual acuity testing to be focused on a spatial subsample of the full chart (e.g., 3 rows of optotypes) instead of presenting all 14 rows of the ETDRS chart. This can be achieved through (… Figure 12A Only a sub-sampled region of the entire chart is displayed. Figure 12B The subsampled area is presented in high contrast relative to the rest of the chart, or ( Figure 12C Only the unblurred subsampled areas are presented, while the rest of the chart is blurred to distinguish the target letters of the concentrated visual acuity test.

[0289] Given a ranking of candidate sensitivity parameters, evaluate and rank the candidate sensitivity charts.

[0290] In some implementations, given the candidate parameters, a table of predicted probabilities for potential outcomes can be constructed for each row of the sensitivity table. The ordering of the sensitivity model parameters is inherent in the probability distribution defined on the model parameters.

[0291] For adaptive algorithms that personalize testing for individual patients and focus testing on individual patients, it may be necessary to compute rankings on a set of candidate sensitivity graphs. It may also be necessary to derive and compute utility functions for ranking. Sensitivity graphs can be evaluated and ranked according to different criteria. In some implementations, the sensitivity graph design can be evaluated for its utility in testing a given set of subjects. Utility functions may include one or more of the following: active learning, variance minimization, entropy minimization, maximum information gain, information gain of graph-specific parameters, information gain of graph-invariant parameters, or combinations thereof.

[0292] Existing techniques (including the QUEST method) only compute single-opposite design optimization. The advantage of the system and method disclosed herein is that they rank sensitivity chart designs on a single optotype, a single row of multiple optotypes, or a full table of optotypes. The goal of this ranking process is to personalize the test for a specific subject by selecting the best or near-best (high-ranking chart or subchart) sensitivity chart or subchart. In some implementations, composite optimization refers to optimizing for more than one optotype, where a composite or total set of optotypes is presented on a chart or subchart. Composite utility functions can be computed over the entire sensitivity chart design. Composite stimulus optimization allows for the selection of the best design for the sensitivity chart, rather than a single letter. This optimization is computed for each component of the composite stimulus. Optimized designs can be selected based on charts or subcharts. In this case, utility functions are computed for each component of the composite and then summed.

[0293] Algorithm A can provide a prospective, predictive analysis of the optimal sensitivity chart applied to sensitivity chart data collected from test subjects. For comparison, Algorithm S can provide a retrospective analysis of the sensitivity chart data collected from subjects. The basic principle of Algorithm A is optimal design selection: analyzing candidate chart sets to discover the best chart for collecting sensitivity chart data for a specific subject.

[0294] To rank sensitivity charts, utility functions can be computed on a set of candidate charts. Algorithm A can select sensitivity charts focused on specific test subjects. This is optionally accomplished by evaluating, ranking, and selecting the candidate sensitivity charts that best support the collected sensitivity chart data. The optimal design choice can be computed on chart-specific or chart-invariant parameters. In some implementations, adaptive sensitivity charts optionally include a combination of Algorithms S and A. Algorithm A has the advantage of a "closed loop" between the scoring and testing phases compared to post-testing of Algorithm S. Algorithm A can improve accuracy by iteratively testing, scoring, and retesting the sensitivity metrics.

[0295] For sensitivity tests with adaptive sensitivity charts, the optimal design choice, implemented by Algorithm A, can be computed through a computationally intensive process that anticipates the outcome of the next trial in the test and evaluates the optimal sensitivity chart design for that trial. A scoring algorithm can be applied after each trial of the test. Furthermore, the sensitivity model and scoring can be used to evaluate the possible outcomes of the next trial. In this case, the sensitivity test may include multiple trials, and in each trial, one or more responses are collected from the test subject.

[0296] In some implementations, the ranking of candidate charts can be used to select a personalized sensitivity subchart (e.g., one with the highest ranking). In this implementation, the candidate chart can include three rows with a normalized series of .10 log units between the rows. To simulate the detection of sensitivity variations (.10 log units / five-letter difference) between, for example, two medical conditions, optimized correlation patterns can be presented for these two medical conditions. The relative shift of the utility function reveals the accuracy of the combined algorithm used for personalized sensitivity testing and for detecting subtle changes in vision.

[0297] In some implementations, candidate charts can be evaluated based on their corresponding utility values ​​using a selected utility function that changes with the size of the target, such as 1407a, 1407b, and a ranking can be generated for each candidate chart. High-ranking sensitivity charts can provide high utility values ​​(e.g., in arbitrary units) and can be used for testing, while low-ranking charts are less useful and may not be usable for testing. Figure 14A The text explains the advantages of using a composite utility function 1407a to select a high-order sensitivity subfunction compared to using a single utility function 1407b. For example... Figures 14A-14B As shown, the composite utility function 1407a may share a similar curve shape as the single utility function 1407b. However, for the same viewpoint size, for example, in the range of 0.2 to 0.4, the utility value is higher. Furthermore, peaks with smaller viewpoint sizes can be obtained compared to the single utility function. In the prior art, including the QUEST method, the utility function and its optimization are based on a single viewpoint size. When considering the optimization of a three-row subgraph, the viewpoint size between rows has a standard constraint order (represented by the space between shaded areas), and the peaks 1408a and 1408b of the single and composite optimized utility functions may appear at different viewpoint sizes. Figure 14B As shown, in some implementations, using a selected utility function that changes with the size of the target (the utility function is selected to determine whether the two conditions show a sensitivity threshold change of .15 relative to .25), the utility function shows different peaks 1408a, 1408b when the stimulus is a single target row or a subgraph of a composite stimulus including multiple target sizes.

[0298] Reference Figure 15 In a particular implementation, an example of a two-slide sequence from an adaptive standard constrained visual acuity test is presented. Exemplary slides n and n+1 of the longer sequence are presented, where branches represent different potential outcomes for reading slide n. The branches are not exhaustive and are used to illustrate a small subset of the potential outcomes presented on slide n 1501. On slide n 1501, the test focuses on the most informative middle area of ​​the ETDRS chart. Following slide n 1501, which includes a 3-line ETDRS chart, subject responses are collected, and response r 1503 is also presented. In this case, the composite three-line response describes how many optotypes (up to five) are correctly reported for each of the three lines. If up to five letters per line are correctly reported on each line (the bottom branch), the test on the next slide 1502 can be adapted to present a sub-chart with smaller optotypes. In this implementation, optionally, when the subject's response does not reach the point of correctly reporting all three lines, subsequent slides present different areas of the test's full acuity chart. In this implementation, the presented optotype subchart adheres to the sensitivity chart design principle: multiple letter rows have size increments of .10 log units. For each subsequent slide, the adaptive testing algorithm focuses the testing on the optotype size most useful for the sensitivity of each individual being tested. This achieves precise and detailed sampling of optotype sizes while maintaining a .10 log unit increment between each row. In some implementations, on each slide, the optotype subchart adheres to the sensitivity chart design principle: multiple letter rows have size increments of .10 log units. For each subsequent slide, the adaptive testing algorithm focuses the testing on the optotype size most useful for the sensitivity of each individual being tested. This achieves precise and detailed sampling of optotype sizes while maintaining a .10 log unit increment between each row.

[0299] Reference Figure 18 In a particular implementation, the convergence of the adaptive sensitivity test is reflected in the Bayesian posterior probability estimate, which is obtained from multiple simulated test runs using a three-slide test. In this implementation, the widest posterior reflects the earliest sensitivity threshold estimate obtained from the patient's reading of the first column of the full 14-row ETDRS chart shown in slide 0. On the next chart presented on slide 1, a smaller five-row chart, following the ETDRS principles and specifically optimized for the patient, is presented to the patient. The Bayesian posterior function obtained after the presentation of slide 1 reflects the Bayesian update of the parameters of the visual acuity model. The decrease in the deviation of the Bayesian posterior function reflects the convergence of the sensitivity estimate. The illustration shows that the test variability decreases rapidly as the number of test slides increases.

[0300] Reference Figure 19In a specific implementation, experimental results for an adaptive graph-based sensitivity test are shown. In this implementation, subjects with normal vision are tested using a standard constrained adaptive test comprising five test slides sequentially, each slide consisting of five rows, each row containing five letters sampled from the Sloan set without replacement. In this implementation, four test runs are performed, and the variability of the test results is calculated across multiple runs. The decreasing variability of the sensitivity threshold estimate as the number of test slides increases reflects convergence.

[0301] Reference Figure 20 In a particular implementation, subjects with normal vision were tested using a standard constrained adaptive test, which consisted of five test slides in sequence, each slide comprising five rows, each row containing five letters sampled from the Sloan set without replacement. To validate the functions estimated using the adaptive test, predetermined stimuli of target size were sampled (constant stimulus method), and the probabilities of correctly reporting different numbers of targets (from 0 to 5) were calculated and plotted. Only two psychometric functions exhibited monotonicity. The probability P(0 / 5) of correctly reporting 0 targets out of 5 decreased monotonically with increasing target size. As the targets became larger, the probability of not correctly reporting any of the five targets converged to 0%. Additionally, the probability of correctly reporting all five targets out of five started at 0% for smaller target sizes and increased to 100% as target sizes increased. The intermediate multi-target psychometric functions P(1 / 5), P(2 / 5), P(3 / 5), and P(4 / 5) exhibited non-monotonic behavior of interest. As predicted by the sensitivity model, these functions can exhibit ordered and staggered peaks relative to the increase in the number of correctly identified targets. This probability function for correctly reporting 0 to 5 targets out of 5 is the multi-target psychometric function 2005. In this implementation, the chart-specific function E(s) is shown as 2006.

[0302] To supplement Figure 20 And demonstrate the effectiveness of the sensitivity model. Figure 21 It presents what can be seen from Figure 20 The cumulative probability functions obtained from the multi-optometric psychometric functions are shown. The relative positions and laminarity of these cumulative functions 2109 support the validity of the sensitivity model as an appropriate data generation model. These data provide evidence that the multi-optometric psychometric functions 605, 805, and 2005 can provide richer information about performance on visual acuity charts than single-optometric psychometric functions and / or sensitivity parameters.

