Method and systems for reducing the burden of visual impairment

Abstract

A method for evaluating or improving a visual system of a subject is disclosed. The method comprises the steps of displaying multiple visual stimuli to the subject over a time period, wherein each visual stimulus comprises a motion signal; and receiving subsequent inputs from the subject indicating the subject's perception of a direction of motion of the motion signal in each visual stimulus.

Description

DOCKET NO: 1134-244 PCTTITLEMETHOD AND SYSTEMS FOR REDUCING THE BURDEN OF VISUAL IMPAIRMENT

[0001] This invention was made with government support under EY027314 awarded by the National Institutes of Health. The government has certain rights in the invention.

[0002] This application claims priority from U.S. Provisional Application No. 63 / 695,532, filed September 17, 2024, which is incorporated herein by reference. FIELD

[0003] This application relates to the field of alleviating visual impairments through visual training.BACKGROUND

[0004] In the United States, cortical visual impairment (CVI) is considered a diagnosis of exclusion and often refers to any form of visual impairment that cannot be explained via ocular disorder alone. An exact definition of CVI remains heavily debated, with the NEI recently leading a Trans-NTH workshop to create a consensus on the condition. Cortical Blindness (CB) and Cortical Visual Impairment (CVI) are two conditions arising from post- chiasmatic brain injury, resulting in a debilitating loss of vision. This loss of conscious vision can be devastating for those affected, reducing quality of life with no clinically accepted treatment available to restore lost vision. These conditions can arise from numerous etiologies, including but not limited to traumatic brain injury, developmental disorders, and tumors. However, CB is most commonly caused by stroke damage to the primary visual cortex, while CVI is more heterogenous in origin. Currently, treatments to reduce the size of this vision loss are not available for either condition.

[0005] A number of research groups have successfully improved visual performance within the CB field through the use of fixation- enforced discrimination training, even reducing the size of the CB deficit and increasing the amount of vision available to the patient. A key limitation of these interventions, however, is their slow, tedious nature. Patients are often required to train several hundred trials per day for months at a time to acquire even slight visual field improvement. In addition, as these tasks tend to utilize basic psychophysical stimuli, patients report training sessions to be boring, resulting in poor compliance and inattentiveness when performing study tasks.

[0006] There is clear room for improvement in task design. In addition, it remains unclear why focused, intentional training is required for the restoration of lost vision, while daily, passive viewing of rich visual information does not yield a similar result. As CB patients are capable of deploying spatial attention to regions of their blind field, even in the absence of awareness at the attended location, the lack of “spontaneous” improvement in this population is paradoxical.SUMMARY

[0007] One aspect of the present application relates to a method for evaluating or improving a visual system of a subject, the method comprising the steps of displaying multiple visual stimuli to the subject over a time period, wherein each visual stimulus comprises a motion signal; and receiving subsequent inputs from the subject indicating the subject's perception of a direction of motion of the motion signal in each visual stimulus and the subject’s reaction time for gaining the perception.

[0008] Another aspect of the present application relates to a non-transitory computer readable medium having stored thereon computer-readable instructions that, when executed by a processor, cause the processor to execute a method for evaluating or improving a visual system of a subject, the instructions comprising: instructions for displaying multiple visual stimuli on a display device over a time period, wherein each visual stimulus comprises a motion signal; and instructions for receiving subsequent inputs from the subject indicating the subject's perception of a direction of motion of the motion signal in each visual stimulus and the subject’s reaction time for gaining the perception.BRIEF DESCRIPTION OF THE DRAWINGS

[0009] FIGS. 1 A-1C show a novel statistical learning paradigm for use in control, CB, and CVI subjects. For the example shown in FIG. 1 A, participants are asked to indicate the perceived direction of motion for a drifting cloud stimulus (5 deg diameter, 100% contrast, 10 deg / s). Two sample directions are indicated. Dark grey background represents the CB field, while light grey indicates the intact field. FIG. IB shows that the cloud moves in one of eight directions, with participants using a number-pad to respond. Auditory feedback indicates correct / incorrect performance. In FIG. 1C, motions are presented in triplets of High or Low probability, with an additional subset never presented during training. After training, participants are asked to indicate their familiarity with all possible triplets, including the Null set. Improved speed and discrimination ability indicates explicit learningof triplets, while greater familiarity for High vs Low vs Null presentation indicates implicit learning.