[0303] In some implementations, a personalized sensitivity chart design is calculated using preliminary information about the patient's vision, representing a high ranking based on the utility function. In other implementations, results from previous sensitivity tests, other ophthalmological examinations, diagnoses of comorbidities, population statistics such as age or ethnicity, or similar information can be collected as preliminary information about the test subjects. Based on this preliminary information, an adaptive algorithm A can select the optimal chart design for the individual for the next trial or the entire testing process.

[0304] The advantage of this adaptive sensitivity testing method lies in its comprehensive framework for evaluating and optimizing the information conveyed by visual acuity testing. This method can compute m-ary responses rather than just binary responses (through the expectation of information gain).

[0305] In some implementations, an initial ranking of candidate sensitivity models can be provided based on the diagnosis of an eye disease.

[0306] In some implementations, sensitivity testing begins with a broader, simpler sensitivity task. For example, the patient may be instructed to read only the first column of a full chart of visual targets, such as... Figures 12A-12C As shown. The last row where the first target can be identified provides a reference point for starting the full testing procedure. Although effective (see below), this practice is conceptually different from obtaining the initial sample from a sensitivity chart library.

[0307] For visual acuity testing, a chart or sub-chart of optotypes is presented, and responses are collected by a technician or self-registered by the patient. In different response modes, the patient may (1) read down the first column; (2) read along the rows until instructed to stop; or (3) read the single optotype presented to them. (During actual testing, gray optotypes are identical to other optotypes; in this illustration, different shades of gray indicate that they are not prioritized for patient reporting as they are for pure black optotypes.)

[0308] The sensitivity model presented in this paper can be adapted to analyze different sensitivity test workflows: for example, reading down the first column, reading the entire chart, or reading only a single viewpoint.

[0309] As disclosed in this article, initial sampling from the chart library can take the form of, but is not limited to, the following:

[0310] a. Present a non-randomized, full-test chart—like previous techniques—a single, standard ETDRS (or Snellen) chart, and the patient attempts to read all fourteen lines. In this case, the initial sample is simply the same chart typically presented to all patients;

[0311] b. Presenting a non-randomized, short-test chart—as described above, a single, standard ETDRS chart, but with the patient only reading the first column; and

[0312] c. Sampling and presentation of randomized, full-test charts—any chart from the library. Fixed target sizes are not required for the standard ETDRS or Snellen charts currently presented to all patients;

[0313] In some implementations, the systems and methods disclosed herein can be used to analyze sensitivity metrics in these three different formats. More specifically, this analysis can be applied to sensitivity data collected on current ETDRS and Snellen charts to produce sensitivity metrics independent of chart design principles. Alternatively, patients can read the first column of an initial chart and then continue testing using a chart or subchart based on an adaptive sensitivity chart as disclosed herein. This analysis can be applied to both the initial chart and subsequent test charts simultaneously.

[0314] Response rating

[0315] In sensitivity testing, response scoring is a crucial component connecting many other component processes—especially those that estimate sensitivity metrics to determine whether to continue or terminate the test. Response scores can be explicitly correlated with test design features. The decision to terminate the test can be related to the inability to see the target. In prior art, the main difference between chart-based methods is row allocation versus letter-by-letter scoring. For single-target computerized testing, scoring is the accuracy of identification: correct or incorrect. In some implementations, detailed scoring is used: the number of letters correctly reported on each row. This detailed, row-based approach is the basis of this invention, distinguishing it from prior art methods that rely on row allocation and letter counting.

[0316] Typical scoring for sensitivity tests can follow one or more heuristics. Non-restrictive examples of such heuristics include: assigning a pass / fail rating to a line based on any letter correctly reported on that line; assigning the last line to be read correctly based on a criterion of 3 correct out of 5 visual indicators; or using a letter-by-letter scoring system to count all correctly reported letters.

[0317] In some implementations, the algorithms disclosed herein focus on visual acuity metrics collected in a single test condition. Measuring and comparing acuity metrics in two test conditions (left eye vs. right eye; low brightness vs. high brightness) is often more important. Therefore, extended algorithms that provide scoring and adaptive testing for acuity testing across multiple conditions can also be considered.

[0318] In some implementations, the scoring algorithm can be applied to a very simple “adaptive” approach: for prospective analyses during testing using established charts, the scoring algorithm can drive decisions to continue testing until the sensitivity estimate reaches a certain level of accuracy. Specifically, the same scoring algorithm can be used to analyze data collected when testing subjects with ETDRS using static, deterministic paper charts. These charts can be used for repeated testing until the scoring algorithm S returns parameter estimates with given accuracy and / or reliability. It might be more accurate to view this as a test termination application for quality control rather than an “adaptive” application.

[0319] In some implementations, the response format described herein depends on the test design, such as whether the patient is presented with a full chart, a chart sample with fewer rows, or a single optotype. The patient can read:

[0320] Read the entire chart, row by row, until it can be read in no more than two letters;

[0321] First column;

[0322] A single letter; or a combination thereof.

[0323] These three types of responses may yield different results. Previous techniques gave the same score for all letters. Technicians can register a patient's response by: identification of one or more targets presented individually or in sequence; judgment of correctness or incorrectness of the presented targets; or a combination thereof.

[0324] In existing technologies, these distinct responses are scored independently. Another approach actually considers the full optotype confusion matrix to evaluate the probability of reporting based on the identity of the different optotypes (e.g., 'O' is easier to read than 'K'). The systems and methods disclosed herein can handle three types of responses within the same theoretical framework. Detailed line-by-line scoring can be generated using relevant response distributions. In some implementations, probabilistic behavior in the sensitivity task is considered. In some implementations, scoring can be based on a relevant distribution of response outcomes according to optotype size. In some implementations, the sensitivity model considers: composite responses, response combinations, probabilistic descriptions, considering and predicting the probability of composite responses or combinations thereof.

[0325] In some implementations, scoring can be accomplished using a candidate sensitivity function data table, a lookup table of predicted probabilities, an empirical table of confusion matrices, or a combination thereof.

[0326] In prior art, many explicit testing strategies were design-dependent because the design principles defining the properties of optotypes also constrained the sensitivity metrics. This explains why these scoring methods resulted in different sensitivity estimates for an ETDRS optotype set with ten alternatives and an HOTV / Lea optotype set with four alternatives. As a result, sensitivity estimates obtained from different sensitivity tests cannot be directly compared. Recent standards have emphasized alternative methods for evaluating performance that attempt to correct for different optotype guessing levels. These standards recognize the utility of using psychometric functions to describe optotype recognition performance. For example, a sensitivity threshold is defined as 50% of the psychometric function for correct guesses.

[0327] Pc = Guessing rate + (1 - Guessing rate) * (Sigma function)

[0328] The use of psychometric functions recognizes that sensitivity performance is probabilistic, and defines a sensitivity threshold as 50% of the time on a correctly guessed psychometric function. Due to varying guessing rates for different sizes of target sets, the 50% threshold for correctly guessing a psychometric function corresponds to different levels of target performance. In summary, the main methods in the prior art for scoring sensitivity chart data to derive sensitivity indices can include:

[0329] a. Decreasing Limit - Row Allocation Method

[0330] b. Letter count – Each letter is counted in 0.02 log MAR units, or

[0331] c. Estimate 50% of the empirical or chart-specific threshold from the least squares or maximum likelihood fit of the guessed corrected psychometric function.

[0332] In this paper, the empirical or equivalent chart-specific thresholds on chart-specific or empirical functions may heavily depend on design criteria, such as the number of guessed alternatives used for visual identifiers. Therefore, the novel application of signal detection theory to characterize psychometric performance using chart-invariant units of sensitivity charts can provide a possible solution for sensitivity indices that do not depend on explicit characteristics of the sensitivity chart. The transformation between the sensitivity psychometric function defined by signal detection theory and the empirical chart-specific psychometric function can depend on the sensitivity chart characteristics proposed in the sensitivity chart criteria.

[0333] Scoring Algorithm (Algorithm S)

[0334] In some implementations, the systems and methods disclosed herein include a scoring algorithm, or algorithm S, an important component of the framework for visual acuity testing. During a sensitivity test, subjects attempt to identify optotypes of different sizes presented on a sensitivity chart or subchart. Applied to retrospective analysis, algorithm S can generate accurate sensitivity metrics from previously collected sensitivity chart data. Although the primary score is based on sensitivity data collected on the full sensitivity chart, where multiple optotypes are presented across multiple rows, the analysis can be simplified to scoring the sensitivity test using subcharts: presenting a series of optotype rows, a single optotype row, or a single optotype.

[0335] In some implementations, algorithm S exhibits the flexibility to analyze sensitivity data as a function of a specific chart design used for testing, or more generally, to analyze sensitivity data and generate a sensitivity metric independent of the chart design. This analysis enables the development of a universal, chart-invariant sensitivity metric or parameter that allows for the comparison and reconciliation of sensitivity results across many chart designs used in clinical practice.

[0336] Figure 1 A flowchart illustrating an exemplary implementation of the scoring algorithm 40 disclosed herein is shown. In this particular implementation, algorithm S (scoring algorithm) begins and determines at least one sensitivity chart design 100. In this implementation, the sensitivity chart design determination 100 is based on information provided by the user and / or information automatically generated by analyzing visual acuity charts or sub-charts used for testing subjects. In the same implementation, responses collected from testing subjects (i.e., sensitivity chart data) are summarized 400, for example, into a sensitivity chart data table. Operation 300 may determine one or more candidate sensitivity parameters (e.g., chart-specific or chart-invariant parameters) that can describe the visual acuity of the testing subject, in parallel or sequentially with the summarization of sensitivity chart responses 400. Operation 300 may occur in parallel with, before, or after operation 200. In operation 200, a sensitivity model may be defined based on chart-specific or chart-invariant parameters (e.g., chart-specific sensitivity thresholds, chart-specific sensitivity ranges, chart-invariant sensitivity thresholds, etc.). In some implementations, a set of candidate sensitivity parameter values ​​is selected as the values ​​of the sensitivity parameters in the sensitivity model determined in operation 200. Alternatively or in combination, values ​​such as those in operation 300 can be generated based on graph design parameters. Figures 6A-6CThe sensitivity model shown is used to predict the sensitivity graph data of the test subjects. In this case, the defined sensitivity model is applied retrospectively to predict the sensitivity graph data using the selected candidate parameters 500, and the probability of the aggregate response is calculated given the sensitivity model and its two parameters as shown in Table 2. Later, in the same implementation, each sensitivity model and its candidate parameters are evaluated and ranked 600 based on a table summarizing the sensitivity graph data 400 collected from the test subjects, as shown in Table 1. The sensitivity parameters can be estimated 700 based on the ranking of the sensitivity models and their candidate parameters. In this case, the sensitivity model with the highest ranking is selected as the estimate of the sensitivity parameters of the test subjects. The results of these statistical estimates are presented to the user and / or the subject 800. Optionally, the algorithm S terminates after the estimation is completed.