[0010] FIG. 2 shows assessment for familiarity following statistical learning. Each of the High / Low / Null probability triplets are presented at a location mirrored from where training was conducted (placing the stimuli in the intact field of CB patients). Triplets are played sequentially in two intervals separated by a 500ms intertrial interval and paired as either High / Low, High / Null, or Low / Null. Participants then indicate which interval contained the more familiar triplet. No discriminations regarding motion are performed during this task. White arrows and dotted lines represent directions of motion and the blind field training location, respectively, and are not presented during the actual task.

[0011] FIG. 3 shows model of visual statistical processing within the CB and intact visual fields. In daily living, complex natural images are constantly viewed, each containing varying scene statistics. Visual features that occur with consistent spatial or temporal organization become associated by the intact visual system. When VI is damaged by a stroke, there is increased internal noise and a reduction in neuronal synchrony, making it more difficult for stimuli to be perceived at impaired locations (red circle). Visual retraining of the blind field (green circle) reduces the elevated internal noise, but not to intact levels (blue circle). Furthermore, the pathways through which visual processing occurs may not be the same as the intact system. The ability of this retrained region to perform the correlations necessary for statistical learning remains unknown.

[0012] FIGS. 4A-4C shows a novel statistical learning paradigm for use in control, CB, and CVI subjects. In the example shown in FIG. 4A, participants indicate the perceived direction of motion for a drifting cow-patterned stimulus (5 deg diameter, 100% contrast, 10 deg / s) using an Access Controller joystick. Dark grey background represents the CVI deficit, light grey indicates the intact field. FIG. 4B shows that the stimuli move in one of eight directions, presented in triplets of High or Low probability, with an additional subset never presented during training. Improved reaction time and discrimination performance during training indicates successful explicit learning. As depicted in FIG. 4C, after training, each High / Low / Null probability triplets are presented at a location mirrored from training (placing stimuli in the intact field). Triplets are played sequentially in two intervals separated by a 200ms interval and paired High / Low, High / Null, or Low / Null. Participants indicate which interval contained the more familiar triplet. No discriminations regarding motion are performed during this task. White arrows and dotted lines represent directions of motion andthe testing / training locations, respectively, and are not present during the actual task. Greater familiarity for High vs Low vs Null presentation indicates successful implicit learning.

[0013] FIGS. 5A-5D show early statistical learning exercises and outcomes. FIG. 5A depicts an exemplary learning task and stimulus exercise. FIG. 5B depicts an exemplary exercise in which stimuli were presented one at a time in triplets of high or low presentation probability. FIG. 5C depicts the result of training participants with cortical blindness on an exemplary learning task. FIG. 5D depicts a decrease in reaction time over time in abovechance performers.

[0014] While the present disclosure will now be described in detail, and it is done so in connection with the illustrative embodiments, it is not limited by the particular embodiments illustrated in the figures and the appended claims.DETAILED DESCRIPTION

[0015] References are made in detail to certain aspects and exemplary embodiments of the application, illustrating examples in the accompanying structures and figures. The aspects of the application are described in conjunction with the exemplary embodiments, including methods, materials and examples, such description is non-limiting and the scope of the application is intended to encompass all equivalents, alternatives, and modifications, either generally known, or incorporated here. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. One of skill in the art will recognize many techniques and materials similar or equivalent to those described here, which could be used in the practice of the aspects and embodiments of the present application.

[0016] As used herein, the singular forms “a”, “an”, and “the” include both singular and plural referents unless the context clearly dictates otherwise.

[0017] As used herein, the terms “about,” “approximate,” “at or about,” and “substantially” can mean that the amount or value in question can be the exact value or a value that provides equivalent results or effects as recited in the claims or taught herein. That is, it is understood that amounts, sizes, parameters, and other quantities and characteristics are not and need not be exact but may be approximate and / or larger or smaller, as desired, reflecting tolerances, conversion factors, rounding off, measurement error and the like, and other factors known to those of skill in the art such that equivalent results or effects are obtained. In some circumstances, the value that provides equivalent results or effects cannot be reasonably determined. In general, an amount, size, parameter or other quantity or characteristic is “about,” “approximate,” or “at or about” whether or not expressly stated tobe such. It is understood that where “about,” “approximate,” or “at or about” is used before a quantitative value, the parameter also includes the specific quantitative value itself, unless specifically stated otherwise.

[0018] It should be noted that ratios and other numerical data can be expressed herein in a range format. It will be further understood that the endpoints of each of the ranges are significant both in relation to the other endpoint, and independently of the other endpoint. It is also understood that there are a number of values disclosed herein, and that each value is also herein disclosed as “about” that particular value in addition to the value itself. For example, if the value “10” is disclosed, then “about 10” is also disclosed. Ranges can be expressed herein as from “about” one particular value, and / or to “about” another particular value. Similarly, when values are expressed as approximations, by use of the antecedent “about,” it will be understood that the particular value forms a further aspect. For example, if the value “about 10” is disclosed, then “10” is also disclosed.

[0019] Where a range is expressed, a further aspect includes from the one particular value and / or to the other particular value. Where a range of values is provided, it is understood that each intervening value, to the tenth of the unit of the lower limit unless the context clearly dictates otherwise, between the upper and lower limit of that range and any other stated or intervening value in that stated range, is encompassed within the disclosure. The upper and lower limits of these smaller ranges may independently be included in the smaller ranges and are also encompassed within the disclosure, subject to any specifically excluded limit in the stated range. Where the stated range includes one or both of the limits, ranges excluding either or both of those included limits are also included in the disclosure. For example, where the stated range includes one or both of the limits, ranges excluding either or both of those included limits are also included in the disclosure, e.g., the phrase “x to y” includes the range from ‘x’ to ‘y’ as well as the range greater than ‘x’ and less than ‘y’. The range can also be expressed as an upper limit, e.g., ‘about x, y, z, or less’ and should be interpreted to include the specific ranges of ‘about x’, ‘about y’, and ‘about z’ as well as the ranges of Tess than x’, less than y’, and Tess than z’. Likewise, the phrase ‘about x, y, z, or greater’ should be interpreted to include the specific ranges of ‘about x’, ‘about y’, and ‘about z’ as well as the ranges of ‘greater than x’, greater than y’, and ‘greater than z’. In addition, the phrase “about ‘x’ to ‘y’”, where ‘x’ and ‘y’ are numerical values, includes “about ‘x’ to about ‘y’”.

[0020] It is to be understood that such a range format is used for convenience and brevity, and thus, should be interpreted in a flexible manner to include not only the numericalvalues explicitly recited as the limits of the range, but also to include all the individual numerical values or sub-ranges encompassed within that range as if each numerical value and sub-range is explicitly recited. To illustrate, a numerical range of “about 0.1% to 5%” should be interpreted to include not only the explicitly recited values of about 0.1% to about 5%, but also include individual values (e.g., about 1%, about 2%, about 3%, and about 4%) and the sub-ranges (e.g., about 0.5% to about 1.1%; about 5% to about 2.4%; about 0.5% to about 3.2%, and about 0.5% to about 4.4%, and other possible sub-ranges) within the indicated range.

[0021] As used herein, the term “contrast” refers to the ratio of a given feature (luminance, color, etc.) between the lowest intensity value and the highest within a stimulus or image, or the relative intensity between a visual stimulus and the background. For example, “100% contrast” refers to pure black on pure white, or vice versa. In some embodiments of the present disclosure, for a black and white stimulus, the within-stimulus contrast level is 100%, but any given pixel in the stimulus is at 50% contrast relative to the mid-gray background. When a “lower contrast” level stimulus is presented to a subject, both the highest intensity (white) and lowest intensity (black) pixels are moved closer to the middle (grey), so both the within stimulus contrast and the contrast relative to the background become closer to zero. As used herein, the term “high contrast” refers to a contrast ratio of greater than 50% relative to background. As used herein, the term “low contrast" refers to a contrast ratio of 50% or lower relative to background.Method for evaluating or improving visual system in a subject

[0022] One aspect of the present application relates to a method for evaluating or improving a visual system of a subject. The method comprises the steps of displaying multiple visual stimuli to the subject over a time period, wherein each visual stimulus comprises a motion signal; and receiving subsequent inputs from the subject indicating the subject's perception of a direction of motion of the motion signal in each visual stimulus and the subject’s reaction time for gaining the perception. In some embodiments, the multiple visual stimuli are grouped based on statistical learning principle.