[0337] Reference Figure 3B In a particular implementation, at least one single-target psychometric function 604, 804 is determined in operation 230 using two chart-specific sensitivity parameters, namely candidate sensitivity parameter sets such as sensitivity threshold and sensitivity range. Figure 1 The determination of the sensitivity model for predicting sensitivity chart data or sensitivity test data is performed in operation 200. In the same case, in operation 320, a family of multi-opinion psychometric functions is calculated by serial multiplication of single-opinion psychometric functions. In the same embodiment, optionally, in operation 420, a chart-specific psychometric function is calculated using a weighted sum of the family of multi-opinion psychometric functions.

[0338] Reference Figure 3A In a particular implementation, two graph-invariant sensitivity parameters are used in operation 230. Figure 4A The sensitivity threshold 401 and / or sensitivity range 403, along with the cumulative Gaussian distribution function (CDF), determine at least one chart-invariant single-target psychometric function 404, to begin. Figure 1 The determination of the sensitivity model used to predict sensitivity chart data or sensitivity test data is described in section 200. In some embodiments, other chart-invariant sensitivity parameters such as sensitivity slope, sensitivity threshold change, etc., can be used. In some embodiments, there can be two, three, four, five, six, seven, or even more single-viewpoint psychometric functions. In this embodiment, in operation 330, the single-viewpoint psychometric function is converted from chart-invariant psychometric function 404 to chart-specific psychometric function 604, 804. In this embodiment, in operation 330, by generating, for example... Figure 5The intermediate chart-invariant function 504 shown transforms the single-viewpoint psychometric function from the chart-invariant function 404 into a chart-specific psychometric function. In the same embodiment, a family of multi-viewpoint psychometric functions is calculated by serial multiplication of the chart-specific single-viewpoint psychometric functions. In the same embodiment, optionally, the chart-specific psychometric function is calculated using a weighted sum of the family of multi-viewpoint psychometric functions.

[0339] Reference Figure 5 In a particular implementation, different coordinates from the transformation from the chart-invariant single-viewpoint psychometric function to the chart-specific single-viewpoint psychometric function 604 can be used to represent the chart-invariant single-viewpoint psychometric function, such as 404. In this implementation, Figure 5 The diagram shows the probability of correctly reporting a single letter as a function of sensitivity, given the number of alternatives. Different functions can exist depending on the number of alternatives (e.g., 2AFC, 10AFC, etc.). For example, the function of the probability of correctly responding to "A" is given when the target set (the set of guessed alternatives) is "A" and "C" (2AFC). The probability of correctly reporting "A" from a target set with 10 alternatives is even lower.

[0340] Adaptive Algorithm (Algorithm A)

[0341] In some cases, the systems and methods disclosed herein include adaptive algorithms. Adaptive algorithms can be used to improve accuracy by focusing the test on individuals. To complement the application of algorithm S in retrospective analysis, adaptive algorithm A, for example, can improve the prospective analysis of sensitivity chart data.

[0342] In some implementations, Algorithm A (Adaptive Algorithm) 20 improves the prospective analysis of sensitivity chart data by selecting the optimal sensitivity chart design. In other words, Algorithm A can select one or more sensitivity chart designs for a particular subject's future sensitivity tests based on the collected sensitivity chart data of that particular subject.

[0343] Standard paper sensitivity charts can be deterministic. The same optotypes are presented at the same size to all subjects tested using the same paper sensitivity chart. Different designs of printed sensitivity charts, presenting the same optotype size but with randomized optotypes, can prevent subjects from simply learning or guessing the chart. However, using the same set of optotype sizes may reduce test flexibility and accuracy.

[0344] Adaptive sensitivity tests using the adaptive algorithms described in this paper differ: in some implementations, the sensitivity test is dynamically adjusted to focus on each subject based on their response. Recognizing the advantages of adaptive testing over deterministic testing, the prior art has shifted towards adaptive sensitivity tests using single optotypes. The dominant adaptive testing methods in the prior art (up / down ladder; Dixon & Mood 1948; homogeneous estimation through sequential tests (ZEST); parametric estimation through sequential tests (PEST); and the QUEST method are only applicable to single optotypes. As a prominent example, the full EDTRS sensitivity chart has been reformulated into an electronic test, e-ETDRS, which combines single optotype presentation with an adaptive letter-by-letter scoring heuristic.

[0345] In some implementations, Algorithm A provides a novel, more powerful, and more general method for adaptively testing visual acuity. Algorithm A can optimize the sensitivity test using single optotypes, single-line optotypes, or multi-line optotypes. Algorithm A can be powerful enough to optimize the full sensitivity chart to be presented to the subject. In some implementations, the adaptive algorithm disclosed herein includes one or more of the following steps: 1) defining a sensitivity chart design; 2) defining a sensitivity model for predicting sensitivity chart data; 3) defining a set of candidate sensitivity model parameters; and 4) preliminary ranking of the candidate sensitivity model parameters. Figure 9 A flowchart illustrating an exemplary implementation of Algorithm A is shown.

[0346] Unlike Algorithm S, which is applied to analyze previously collected sensitivity data, Algorithm A can be applied to proactively improve and personalize the collection of sensitivity data based on prior information. To personalize sensitivity charts to improve the quality of data collected from specific subjects, Algorithm S typically involves several steps.

[0347] In some implementations, a sensitivity model for algorithm S can be defined in step 2). Algorithm A may include a sensitivity model to predict the optimal conditions for collecting future sensitivity chart data, rather than applying the sensitivity model to predict previously collected data. Algorithm A can provide which candidate sensitivity chart should be used to test a prospective analysis of the subjects. In some implementations, a table of sensitivity model parameters, as used in algorithm S, can be defined in step 3). This table contains candidate parameters that may describe the sensitivity chart data to be collected from the subjects. The sensitivity model parameters can be chart-specific or chart-invariant. In algorithm S, the sensitivity model parameters can be ordered based on likelihood values, which describe the prediction of the sensitivity chart data as a function of the sensitivity model parameters. In some implementations, the initial ordering of the sensitivity model parameters in step 4) can be uniform or defined at arbitrary proportions. In some implementations, the ordered values ​​should be normalized to 1. This effectively makes the list of ordered values ​​a probability distribution. All candidate parameters can have uniform probabilities before testing. Given a candidate parameter, a probability distribution describing the probability of the combination of parameters that describes the subject can be constructed. To begin Bayesian adaptive inference, the ranking of candidate parameters can be transformed into a probability distribution. The initial ranking of candidate sensitivity parameters can be uniform. Alternatively, preliminary prior information can be used to seed knowledge about the subject's vision. Prior information can come from, but is not limited to:

[0348] a. Group statistics, such as age;

[0349] b. Comorbid conditions such as cataracts, AMD, or other known conditions affecting the visual system;

[0350] c. Different vision tests;

[0351] d. The current test's reduction and filtering modes. For example, subjects can read down to the first column instead of reading the entire table; or a combination thereof.

[0352] Given the conversion of the ranking of candidate parameters into prior probabilities, testing can begin. Before the first test trial, there is no sensitivity data to consider and score. Therefore, the first application can take into account the history of the collected data. In this case, a uniform ranking leads to a flat prior.

[0353] In some implementations, a single-opposite psychometric function is generated using sensitivity chart data. Then, based on the sensitivity chart data predicted by the generated single-opposite psychometric function and given the goodness of fit of the data, an estimate of the subject's sensitivity threshold and range can be obtained.

[0354] Reference Figure 9In a particular implementation, algorithm A (adaptive algorithm) 20 optionally begins by determining sensitivity chart design parameters 100. After the chart design has been determined, candidate parameters (e.g., one or more sensitivity thresholds, sensitivity slopes, and / or sensitivity ranges) are optionally selected to define, for example,... Figure 1 as well as Figure 3A and / or Figure 3B Sensitivity models 200 and 300 are similarly described. Sensitivity models, as shown in Table 2, can be defined based on preliminary information from the test subjects to predict sensitivity chart data. In this case, based on... Figure 1 The sensitivity model in operation 200, the parameters in operation 300, and / or the responses collected in operation 400 generate an initial ranking 410 of the sensitivity model. A set of candidate sensitivity charts 510 to be presented to the subjects is determined at least in part by the sensitivity chart design features. Alternatively or in combination, the set of candidate sensitivity charts 510 to be presented to the subjects is determined by sensitivity chart data or sensitivity test data collected from previous tests. In some embodiments, the set of candidate sensitivity charts to be presented to the subjects is determined by other information about the test subjects, such as population statistics, previous sensitivity parameter estimates, previous sensitivity test results, etc. Optionally, in this embodiment, each candidate sensitivity chart in 510 is evaluated and ranked 610 based on the initial ranking of the sensitivity model in 410 and the sensitivity parameters. In the same case, one of the highest-ranked (i.e., top 1, top 2, and / or top 3) sensitivity charts is selected to be presented to the test subjects 710, and the adaptive algorithm ends. Optionally, a sensitivity chart can be randomly selected from the first tenths or quartiles of the sorted sensitivity charts.

[0355] For example, in addition to defining the sensitivity chart design and defining the sensitivity model based on that design, it is also necessary to define a set of candidate sensitivity model parameters to describe the sensitivity chart data to be collected from patients. Figure 9 As shown, following several typical steps of Algorithm S—defining the sensitivity chart design 100 and the sensitivity model 200, defining candidate sensitivity model parameters 300, and steps 410, 510, and 610 that differ from those in Algorithm S—the analysis of candidate sensitivity charts yields an optimal sensitivity chart design 710 that can be used to test the visual acuity of the test subjects. For example, the optimal sensitivity chart may focus on target sizes that are likely close to the subject's sensitivity threshold or within the subject's sensitivity range.

[0356] Combinatorial Algorithm

[0357] The systems and methods disclosed herein may include combinatorial algorithms. Figures 16-17Exemplary implementations of the combined algorithms disclosed herein are illustrated. The systems and methods disclosed herein address the complexity of visual acuity metrics by applying a combination of signal detection theory and probability theory to add more psychometric functions. In some implementations, the analysis herein considers how an implicit graph-invariant psychometric function can be defined by two sensitivity parameters, such as a sensitivity threshold and a sensitivity slope, or a sensitivity threshold and a sensitivity range, to generate a whole family of explicit psychometric functions observed in visual acuity tests.

[0358] Reference Figure 16 In a particular implementation, the sensitivity test optionally begins with collecting preliminary information 10 about the subject to be tested. In the same implementation, algorithm A20 is applied and a sensitivity chart is presented to the test subject 30, and algorithm S 40 is applied to generate a score for the sensitivity test. If the result from algorithm S meets the termination criteria (e.g., if the model parameter is above a determined threshold probability or if the deviation of the Bayesian posterior is below a determined level of variability), the sensitivity test no longer continues 50 and the test result of estimated visual acuity (e.g., the value of the subject's estimated visual acuity parameter) 60 is displayed. Optionally, if the result from algorithm S does not meet the termination criteria, the sensitivity test continues 50 by repeating one or more previously performed steps.

[0359] Reference Figure 17 In a particular implementation, the sensitivity test optionally begins with collecting preliminary information 10 about the subject to be tested. An initial sensitivity chart selected based on the initial information can then be presented to the subject 70, and a scoring algorithm S 40 can be applied to generate a score for the subject's visual acuity. In the same implementation, algorithm A 20 is subsequently applied to the score and / or the user's preliminary information, and a sensitivity chart adaptively generated by algorithm A is presented to the test subject 30. After responses have been collected from the subject, algorithm S 40 is applied to generate a score for the subject's sensitivity test. If the result from algorithm S meets the termination criteria, the sensitivity test no longer continues 50, and the test result 60 is displayed. Optionally, if the result from algorithm S does not meet the termination criteria, the sensitivity test continues 50 by repeating one or more previously performed steps.

[0360] Signal Detection Theory (SDT) was developed to describe the wartime performance of radar operators, distinguishing between perception level factors that make operators sensitive to target detection and decision level factors that allow operators to be free or conservative in their response to targets (Green & Swets, 1961; McMillan & Creelman, 2004, which are incorporated herein by reference in their entirety). The SDT framework provides a richer quantification beyond “correct” and “incorrect” detection behavior. SDT analysis considers how perception sensitivity (d') and decision level parameters contribute to predicting the probabilities of “hit,” “missed,” “false alarm,” and “correct rejection.” The distinction between sensitivity and decision factors and their quantification are crucial to the development of modern psychophysics. It provides a framework for identifying the contribution of factors independent of the design of the psychophysical task, relative to factors specific to task design and subject response biases and decision criteria. Lesmes et al. (2015, incorporated herein by reference in its entirety) combined Bayesian adaptive algorithms with signal detection theory to develop an adaptive method that provides threshold estimation independent of psychophysical tasks such as forced choice, simple detection, cue detection, and graded detection. In this invention, the adaptive algorithm is formulated in detail to account for the complex design of sensitivity charts and provides a design-invariant measurement of visual acuity that is independent of task factors and design principles specific to each sensitivity chart.

[0361] When combined with a design-dependent adaptive testing algorithm—tuned to be optimized for designs specific to a particular chart—it can optimally sample a library of sensitivity charts to accelerate and improve the testing and analysis of sensitivity and its variations.

[0362] Although the predictive or sensitivity models disclosed in this article have some complexity, they can be simplified and presented to patients and doctors as a display of test results.

[0363] Validation of the combinatorial algorithm

[0364] Figure 18Some exemplary aggregation simulation results of the combined algorithm are shown. The convergence of the adaptive sensitivity test is reflected in the Bayesian posterior estimates obtained from a large number of simulation test runs. The widest posterior reflects the initial sensitivity threshold estimate obtained from the patient's reading of the first column of the full 14-row ETDRS chart shown in slide 0. On the next chart presented in slide 1, a smaller five-row subchart adhering to the ETDRS principles is presented to the patient, but still represents a subchart optimized for the patient specifically. The Bayesian posterior obtained after testing with slide 1 reflects the Bayesian update of the parameters of the visual acuity model. The reduction in the deviation of the Bayesian posterior reflects the success of the information gain criterion used for the adaptive test. The convergence of sensitivity estimates continues with information obtained from the optimized presentation of slide 2. The illustration shows that the test variability decreases rapidly as the number of test slides increases.

[0365] Figure 19 Experimental results of an adaptive graph-based sensitivity test are presented. Individuals with normal vision were optionally tested using a standard constrained adaptive testing algorithm that sequentially presents five graph slides, each consisting of five rows, where each row contains five letters sampled from the Sloan set without replacement. Four test runs were performed, and the variability of the test results was calculated across multiple runs. The decreasing variability of the sensitivity threshold estimate as the number of test slides increases reflects the successful convergence of this adaptive test.

[0366] Test and Response

[0367] In some implementations, the visual acuity test described herein includes at least one test run or at least one test slide. In each test run or test slide, a chart or sub-chart of visual targets is presented to the subject, who is required to respond appropriately to test questions in order to properly record, evaluate, score, estimate acuity parameters, and continue or exit the current test.

[0368] In some implementations, the test response is recorded separately for each test run within a single test.

[0369] In some implementations, targets are sampled repeatedly from a target library. In some implementations, targets are sampled non-repeatingly from a target library.

[0370] In some implementations, the at least one visual symbol in the visual chart or subchart is generated using random sampling from a symbol library. In some implementations, the at least one visual symbol in the visual chart or subchart is generated using adaptive sampling from a symbol library.

[0371] In some implementations, the expected information gain is calculated for all possible charts or subcharts to be presented in the next trial. In some implementations, the expected entropy is calculated for all possible charts or subcharts to be presented in the next trial. In some implementations, the chart, subchart, or slide with the minimum expected entropy is presented to the subject.

[0372] termination

[0373] In some implementations, the test is terminated after at least one termination criterion is met. In some implementations, the visual acuity test is terminated after testing a predetermined number of test charts, subcharts, or slides. In some implementations, the visual acuity test is terminated after a preset accuracy level is met.

[0374] In some implementations, to shorten testing time, the test can be terminated after only one chart. Alternatively, for accuracy, the test can continue with a precise, optimized selection, as the more information collected, the more focused the test becomes on the individual. The termination rule for reading ETDRS charts can be heuristic: encourage the patient to proceed until they can no longer read 3 out of 5 targets per line.

[0375] In some implementations, the test can continue to adapt its focus to the individual rather than terminating the test. Optimization algorithms can improve the test by focusing it on the most informative area of ​​the sensitivity chart. If the test does not terminate after reading the first chart (in the first trial), algorithm A can be applied to present another sensitivity chart. If the test is to be terminated, the system and method can display the test results to the subject.

[0376] Display of test results

[0377] Reference Figure 1 The results, including model parameters, can be displayed to users and / or subjects. In some implementations, the results of graph-based sensitivity tests are visualized in a way that allows test subjects, doctors, and / or caregivers to intuitively understand the results. (See also...) Figures 22A-22D In some implementations, the Bayesian posterior probability functions of the sensitivity parameters generated at different sensitivity model parameters, i.e., the sensitivity thresholds and / or ranges, are overlaid as heatmaps on the ETDRS sensitivity graph. Figures 22A-22B In some implementations, the heatmap of the sensitivity parameter uses the darkest gray to represent the average estimate of the sensitivity threshold parameter, and the shade of gray gradually decreases as the target size gradually differs from the average target size. In some implementations, the gradient of color change in the heatmap reflects the slope of the sensitivity. In some implementations, the progression of the disease or the remedy through treatment is represented by multiple heatmaps presented on the same chart. Figure 22AIn a particular implementation, relative to Figure 22B The presented results reflect a low confidence level for the sensitivity threshold parameter. This can provide a useful representation of sensitivity results corresponding to those presented on the ETDRS standard. In another specific implementation, the Bayesian posterior probability is represented by a graphical box plot (…). Figure 22C This indicates the mean, interquartile range, and final 95% confidence interval of the sensitivity threshold estimate. Figure 22D In another embodiment, the progression of visual acuity loss is indicated by acuity measurements obtained at different assessment times. "Time 1" represents the visual acuity threshold tested at a first time point, and "Time 2" represents the visual acuity threshold tested at a second time point.

[0378] Reference Figure 23 In a particular implementation, the predicted performance is overlaid on a visual acuity chart for display. In this implementation, the probability of correctly identifying fewer than 3 targets out of a total of five targets in each row is shown in solid black directly above each row of the acuity chart. In this implementation, the probability of correctly identifying 3 targets out of a total of five targets in each row is shown in dark gray. In this implementation, the probability of correctly identifying more than 3 targets out of a total of five targets in each row is shown in light gray, and the probability of correctly identifying exactly 3 targets out of a total of five targets in each row is shown in medium gray. The probability is quantified by dividing the length of each gray bar by the length of all three bars below each row of targets.

[0379] In some implementations, visual acuity parameters can be estimated and presented on an acuity chart. Furthermore, several test results over time can be presented, conveniently showing visual progression over time.