[0023] Statistical learning is a type of representational learning that occurs automatically in the absence of an explicit task and feedback, as is used in traditional training interventions for CB. Instead, participants are presented with repeated exposures to visual information linked via spatial or temporal patterns. Overtime, participants create internal representations of these patterns, implicitly learning the correlations without the need for directed training. Such learning can also result in the participant gaining explicit knowledgeof the statistical parameters of the stimuli, leading to a conscious ability to utilize this information when performing a task, in addition to the implicit benefit of unconsciously increased familiarity with test stimuli.

[0024] This automatic process is thought to be the system by which informational processing naturally develops in infants. Furthermore, statistical learning 1) is not reliant on selective attention, allowing for a potentially less strenuous intervention for CB patients, 2) results in long-term performance enhancements, and 3) can occur without the participant being aware of the patterns being learned, making it ideal for training within the visual deficit.

[0025] While promising, it is currently unclear if statistical learning can occur in the absence of awareness for the stimuli themselves, and there is debate in the field regarding the extent of its automaticity. Furthermore, as this type of learning has never been attempted in CB fields, it is unknown if successful learning within a deficit would alleviate vision loss in the same way directed perceptual training does.

[0026] The present application shows statistical learning is possible within the retrained, cortically blind field. The discovery provides key insights into viability of such a training task for CB patients, generate evidence for or against the fully automatic nature of statistical learning in vision, and deepen the understanding of how the visual system reacts to insult.

[0027] In some embodiments, the present application relates to a novel statistical learning paradigm using multiple visual stimuli to a subject over a time period, wherein each visual stimulus comprises a motion signal (e.g., a drifting cloud) and wherein the multiple visual stimuli are grouped based on statistical learning principal. Visually-intact controls are trained on this task to confirm that the paradigm is viable. In some embodiments, the motion signal moves in one of eight directions. Cortically blind patients are then tested on the same task at previously retrained locations within their afflicted hemifield.

[0028] In some embodiments the multiple visual stimuli are displayed in groups of 3, 4, 5, 6, 7, 8 or 9 visual stimuli per group, with an interval time period between two adjacent visual stimuli and a separate time period between two adjacent groups. In some embodiments, the interval time period is between 50-1000 ms. In some embodiments, the interval time period is 200 ms. In some embodiments, the separate time period is between 100-2000 ms. In some embodiments, the separate time period is 500 ms.

[0029] In some embodiments, some groups are displayed greater than 5 times (High) during a test, some groups are displayed 2-5 times (Low) during the test, and some groups aredisplayed only one time (Null) during the test. In some embodiments, wherein the stimuli groups are displayed sequentially in pairs of High / Low, High / Null or Low / Null.

[0030] In some embodiments, the motion signal is a high-contrast signal. In some embodiments, the motion signal is a low-contrast signal. In some embodiments, the motion signal has between about 1% and 100%, about 10% and 90%, about 20% and 80%, about 30% and 70%, or about 40% and 60% contrast. In some embodiments, the contrast of the motion signal is about 1%, 10%, 20%, 30%, 40%, 45%, 50%, 55%, 60%, 70%, 80%, 90%, 95%, or 100%.

[0031] In some embodiments, the motion signal has a size of 2-20 deg diameter. In some embodiments, the motion signal moves at a speed in the range of 2-30 deg / s.

[0032] In some embodiments, the motion signal is located within the subject’s vision. In some embodiments, input from the subject is received from a joystick, numeric keypad, touch pad, or keyboard. In some embodiments, the method further comprises the step of providing an evaluation of the visual system of the subject based on inputs received from the subject.

[0033] In some embodiments, the subject is an adult with cortical blindness. In some embodiments, the subject is a pediatric patient with cortical visual impairment.