[0380] Visualization of test results is important for patient engagement and education. For progressive vision loss, the degree and rate of decline are crucial factors. Particularly for diabetes-related vision loss, where adherence to and control of blood glucose levels directly impacts vision, visualization of test results can include a wealth of statistical information related to visual acuity. This statistical information can include one or more of the following: posterior probability distribution, parametric box plots, mean plus confidence, visual sensitivity parameters, visual threshold parameters / mean plus confidence, empirical or psychometric thresholds and sensitivity thresholds, mean / confidence intervals, sensitivity thresholds and sensitivity ranges, sensitivity thresholds and sensitivity slopes, degree of variation, overlap of probabilities of presenting variations, and overlap of sensitivity probability distributions.

[0381] Digital processing equipment

[0382] In some embodiments, the platforms, systems, media, and methods described herein include digital processing devices or their use. In further embodiments, the digital processing device includes one or more hardware central processing units (CPUs) or general-purpose graphics processing units (GPGPUs) that perform device performance. In even further embodiments, the digital processing device also includes an operating system configured to execute executable instructions. In some embodiments, the digital processing device is optionally connected to a computer network. In even further embodiments, the digital processing device is optionally connected to the Internet so that it can access the World Wide Web. In even further embodiments, the digital processing device is optionally connected to a cloud computing infrastructure. In other embodiments, the digital processing device is optionally connected to an intranet. In other embodiments, the digital processing device is optionally connected to a data storage device.

[0383] Based on the description herein, suitable digital processing devices, by way of non-limiting example, include: server computers, desktop computers, laptop computers, notebook computers, sub-notebook computers, netbook computers, notepad computers, set-top computers, media streaming devices, handheld computers, internet-connected appliances, mobile smartphones, tablet computers, personal digital assistants, video game consoles, and media devices. Those skilled in the art will recognize that many smartphones are suitable for the systems described herein. Those skilled in the art will also recognize that optional televisions, video players, and digital music players with optional computer network connectivity are suitable for the systems described herein. Suitable tablet computers include those known to those skilled in the art that have brochures, tablets, and convertible configurations.

[0384] In some implementations, the digital processing device includes an operating system configured to execute executable instructions. For example, an operating system is software that contains programs and data, manages the device's hardware, and provides services for executing applications.

[0385] In some embodiments, the device includes a storage and / or memory device. A storage and / or memory device is one or more physical devices for temporarily or permanently storing data or programs. In some embodiments, the device is volatile memory and requires power to maintain the stored information. In some embodiments, the device is non-volatile memory and retains the stored information when the digital processing device is not powered. In further embodiments, the non-volatile memory includes flash memory. In some embodiments, the non-volatile memory includes dynamic random access memory (DRAM). In some embodiments, the non-volatile memory includes ferroelectric random access memory (FRAM). In some embodiments, the non-volatile memory includes phase-change random access memory (PRAM). In other embodiments, as a non-limiting example, the device is a storage device including: CD-ROM, DVD, flash memory device, disk drive, tape drive, optical disk drive, and cloud-based storage. In further embodiments, the storage and / or memory device is a combination of devices such as those disclosed herein.

[0386] In some embodiments, the digital processing device includes a display that delivers visual information to a user. In some embodiments, the display is a liquid crystal display (LCD). In further embodiments, the display is a thin-film transistor liquid crystal display (TFT-LCD). In some embodiments, the display is an organic light-emitting diode (OLED) display. In various further embodiments, the OLED display is a passive-matrix OLED (PMOLED) or an active-matrix OLED (AMOLED) display. In some embodiments, the display is a plasma display. In other embodiments, the display is a video projector. In still other embodiments, the display is a head-mounted display, such as a VR headset, that communicates with the digital processing device.

[0387] In some embodiments, the digital processing device includes an input device for receiving information from a user. In some embodiments, the input device is a keyboard. In some embodiments, as a non-limiting example, the input device is a pointing device including a mouse, trackball, touchpad, joystick, game controller, or stylus. In some embodiments, the input device is a touchscreen or multi-touch screen. In other embodiments, the input device is a microphone that captures voice or other sound input. In other embodiments, the input device is a camera or other sensor that captures motion or visual input. In further embodiments, the input device is Kinect, Leap Motion, etc. In still further embodiments, the input device is a combination of devices such as those disclosed herein.

[0388] Reference Figure 24In a particular embodiment, the exemplary digital processing device 2401 is programmed or otherwise configured to estimate the visual acuity of a subject. Device 2401 can manage various aspects of the algorithmic and methodological steps of this disclosure. In this embodiment, digital processing device 2401 includes a central processing unit (CPU, also referred to herein as a “processor” and “computer processor”) 2405, which may be a single-core or multi-core processor, or multiple processors for parallel processing. Digital processing device 2401 also includes memory or storage location 2410 (e.g., random access memory, read-only memory, flash memory), electronic storage unit 2415 (e.g., hard disk), communication interface 2420 for communicating with one or more other systems (e.g., network adapter), and peripheral devices 2425, such as cache, other memory, data storage, and / or electronic display adapters. Memory 2410, storage unit 2415, interface 2420, and peripheral devices 2425 communicate with CPU 2405 via a communication bus (solid line) such as a motherboard. Storage unit 2415 may be a data storage unit (or data repository) for storing data. Computer system 2401 may be operatively coupled to computer network (“network”) 2430 via communication interface 2420. Network 2430 may be the Internet, the Internet of Things and / or an extranet, or an intranet and / or extranet communicating with the Internet. In some cases, network 2430 is a telecommunications and / or data network. Network 2430 may include one or more computer servers that can enable distributed computing, such as cloud computing. In some cases, network 2430 may enable a peer-to-peer network via device 2401, which allows devices coupled to device 2401 to act as clients or servers.

[0389] Continue to refer to Figure 24 The CPU 2405 can execute a series of machine-readable instructions, which may be embodied in a program or software. These instructions may be stored in a storage location such as memory 2410. The instructions may be directed to the CPU 2405, and may subsequently be programmed or otherwise configured to implement the methods of this disclosure. Examples of operations performed by the CPU 2405 may include acquiring, decoding, executing, and writing back. The CPU 2405 may be part of circuitry such as an integrated circuit. The circuitry may include one or more other components of device 2401. In some embodiments, the circuitry is an application-specific integrated circuit (ASIC) or a field-programmable gate array (FPGA).

[0390] Continue to refer to Figure 24Storage unit 2415 can store files, such as drivers, libraries, and saved programs. Storage unit 2415 can store user data, such as user preferences and user programs. In some cases, digital processing device 2401 may include one or more additional data storage units located externally, such as on a remote server that communicates with digital processing device 2401 via an intranet or the Internet.

[0391] Continue to refer to Figure 24 Digital processing device 2401 can communicate with one or more remote computer systems via network 2430. For example, computer system 2401 can communicate with a user's remote computer system. Examples of remote computer systems include personal computers (e.g., portable PCs), tablets, or tablet PCs (e.g., tablet PCs). iPad Galaxy Tab), telephone, smartphone (e.g., iPhone, Android-compatible devices (or personal digital assistant)

[0392] The methods described herein can be implemented using machine-executable code stored in an electronic storage location of digital processing device 2401, such as, for example, in memory 2410 or electronic storage unit 2415. The machine-executable or machine-readable code can be provided in software form. During use, the code can be executed by processor 2405. In some embodiments, the code can be retrieved from storage unit 2415 and stored in memory 2410 for access by processor 2405. In some cases, electronic storage unit 2415 can be excluded, and machine-executable instructions are stored in memory 2410.

[0393] Non-transitory computer-readable storage medium

[0394] In some embodiments, the platforms, systems, media, and methods disclosed herein include one or more non-transitory computer-readable storage media encoded with a program, the program including instructions executable by an operating system of an optionally networked digital processing device. In further embodiments, the computer-readable storage medium is a tangible component of the digital processing device. In still further embodiments, the computer-readable storage medium is optionally removable from the digital processing device. In some embodiments, as non-limiting examples, the computer-readable storage medium includes CD-ROMs, DVDs, flash memory devices, solid-state storage, disk drives, magnetic tape drives, optical disc drives, and cloud computing systems and services, etc. In some embodiments, the program and instructions are permanently, substantially permanently, semi-permanently, or non-transitory encoded on the medium.

[0395] Computer program

[0396] In some embodiments, the platforms, systems, media, and methods disclosed herein include at least one computer program or its intended use. A computer program includes a set of instructions executable in the CPU of a digital processing device, which are written to perform a specified task. Computer-readable instructions can be implemented as program modules that perform a specific task or implement a specific abstract data type, such as functions, subjects, application programming interfaces (APIs), data structures, etc. Based on the disclosure provided herein, those skilled in the art will recognize that computer programs can be written in various versions of various languages.

[0397] The representation of computer-readable instructions can be combined or distributed as needed in various environments. In some embodiments, a computer program includes a sequence of instructions. In some embodiments, a computer program includes multiple sequences of instructions. In some embodiments, the computer program is provided from one location. In other embodiments, the computer program is provided from multiple locations. In various embodiments, the computer program includes one or more software modules. In various embodiments, the computer program includes, in part or in whole, one or more web applications, one or more mobile applications, one or more standalone applications, one or more web browser plugins, extensions, add-ons, or accessories, or combinations thereof.

[0398] Network Applications

[0399] In some implementations, the computer program includes a web application. Based on the disclosure provided herein, those skilled in the art will recognize that, in different implementations, the web application utilizes one or more software frameworks and one or more database systems. In some implementations, the web application is in a context such as... Built on software frameworks such as .NET or Ruby on Rails (RoR). In some implementations, as a non-limiting example, the web application utilizes one or more database systems, including relational, non-relational, subject-oriented, associational, and XML database systems. In further implementations, as a non-limiting example, suitable relational database systems include SQL Server, MySQL TM and Those skilled in the art will also recognize that, in different implementations, the web application is written in one or more languages ​​and versions thereof. The web application may be written in one or more markup languages, presentation definition languages, client-side scripting languages, server-side coding languages, database query languages, or combinations thereof. In some implementations, the web application is written to a certain extent in a markup language such as Hypertext Markup Language (HTML), Extensible Hypertext Markup Language (XHTML), or Extensible Markup Language (XML). In some implementations, the web application is written to a certain extent in a presentation definition language such as Cascading Style Sheets (CSS). In some implementations, the web application is written to a certain extent in a language such as Asynchronous JavaScript and XML (AJAX). Actionscript, Javascript or Web applications are written in client-side scripting languages ​​such as ASP. In some implementations, web applications are based to a certain extent on client-side scripting languages ​​such as ASP. Perl, Java TM Java Server Pages (JSP), Hypertext Preprocessor (PHP), Python TM Ruby, Tcl, Smalltalk Or it can be written in a server-side coding language such as Groovy. In some implementations, the web application is written to some extent in a database query language such as Structured Query Language (SQL). In some implementations, the web application integrates technologies such as... Lotus Enterprise server products, etc. In some implementations, the network application includes a media player element. In various further implementations, as a non-limiting example, the media player element utilizes one or more of a variety of suitable multimedia technologies, including: HTML 5 Java TM and

[0400] Reference Figure 25In a particular implementation, the application provisioning system includes one or more databases 2500 accessed by a relational database management system (RDBMS) 2510. Suitable RDBMSs include Firebird, MySQL, PostgreSQL, SQLite, Oracle databases, Microsoft SQL Server, IBM DB2, IBM Informix, SAP Sybase, SAP Sybase, Teradata, etc. In this implementation, the application provisioning system also includes one or more application servers 2520 (such as Java servers, .NET servers, PHP servers, etc.) and one or more web servers 2530 (such as Apache, IIS, GWS, etc.). The web servers optionally expose one or more web services through an application programming interface (API) 2540. The system provides a browser-based and / or mobile local user interface via a network such as the Internet.

[0401] Reference Figure 26 In a particular implementation, the application provisioning system alternatively has a distributed, cloud-based architecture 2600 and includes elastically load-balanced, autoscaling web server resources 2610 and application server resources 2620, as well as a synchronously replicated database 2630.

[0402] Software Module

[0403] In some embodiments, the platforms, systems, media, and methods disclosed herein include software, server, and / or database modules or their use. Given the disclosure provided herein, software modules are created using techniques known to those skilled in the art, employing machines, software, and languages ​​known in the art. The software modules disclosed herein are implemented in various ways. In different embodiments, software modules include files, code segments, programming objects, programming constructs, or combinations thereof. In further different embodiments, software modules include multiple files, multiple code segments, multiple programming objects, multiple programming constructs, or combinations thereof. In different embodiments, as a non-limiting example, one or more software modules include web applications, mobile applications, and standalone applications. In some embodiments, the software module is within a computer program or application. In other embodiments, the software module is within more than one computer program or application. In some embodiments, the software module is hosted on a single machine. In other embodiments, the software module is hosted on more than one machine. In a further embodiment, the software module is hosted on a cloud computing platform. In some embodiments, the software module is hosted on one or more machines in one location. In other embodiments, the software module is hosted on one or more machines in more than one location.

[0404] database

[0405] In some embodiments, the platforms, systems, media, and methods disclosed herein include one or more databases or their use. Based on the disclosure provided herein, those skilled in the art will recognize that many databases are suitable for storing and retrieving sensitivity charts, sensitivity subcharts, preliminary information about subjects, chart data on subjects, inputs and / or outputs of the algorithms described herein, etc. In various embodiments, suitable databases, as non-limiting examples, include: relational databases, non-relational databases, subject-oriented databases, subject databases, entity-relational model databases, association databases, and XML databases. Further non-limiting examples include SQL, PostgreSQL, MySQL, Oracle, DB2, and Sybase. In some embodiments, the database is Internet-based. In further embodiments, the database is network-based. In still further embodiments, the database is cloud-based. In other embodiments, the database is based on one or more local computer storage devices.

[0406] While certain embodiments and examples have been provided in the foregoing description, the subject matter of the invention extends beyond the specific disclosed embodiments to other alternative embodiments and / or uses, and to modifications and equivalents thereof. Therefore, the scope of the appended claims is not limited to any specific embodiment described herein. For example, in any method disclosed herein, operations can be performed in any suitable order and are not necessarily limited to any particular disclosed order. Various operations may be described sequentially as a plurality of discrete operations in a manner conducive to understanding certain embodiments; however, the order of description should not be construed as implying that these operations are sequentially related. Furthermore, the systems and / or devices described herein may be embodied as integrated components or independent components.

[0407] This invention provides, but is not limited to, the following embodiments:

[0408] 1. A computer-implemented system for generating an acuity model that scores the visual acuity of a subject, the system comprising:

[0409] A digital processing device, the digital processing device including an operating system and a memory configured to execute executable instructions; and

[0410] A computer program, comprising instructions executable by the digital processing device to create a scoring application including software modules:

[0411] a) Obtain one or more sensitivity chart design parameters;

[0412] b) Collect sensitivity test data from the subjects;

[0413] c) Select the first set of chart-specific sensitivity parameters from one or more sets of candidate sensitivity parameters;

[0414] d) Generate a sensitivity model, including:

[0415] i. Generate a single-viewpoint psychometric function, the single-viewpoint psychometric function including the...

[0416] The first set of specific sensitivity parameters for the chart;

[0417] ii. Calculate a family of multi-target psychometric functions using the single-target psychometric function; and

[0418] iii. Using the aforementioned family of multi-opsight psychometric functions, generate chart-specific psychometric functions, wherein the chart-specific psychometric functions include a second set of chart-specific sensitivity parameters.

[0419] The sensitivity model is configured to estimate the probability of obtaining the sensitivity test data of the subject based on one or more sensitivity chart design parameters.

[0420] 2. The system as described in embodiment 1 further includes: after (c) and before (i), generating a sensitivity-based psychometric function comprising a first set of chart-invariant sensitivity parameters, wherein the first set of chart-invariant sensitivity parameters comprises a sensitivity threshold and a sensitivity range, and wherein the first set of chart-invariant sensitivity parameters is generated based on a first set of chart-specific sensitivity parameters.

[0421] 3. The system as described in Embodiment 2, wherein the sensitivity-based psychometric function is independent of the one or more sensitivity chart design parameters.

[0422] 4. The system as described in Embodiment 2, wherein the sensitivity-based psychometric function is generated based on the one or more sensitivity chart design parameters and one or more additional parameters of the subject, the one or more additional parameters being chart-invariant.

[0423] 5. The system as described in Embodiment 2, wherein the sensitivity-based psychometric function is generated based on signal detection theory.

[0424] 6. The system as described in Embodiment 2, wherein the sensitivity-based psychometric function is configured to describe the subject’s visual acuity performance as a d' function of one or more optotype sizes, and is independent of the one or more chart design parameters.

[0425] 7. The system as described in Embodiment 2 further includes, prior to (i), converting the sensitivity-based psychometric function into the single-target psychometric function.

[0426] 8. The system of any one of embodiments 1-7, wherein the first set of graph-specific sensitivity parameters includes a sensitivity threshold and a sensitivity range.

[0427] 9. The system as described in any one of embodiments 1-8, wherein the single-target psychometric function is chart-specific.

[0428] 10. The system of any one of embodiments 1-9, wherein the second set of the chart-specific sensitivity parameters includes a sensitivity threshold and a sensitivity range.

[0429] 11. The system of any one of embodiments 1-10, wherein the first or second set of the chart-specific sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between two test conditions, a sensitivity range change between two test conditions, a sensitivity slope change between two test conditions, or a combination thereof.

[0430] 12. The system of any one of embodiments 1-11, wherein the first or second set of the chart-specific sensitivity parameters comprises one or more numerical values.

[0431] 13. The system of any one of embodiments 1-12, wherein the first or second set of graph-specific sensitivity parameters includes at least one posterior probability density function for the parameters.

[0432] 14. The system of any one of embodiments 1-13, wherein the sensitivity test data of the subject includes data from at least two different test conditions.

[0433] 15. The system of embodiment 14, wherein the first or second set of the chart-specific sensitivity parameters comprises a posterior probability density function from sensitivity thresholds for the at least two different test conditions.

[0434] 16. The system of any one of embodiments 1-15, wherein the one or more sensitivity chart design parameters include: total number of targets, number of targets per row, target sampling method, target size, number of rows, and the subject's response or a combination thereof.

[0435] 17. The system of any one of embodiments 1-16, wherein the single optotype psychometric function is based on at least one optotype size.

[0436] 18. The system of any one of embodiments 1-17, wherein each function of the family of multi-optotype psychometric functions is based on two or more different optotype sizes, and wherein the family of multi-optotype psychometric functions includes two, three, four, five, six or any other integer number of multi-optotype psychometric functions.

[0437] 19. The system of any one of embodiments 1-18, wherein the single-target psychometric function includes the subject's sensitivity threshold and sensitivity range.

[0438] 20. The system of any one of embodiments 1-19, wherein one or more functions in the family of multi-opsight psychometric functions are based on multiple optotypes in each row of a sensitivity chart or sub-chart.

[0439] 21. The system of any one of embodiments 1-20, wherein (iii) further comprises using a weighted summation of the family of multi-target psychometric functions.

[0440] 22. The system of embodiment 21, wherein the weighting of the family of psychometric functions is based on the sensitivity chart data of the subject.

[0441] 23. The system as described in embodiment 21, wherein the weighting is determined manually by the user or automatically by a computer program.

[0442] 24. The system of any one of embodiments 1-23, wherein (b) further comprises displaying one or more sensitivity charts generated based on the one or more sensitivity chart design parameters for collecting the sensitivity test data.

[0443] 25. The system of embodiment 24, wherein each of the one or more sensitivity charts includes at least one optotype, wherein the at least one optotype is selected from an optotype library.

[0444] 26. The system of any one of embodiments 24-25, wherein the contrast of the one or more sensitivity charts is about 25%, 2.5% or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart.

[0445] 27. The system of any one of embodiments 24-26, wherein the at least one visual identifier is a letter, number, or symbol.

[0446] 28. The system of any one of embodiments 1-27, wherein one or more sets of the candidate sensitivity parameters are generated based on the one or more sensitivity chart design parameters, the sensitivity test data of the subject, or both.