[0034] In some embodiments, visually intact participants (controls) are trained on a statistical learning paradigm in which a high-contrast cloud stimulus drifts in one of eight base direction (FIGS. 1 A and IB). Controls are asked to discriminate the direction of motion presented, and provided feedback based on performance (correct or incorrect). These stimuli are placed at 7 deg eccentricity, roughly equivalent to typical training locations for CB patients. In addition, the directions of motion are grouped in triplets (FIG. 1C), categorized by High (70%) or Low prevalence (30%), or never shown (0%). Four triplets are generated for each category by randomly assigning motion directions.

[0035] In some embodiments, the training task as an intervention only contains High / Low probability stimuli, grouped into triplets, and randomly presented. After training, participants are presented with each triplet in a mirror-symmetric location, with triplets paired by presentation prevalence and shown sequentially in two intervals (FIG. 2). This is a separate task for assessment of learning. During this familiarity assessment, participants indicate which interval contained the most familiar triplet. Correctly categorizing triplets by presentation prevalence during the assessment task indicates successful implicit learning, while improved discrimination speed and performance for highly prevalent stimuli during the training task indicates successful explicit learning.Statistical learning within retrained cortically blind fields

[0036] In some embodiments, CB participants are evaluated with the statistical learning paradigm described above. The stimulus stream is placed within the participants’ impaired field, at a location which has undergone prior successful visual training and recovered measurable visual performance. Ability to perform prior direction discriminations, as well as identifying the direction of motion in the statistical learning task, confirms that the participant is processing the presented stimuli. Participant familiarity with multiple stimulus patterns is then assessed, in the same manner as above, to determine if statistical learning, either explicitly or implicitly, has occurred within the retrained visual deficit. Extent of learning can additionally be compared against motion discrimination performance as a means of assessing the impact of conscious awareness on statistical learning outcomes.

[0037] Successful development of a statistical learning paradigm for cortical blindness has both clinical and scientific value. The field of statistical learning has primarily focused on shape and form processing, with limited investigation into the statistical learning of motion. In addition, this approach provides essential information about the nature - and limitations - of statistical learning with reduced awareness of the stimulus, as the retrained CB field is not completely normal in terms of awareness or discrimination performance (FIG. 3).

[0038] As such, evidence generated for or against the efficacy of statistical learning for unconsciously processed or degraded visual information helps address ongoing discussions in the field regarding the automatic nature of statistical learning. These results also represent one of the first investigations into statistical learning within an abnormal visual system, as previous work has focused on learning disabilities such as dyslexia and autism, which reported unique processing differences for the given condition.

[0039] As such, understanding how statistical learning functions in CB is critical for understanding visual development, which is thought to utilize statistical learning in normal development, in patients with non-optimal visual processing such as childhood cataracts or neurologic conditions. Finally, successful statistical learning within a retrained CB field provides a basis for the creation of a novel training intervention in CB by establishing the viability of this type of learning in the damaged visual system. Such an intervention, given its reduced training burden and automatic nature, creates a type of training that is more easily deployable to this vulnerable and underserved patient population. As such, the data generated is used for direct training of the CB field with the statistical learning paradigm herein, with the addition of attentional manipulations and in-depth post-training assessments.

[0040] Statistical learning paradigms have traditionally relied upon form discrimination, audition, and language processing, with limited investigation into the statistical learning of visual motion. The few prior studies containing a motion paradigm have also presented motion by moving the stimulus through multiple locations, inducing retinal motion, rather than stationary stimuli that contain a motion signal. As such, these designs are not appropriate for use in CB patients, who are required to fixate centrally while stimuli are repeatedly presented at a singular location within the deficit.