[0447] 29. A computer system for generating a visual acuity model that scores the visual acuity of a subject, the system comprising:

[0448] A digital processing device, including an operating system and memory configured to execute executable instructions; and

[0449] A computer program, including instructions executable by the digital processing device to create a scoring application comprising software modules, the software modules being:

[0450] a) Obtain one or more sensitivity chart design parameters;

[0451] b) Collect sensitivity test data from the subjects;

[0452] c) Generate one or more sets of candidate sensitivity parameters based on the one or more sensitivity chart design parameters, the sensitivity test data of the subject, or both;

[0453] d) Generate a sensitivity model, including generating one or more chart-specific psychometric functions, each chart-specific psychometric function including one or more sets of chart-specific sensitivity parameters.

[0454] e) Using the sensitivity model and one or more sets of the candidate sensitivity parameters, generate the probability of observing the sensitivity test data;

[0455] f) Rank one or more sets of the candidate sensitivity parameters of the subject based on the probability; and

[0456] g) Based on the ranking, select a set from one or more sets of candidate sensitivity parameters of the subject.

[0457] 30. The system of embodiment 29, wherein each set of the candidate sensitivity parameters includes: a sensitivity threshold and a sensitivity range.

[0458] 31. The system of any one of embodiments 29-30, wherein each set of the candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between two test conditions, a sensitivity range change between two test conditions, a sensitivity slope change between two test conditions, or a combination thereof.

[0459] 32. The system of any one of embodiments 29-31, wherein each set of the candidate sensitivity parameters includes a posterior probability density function of the one or more sensitivity parameters.

[0460] 33. The system of any one of embodiments 29-32, wherein each set of the candidate sensitivity parameters includes a posterior probability density function of each of the one or more sensitivity parameters.

[0461] 34. The system of any one of embodiments 29-33, wherein (b) includes summarizing sensitivity test data from the subject and storing the sensitivity test data in a data table in the memory.

[0462] 35. The system of any one of embodiments 29-34 further includes, after (g), displaying the sensitivity test results to the subject or the user using a digital display, the results including a selected set of candidate sensitivity parameters.

[0463] 36. The system of any one of embodiments 29-35, wherein the sensitivity test data of the subject includes data from only the first test condition.

[0464] 37. The system of any one of embodiments 29-36, wherein the sensitivity test data of the subject includes data from a first test condition and a second test condition.

[0465] 38. The system of embodiment 37, wherein the selected set of candidate sensitivity parameters includes a first posterior probability density function of a sensitivity threshold or a sensitivity range.

[0466] 39. The system as described in embodiment 38, wherein b) includes selecting data having the first test condition.

[0467] 40. The system as described in embodiment 39 further includes, after g), selecting data having the second test condition from the sensitivity test data of the subject; and repeating c) to g).

[0468] 41. The system of embodiment 40, wherein the selected set of candidate sensitivity parameters includes a second posterior probability density function of the sensitivity threshold or the sensitivity range.

[0469] 42. The system of embodiment 41 further includes obtaining a difference distribution using the first posterior probability density function and the second posterior probability density function.

[0470] 43. The system as described in embodiment 42 further includes calculating the change in sensitivity threshold or sensitivity range between the first test condition and the second test condition based on the difference distribution.

[0471] 44. The system as described in embodiment 35, wherein the visual acuity test result is either chart-invariant or chart-specific.

[0472] 45. The system as described in any one of embodiments 29-44, wherein (e) is based on Bayes' theorem or maximum likelihood estimation.

[0473] 46. ​​The system of any one of embodiments 32 and 33, wherein the posterior probability density function is a Bayesian posterior probability density function.

[0474] 47. The system of any one of embodiments 38 and 41, wherein the first posterior probability density function or the second posterior probability density function is a Bayesian posterior probability density function.

[0475] 48. The system of any one of embodiments 29-47, wherein (d) comprises:

[0476] Generate a single-viewpoint psychometric function, the single-viewpoint psychometric function comprising a first set of chart-specific sensitivity parameters;

[0477] Calculate a family of multi-target psychometric functions using the single-target psychometric function; and

[0478] The family of multi-viewpoint psychometric functions is used to generate chart-specific psychometric functions, which include a second set of chart-specific sensitivity parameters.

[0479] 49. The system as described in embodiment 48 further includes, before generating the single-target psychometric function, generating a first set of graph-invariant sensitivity parameters.

[0480] 50. The system as described in embodiment 49, wherein the sensitivity-based psychometric function is independent of the one or more sensitivity chart design parameters.

[0481] 51. The system of any one of embodiments 49-50, wherein the first set of graph-invariant sensitivity parameters includes a sensitivity threshold and a sensitivity range.

[0482] 52. The system of any one of embodiments 49-51, wherein the sensitivity-based psychometric function is generated based on the one or more sensitivity chart design parameters and one or more additional parameters based on the subject, the one or more additional parameters being chart-invariant.

[0483] 53. The system of any one of embodiments 49-52, wherein the sensitivity-based psychometric function is generated based on signal detection theory.

[0484] 54. The system of any one of embodiments 50-53, wherein the sensitivity-based psychometric function is configured to describe the visual acuity performance of the subject as a d' function of one or more optotype sizes, and is independent of the one or more chart design parameters.

[0485] 55. The system as described in any one of embodiments 49-54 further includes converting the sensitivity-based psychometric function into the single-target psychometric function.

[0486] 56. The system of any one of embodiments 29-55, wherein the chart-specific sensitivity parameter set includes a sensitivity threshold and a sensitivity range.

[0487] 57. The system as described in embodiment 48, wherein the single-target psychometric function is chart-specific.

[0488] 58. The system of any one of embodiments 29-57, wherein the chart-specific sensitivity parameter set includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between two test conditions, a sensitivity range change between two test conditions, a sensitivity slope change between two test conditions, or a combination thereof.

[0489] 59. The system of any one of embodiments 29-58, wherein the chart-specific sensitivity parameter set comprises one or more numerical values.

[0490] 60. The system of any one of embodiments 29-59, wherein the graph-specific sensitivity parameter includes at least one posterior probability density function for the parameter.

[0491] 61. The system of any one of embodiments 29-60, wherein the sensitivity test data of the subject includes data from at least two different test conditions.

[0492] 62. The system of embodiment 61, wherein the chart-specific sensitivity parameter includes a posterior probability density function for sensitivity thresholds under the at least two different test conditions.

[0493] 63. The system of any one of embodiments 29-62, wherein the one or more sensitivity chart design parameters include: total number of targets, number of targets per row, target sampling method, target size, number of rows, and the subject's response or a combination thereof.

[0494] 64. The system as described in embodiment 48, wherein the single optotype psychometric function is based on at ...

Claims

1. A computer system for generating a visual acuity model that scores the visual acuity of a subject, the system comprising: Digital processing device, including an operating system and memory configured to execute executable instructions; as well as A computer program, including instructions executable by the digital processing device to create a scoring application comprising software modules configured to: a) Obtain one or more sensitivity chart design parameters; b) Collect sensitivity test data from the subjects; c) Generate one or more sets of candidate sensitivity parameters based on the one or more sensitivity chart design parameters, the sensitivity test data of the subject, or both; d) Generate a sensitivity model, including generating one or more chart-specific psychometric functions, each chart-specific psychometric function including one or more sets of chart-specific sensitivity parameters. e) Use the sensitivity model and one or more sets of the candidate sensitivity parameters to generate the probability of observing the sensitivity test data; f) Rank one or more sets of the candidate sensitivity parameters of the subject based on the probability; as well as g) Based on the ranking, select a set from one or more sets of candidate sensitivity parameters of the subject.

2. The system of claim 1, wherein each set of the candidate sensitivity parameters comprises: Sensitivity threshold and sensitivity range.

3. The system of claim 1, wherein each set of the candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between two test conditions, a sensitivity range change between two test conditions, a sensitivity slope change between two test conditions, or a combination thereof.

4. The system of claim 1, wherein each set of the candidate sensitivity parameters comprises a posterior probability density function of one or more sensitivity parameters.

5. The system of claim 1, wherein each set of the candidate sensitivity parameters comprises a posterior probability density function of each of one or more sensitivity parameters.

6. The system of claim 1, wherein (b) includes summarizing sensitivity test data from the subject and storing the sensitivity test data in a data table in the memory.

7. The system as claimed in claim 1, wherein, The software module is also configured to, after (g), display the sensitivity test results to the subject or user using a digital display, the results including a selected set of candidate sensitivity parameters.

8. The system of claim 1, wherein the sensitivity test data of the subject includes data from only the first test condition.

9. The system of claim 1, wherein the sensitivity test data of the subject includes data from a first test condition and a second test condition.

10. The system of claim 9, wherein the selected set of candidate sensitivity parameters includes a first posterior probability density function of a sensitivity threshold or a sensitivity range.

11. The system of claim 10, wherein b) includes selecting data having the first test condition.

12. The system of claim 11, wherein, The software module is also configured to, after g), select data with the second test condition from the subject's sensitivity test data; and repeat c) to g).

13. The system of claim 12, wherein the selected set of candidate sensitivity parameters includes a second posterior probability density function of the sensitivity threshold or the sensitivity range.

14. The system of claim 13, wherein, The software module is also configured to obtain the difference distribution using the first posterior probability density function and the second posterior probability density function.

15. The system of claim 14, wherein, The software module is also configured to calculate the change in sensitivity threshold or sensitivity range between the first test condition and the second test condition based on the difference distribution.

16. The system of claim 7, wherein the visual acuity test result is either chart-invariant or chart-specific.

17. The system of claim 1, wherein (e) is based on Bayes' theorem or maximum likelihood estimation.

18. The system of claim 4 or 5, wherein the posterior probability density function is a Bayesian posterior probability density function.

19. The system of claim 13, wherein the first posterior probability density function or the second posterior probability density function is a Bayesian posterior probability density function.