[0041] The statistical learning paradigm of the present application is better tailored to this patient population. The paradigm first develops a task that is localized to a single location and uses large, high-contrast stimuli with a motion signal to elicit optimal performance in CB. To ensure participants remain engaged with the task, the presentation of visual stimuli such as triplets is embedded within a motion discrimination task, which additionally serves to measure discrimination ability and stimulus processing (conscious or unconscious) at the target location. Given the unique demands of the CB visual system and the resulting modifications to a traditional statistical learning paradigm, this study seeks to establish the efficacy and viability of this novel paradigm in visually-intact controls prior to deployment in CB patients, in addition to providing new evidence for the features of statistical learning in the motion domain.Computer readable medium

[0042] Another aspect of the present application relates to a non-transitory computer readable medium having stored thereon computer-readable instructions that, when executed by a processor, cause the processor to execute a method for evaluating or improving a visual system of a subject using the statistical learning paradigm of the present application. The instructions comprising: instructions for displaying multiple visual stimuli on a display device over a time period, wherein each visual stimulus comprises a motion signal; and instructions for receiving subsequent inputs from the subject indicating the subject's perception of a direction of motion of the motion signal in each visual stimulus and the subject’s reaction time for gaining the perception.

[0043] The present application is further illustrated by the following examples that should not be construed as limiting. The contents of all references, patents, and published patent applications cited throughout this application, as well as the Figures and Tables, are incorporated herein by reference.EXAMPLESEXAMPLE 1: Learning Task and Assessment

[0044] Ten visually-intact controls are recruited. Should significant modifications to the training task be required, additional controls are recruited to ensure statistically significant power of final analyses. Participants are recruited with the goal of closely approximating cortically-blind participants recruited. As such, participants are aged 21-75 years old, with equal proportions of males and females.

[0045] Controls are free of ocular (e.g., glaucoma or macular degeneration), visual (e.g., any visual field cuts or best-corrected acuity worse than 20 / 40), or neurological conditions (e.g., neglect or dementia) that may interfere with performing the learning task. Recruitment is performed following pre-screening and informed consent.

[0046] Controls perform the task described in Fig 1, in which they are presented with a cloud stimulus generated from non-oriented noise, drifting in one of eight directions. A high-frequency monitor (360hz refresh rate) with a calibrated, uniformly-backlit IPS display panel are utilized to ensure precise stimulus timing. Such timing is critical for the success of this study, as measurement of response time is required to determine the presence of explicit learning following training.

[0047] Participants are instructed that their reaction time are measured, in addition to task performance, to avoid confounds in response time analysis due to lags in participant responses. Stimulus parameters including size (5 deg diameter), speed (10 deg / s), and contrast (100%) are designed for optimal detection by the cortically blind visual system. The statistical learning task are performed under gaze-contingent conditions using an Eyelink Portable Duo eyetracker (SR. Research), requiring that participants maintain fixation within a 1 deg window around a central target before and during stimulus presentation. Stimuli is placed at ~7 deg eccentricity around cartesian coordinates (5,5), approximating typical CB training locations. Each of the High probability stimulus triplets is presented 25 times, randomly interleaved with 7 presentations for each of the Low probability triplets, for a total of 128 triplet presentations or 384 total trials (300 trials of High probability plus 84 trials of Low probability).

[0048] This trial count was designed to imitate previously successful motion learning studies, in which triplets were presented with 24 repetitions, as well as approximating the number of trials previously used in successful CB interventions (300 trials per location). Stimuli presentation is pseudo-random, with care taken to ensure unintended statistical formations do not occur between or across triplet presentations. Total motion signal acrosseach triplet is identical to avoid biasing triplet familiarity with signal strength. Presentation order is identical across participants, though reshuffling may occur if significant modifications to the task are made while developing the task.

[0049] Following the training session, controls are presented with an assessment task (FIG. 2). During assessment, triplets are presented at a location mirror-symmetric to where training was conducted [i.e controls trained at (5,5) are assessed at (-5,5)]. These triplets appear in two intervals, pairing High and Low probability, High and Null probability, or Low and Null probability, resulting in a total of 72 comparisons. To account for potential lapse rate, each comparison is conducted twice, for a total of 144 trials. Should inconsistent performance be observed, additional repetitions of assessment comparisons are performed. As in the training task, comparison order is identical across all participants in both aims. Participants are asked to indicate which presentation interval contained a more familiar triplet, with no feedback provided until the completion of the assessment. Performance on the assessment task is then compared to performance on the motion discrimination task.

[0050] The task design induces both explicit and implicit learning. Control performance on the motion discrimination task is expected to be near 100% correct, as discriminations of high contrast targets with long durations is exceptionally easy for the intact visual system. As such, the method of assessing explicit learning is to compare participant reaction time when discriminating the third piece of a given triplet vs all other discriminations.