20. The system of claim 1, wherein (d) comprises: Generate a single-viewpoint psychometric function, the single-viewpoint psychometric function comprising a first set of chart-specific sensitivity parameters; The family of multi-target psychometric functions is calculated using the single-target psychometric function; as well as The family of multi-viewpoint psychometric functions is used to generate chart-specific psychometric functions, which include a second set of chart-specific sensitivity parameters.

21. The system of claim 20, wherein, The software module is also configured to generate a sensitivity-based psychometric function, including a first set of chart-invariant sensitivity parameters, before generating the single-target psychometric function.

22. The system of claim 21, wherein the sensitivity-based psychometric function is independent of the one or more sensitivity chart design parameters.

23. The system of claim 21, wherein the first set of graph-invariant sensitivity parameters includes a sensitivity threshold and a sensitivity range.

24. The system of claim 21, wherein the sensitivity-based psychometric function is generated based on the one or more sensitivity chart design parameters and one or more additional parameters based on the subject, the one or more additional parameters being chart-invariant.

25. The system of claim 21, wherein the sensitivity-based psychometric function is generated based on signal detection theory.

26. The system of claim 22, wherein the sensitivity-based psychometric function is configured to describe the subject’s visual acuity performance as a d' function of one or more optotype sizes, and is independent of the one or more acuity chart design parameters.

27. The system of claim 21, wherein, The software module is also configured to convert the sensitivity-based psychometric function into the single-target psychometric function.

28. The system of claim 1, wherein the chart-specific sensitivity parameter set includes a sensitivity threshold and a sensitivity range.

29. The system of claim 20, wherein the single-target psychometric function is chart-specific.

30. The system of claim 1, wherein the chart-specific sensitivity parameter set includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between two test conditions, a sensitivity range change between two test conditions, a sensitivity slope change between two test conditions, or a combination thereof.

31. The system of claim 1, wherein the chart-specific sensitivity parameter set comprises one or more numerical values.

32. The system of claim 1, wherein the chart-specific sensitivity parameter comprises at least one posterior probability density function for the parameter.

33. The system of claim 1, wherein the sensitivity test data of the subject includes data from at least two different test conditions.

34. The system of claim 33, wherein the chart-specific sensitivity parameter comprises a posterior probability density function for sensitivity thresholds under the at least two different test conditions.

35. The system of claim 1, wherein the one or more sensitivity chart design parameters include: The total number of targets, the number of targets per row, the target sampling method, the target size, the number of rows, and the subject's response or a combination thereof.

36. The system of claim 20, wherein the single optotype psychometric function is based on at least one optotype size.

37. The system of claim 20, wherein each function of the family of multi-target psychometric functions is based on two or more different target sizes.

38. The system of claim 20, wherein the single-target psychometric function includes the subject's sensitivity threshold and sensitivity range.

39. The system of claim 20, wherein one or more functions in the family of multi-ops psychometric functions are based on multiple optotypes in each row of a sensitivity chart or subchart.

40. The system of claim 20, wherein generating the chart-specific psychometric function further includes using a weighted sum of the family of multi-opsight psychometric functions.

41. The system of claim 40, wherein the weighted sum of the family of psychometric functions is based on the subject's sensitivity chart data.

42. The system of claim 40, wherein the weighting is determined manually by a user or automatically by a computer program.

43. The system of claim 1, wherein (b) further comprises displaying to the subject one or more sensitivity charts generated based on the one or more sensitivity chart design parameters.

44. The system of claim 43, wherein each of the one or more sensitivity charts includes at least one optotype, wherein the at least one optotype is selected from an optotype library.

45. The system of claim 43, wherein the contrast of the one or more sensitivity charts is 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart.

46. ​​The system of claim 44, wherein the at least one visual identifier is a letter, number, or symbol.

47. The system of claim 1, wherein, The software module is also configured to: a) Generate multiple candidate sensitivity charts based on the one or more sensitivity chart design parameters; as well as b) Based on the ranking of the plurality of candidate sensitivity charts, select one or more of the plurality of candidate sensitivity charts for the subject's sensitivity test, wherein the ranking is based on a selected set of the subject's candidate sensitivity parameters.

48. The system of claim 47, wherein, The software module is also configured to, after (b), collect data from the subject using one or more selected candidate sensitivity charts; and, if the stopping criteria are not met, repeat steps c) to g) of claim 1 and steps a) to b) of claim 47 until the stopping criteria are met.

49. A system for adaptively evaluating the visual acuity of a subject, the system comprising: Digital processing device, including an operating system and memory configured to execute executable instructions; as well as A computer program, including instructions executable by the digital processing device to create a scoring application comprising software modules, the software modules being: a) Obtain one or more sensitivity chart design parameters; b) Collect sensitivity test data from the subjects; c) Generate one or more sets of candidate sensitivity parameters based on one or more sensitivity chart design parameters and the sensitivity test data of the subject, each of the one or more sets of candidate sensitivity parameters including a ranking obtained by analyzing the sensitivity test data of the subject; d) Generate multiple candidate sensitivity charts based on the one or more sensitivity chart design parameters; e) Rank the plurality of candidate sensitivity charts based on the ranking of each of one or more sets of the candidate sensitivity parameters; as well as f) Select one or more sensitivity charts from the plurality of candidate sensitivity charts based on their ranking to be presented to the subject for sensitivity testing.

50. The system of claim 49 further includes, after f), presenting one or more selected sensitivity charts to the subject using a digital display.

51. The system of claim 50, further comprising, after f), using the one or more selected sensitivity charts, updating the sensitivity test data with new data collected from the subject.

52. The system of claim 49, further comprising repeating: presenting the one or more selected sensitivity charts to the subject using a digital display; updating the sensitivity test data with new data collected from the subject using the one or more selected sensitivity charts; and steps b) to f) until a stopping criterion is met.

53. The system of claim 49, wherein each of the one or more sets of candidate sensitivity parameters comprises: Sensitivity threshold and sensitivity range.

54. The system of claim 49, wherein each of the set of one or more candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between two test conditions, a sensitivity range change between two test conditions, a sensitivity slope change between two test conditions, or a combination thereof.

55. The system of claim 49, wherein each of the one or more sets of candidate sensitivity parameters comprises a posterior probability density function of one or more sensitivity parameters.

56. The system of claim 49, wherein each of the one or more sets of candidate sensitivity parameters includes a posterior probability density function for each of the one or more sensitivity parameters.

57. The system of claim 49, wherein (b) includes summarizing sensitivity test data from the subject.

58. The system of claim 49, wherein the sensitivity test data of the subject includes data from only the first test condition.

59. The system of claim 49, wherein the sensitivity test data of the subject includes data from a first test condition and a second test condition.

60. The system of claim 49, wherein each of one or more sets of the candidate sensitivity parameters includes a sensitivity threshold, a sensitivity range, a sensitivity slope, a sensitivity threshold change between two test conditions, a sensitivity range change between two test conditions, a sensitivity slope change between two test conditions, or a combination thereof.

61. The system of claim 49, wherein the sensitivity test data of the subject includes data from at least two different test conditions.

62. The system of claim 49, wherein each of one or more sets of the candidate sensitivity parameters includes a posterior probability density function of a sensitivity threshold or sensitivity range.

63. The system of claim 49, wherein the one or more sensitivity chart design parameters include: The total number of targets, the number of targets per row, the target sampling method, the target size, the number of rows, and the subject's response or a combination thereof.

64. The system of claim 49, wherein each of the candidate sensitivity charts includes at least one optotype, wherein the at least one optotype is selected from a optotype library.

65. The system of claim 49, wherein the contrast of one or more candidate sensitivity charts is 25%, 2.5%, or 1.5% of the normal contrast in a standard ETDRS chart or a Sloane chart.

66. The system of claim 64, wherein the at least one visual identifier is a letter, number, or symbol.

67. A computer-implemented method for generating a visual acuity model that scores the visual acuity of a subject, the method comprising: a) Obtain one or more sensitivity chart design parameters; b) Collect sensitivity test data from the subjects; c) Select the first set of chart-specific sensitivity parameters from one or more sets of candidate sensitivity parameters; d) Generate a sensitivity model, including: i) Generate a single-viewpoint psychometric function, the single-viewpoint psychometric function including a first set of chart-specific sensitivity parameters; ii) Calculate a family of multi-target psychometric functions using the single-target psychometric function; and iii) Generate chart-specific psychometric functions using the aforementioned family of multi-viewpoint psychometric functions, wherein the chart-specific psychometric functions include a second set of chart-specific sensitivity parameters. The sensitivity model is configured to estimate the probability of obtaining the sensitivity test data of the subject based on one or more sensitivity chart design parameters.

68. A computer-implemented method for scoring the visual acuity of a subject, the method comprising: a) Obtain one or more sensitivity chart design parameters; b) Collect sensitivity test data from the subjects; c) Generate one or more sets of candidate sensitivity parameters based on the one or more sensitivity chart design parameters, the sensitivity test data of the subject, or both; d) Generate a sensitivity model, including generating one or more chart-specific psychometric functions, each chart-specific psychometric function including one or more sets of chart-specific sensitivity parameters. e) Use the sensitivity model and one or more sets of the candidate sensitivity parameters to generate the probability of observing the sensitivity test data; f) Rank one or more sets of the candidate sensitivity parameters of the subject based on the probability; as well as g) Based on the ranking, select a set from one or more sets of candidate sensitivity parameters of the subject.

69. A computer-implemented method for adaptively evaluating the visual acuity of a subject, the method comprising: a) Obtain one or more sensitivity chart design parameters; b) Collect sensitivity test data from the subjects; c) Generate one or more sets of candidate sensitivity parameters based on one or more sensitivity chart design parameters and the sensitivity test data of the subject, each of the one or more sets of candidate sensitivity parameters including a ranking obtained by analyzing the sensitivity test data of the subject; d) Generate multiple candidate sensitivity charts based on the one or more sensitivity chart design parameters; e) Rank the plurality of candidate sensitivity charts based on the ranking of each of one or more sets of the candidate sensitivity parameters; as well as f) Select one or more sensitivity charts from the plurality of candidate sensitivity charts based on their ranking to be presented to the subject for sensitivity testing.

Images