[0051] If the control participant has gained explicit knowledge of the triplet, they are cued as to the likely final presentation after viewing the first two-thirds of the triplet. This results in reduced reaction time when responding to this final stimulus. Meanwhile, implicit knowledge is assessed through the assessment task in FIG. 2, where enhanced familiarity with the High probability triplets versus the Low or Null probability triplets indicates an implicit knowledge of the statistics behind stimulus presentation.

[0052] Statistical learning is known to be a fast process, with observable effects expected within a single training session, which is detected by comparing reaction times for the first and final presentation of each triplet. During the assessment task, the study expects control participants to correctly indicate better familiarity with the presented triplets, with High prevalence triplets further delineated from Low prevalence.

[0053] Optimization occurs by adjusting intertrial intervals, stimulus presentation time, and the number of triplet repetitions performed. Extensive changes to the task may warrant increased recruitment to ensure a statistically significant power.

[0054] Furthermore, the study directly asks control participants their impressions of the study and stimuli used in order to better assess any potential flaws in the paradigm. Example 2: Statistical learning occurs within retrained cortically blind fields

[0055] Following successful development of a novel statistical learning paradigm as above, the study deploys this task in CB participants to determine its efficacy at eliciting learning within the target population. This type of learning, as opposed to direct perceptual training, has never been attempted within the CB field. By utilizing participants that have undergone prior visual training, the study ensures that patients are able to perceive the presented stimuli and have the best odds for successful training, despite a classification of blindness at the region of interest. Unlike previous CB interventions, this training consists of a single session conducted in lab under gaze-contingent stimulus presentation, with no homebased training or return visits.

[0056] However, should participants become available for return testing at a later date, the study endeavors to do so as a means of assessing the persistence of the training effect. Previous motion discrimination performance data is compared to outcomes of the direction discrimination task utilized during training to assess potential differences between study designs, as these previous studies contain numerous design differences from the proposed task (ex. 2AFC vs 8AFC, staircased task difficulty, random dot stimuli, etc.). Performance on both tasks is used to measure perception of motion targets at the training location for later comparisons with extent of implicit and explicit statistical learning..

[0057] Ten cortically blind participants are recruited. Recruitment utilizes a pool of participants that have previously completed visual retraining programs, allowing the study to ensure an ability to fixate and perform psychophysical tasks. Furthermore, the study performs the statistical learning paradigm within a portion of the visual deficit that has previously trained and recovered visual performance, which are established during the participant’s time in the laboratory. As above, participants are aged 21-75 years old, with equal proportions of males and females, and are free of ocular (e.g., glaucoma or macular degeneration), visual (e.g., any visual field cuts or best-corrected acuity worse than 20 / 40), or neurological conditions (e.g., neglect or dementia) that may interfere with performing the learning task. Recruitment is performed following pre-screening and informed consent.

[0058] CB participants undergo baseline collection of Humphrey Visual Fields (Zeiss Meditec) in order to precisely map the boundaries of their visual deficit. Training locations for this aim are chosen based on ability to perform motion discriminations, established during enrollment in prior studies, while still meeting a classification of blindness according toHumphrey perimetry. As such, all testing locations contain a maximum luminance sensitivity of <10dB, a criteria for impairment established by the Social Security Administration. This determination is made by combining 10-2 and 24-2 Humphrey testing patterns and interpolating between testing points in a manner developed previously in the lab to generate high-resolution maps of combined-eye visual fields. While these target locations have been previously trained and thus have reliable motion discrimination abilities, the prior work indicates that internal noise at trained locations remains elevated, limiting the ability to perform detailed discriminations and degrading the quality of vision. These locations are thus ideal for piloting the statistical learning paradigm, as CB patients likely detect motion signal in the stimuli, but are unlikely to process the signal in a traditional manner (FIG. 3).

[0059] After establishing the target location, participants undergo the same training task described above and FIG. l. Training location eccentricity varies depending on the shape of the visual deficit and extent of recovery, but the study endeavors to remain at a similar eccentricity across all patients and controls (i.e ~7 deg). All other task parameters are identical to those established above, including stimulus size, speed, and contrast. Triplets used for all three presentation prevalences (High / Low / Null) are identical across all patients and controls. Auditory cues during stimulus presentation in the blind field are also used to ensure participants are aware of stimulus onset during the training task. Post-training assessment will also be identical to above, with stimulus streams placed at intact field locations mirror-symmetric to blind field training. At the conclusion of all training and testing, participants are also queried as to their subjective awareness of the triplet presentation and whether that awareness had any bearing on their performance.

[0060] Statistical learning has never been attempted within the CB field, and as such no preliminary data exists to support predications regarding the success of this study. However, retrained CB patients in this study are able to consciously perceive moving visual stimuli at the training location and thus have a high probability or successful statistical learning. Given the degraded nature of the retrained vision, and the heterogenous nature of CB deficits, the extent of this awareness is likely to vary and result in mixed magnitudes of learning effects. CB patients with greater baseline discrimination performance at training locations are expected to have better learning outcomes and are more likely to become consciously aware of the statistical connection within stimuli triplets (i.e explicit learning).

[0061] Baseline visual fields are stored for potential future comparisons should the participant return for additional testing. Unlike control participants, CB patients are expected to perform well below ceiling in the discrimination task. As such, explicit learning of tripletswill result in both improved reaction times and enhanced discrimination accuracy. Any enhancement to performance as a result of statistical learning is most readily apparent when comparing performance on the final stimulus in a triplet vs all other discriminations.

[0062] Furthermore, if training is able to enhance performance or reaction time, these effects are noticed by comparing performance between the first and final presentation for each triplet presented during training. Implicit knowledge of the triplets is present as enhanced familiarity with more commonly presented stimuli, as in control participants. Finally, CB patients have significant experience processing and interpreting unreliable visual information within their deficit and are expected to have insightful subjective impressions of the task and its impact on their processing ability. This information is utilized in future task modifications.

[0063] The study combines a passive statistical learning paradigm with an active motion discrimination task. This addition serves to keep patients engaged with the task, causes them to deploy spatial attention to the location of interest, and provides some modicum of visual training regardless of the success of the passive, statistical learning. Furthermore, by utilizing previously trained patients with visual recovery, the study ensures that all recruited patients are capable of performing the task and have motion processing present at the location of training. As such, the study avoids the potential confound of patients failing to learn through a failure to see, rather than resulting from unique features of the CB system.

[0064] Finally, anecdotal reports from CB patients suggest that perception of motion within the blind field, even at retrained locations with good performance, is not always accurate or consistent. Participants may have better detection of certain directions of motion (i.e Up vs Down) or perceive motion in a direction other than what is being presented. This may cause patients to develop incorrect correlations and a sense of familiarity with triplets that were not actually presented during training. To account for this possibility, the study combines multiple forms of motion discrimination data collected previously in each patient with the motion discrimination data collected during this study to determine if any directional biases are present. The study also asks patients about their perception of the motion stimuli and documents any reports of mismatched presentation and perception. If unusual or inconsistent familiarity is reported by a patient, the study attempts to correlate their impressions with performance to determine if the familiarity can be explained by altered percepts.

[0065] The outcomes of the study are not limited in their application to development of a training task. In addition, they provide novel evidence regarding how the intact and damaged visual systems process statistical data. First, the above demonstrates statistical learning in a new motion processing paradigm. As the majority of statistical learning studies have focused on form and language as modalities of interest, this study is one of the few to investigate how this process functions in this specific domain. Furthermore, the task design is unique in its focus on singular locations within the central, but not foveal, visual field. As such, this study provides new evidence and insight into statistical learning for the intact visual system.

[0066] In addition, the success or failure of statistical learning within the CB deficit has significant implications for the role of VI in statistical learning. It is thought that feedback to VI from higher-order visual areas may be a key component of visual awareness, which may explain why damage to VI heavily impairs awareness, despite the area’s lower- order hierarchical status. Should the study fail to elicit statistical learning at a location that has 1) previously been successfully trained on a motion discrimination task and 2) demonstrates strong discrimination performance, then the results would suggest that VI, either directly or through higher-order feedback, strongly contributes to this form of learning for moving visual stimuli.

[0067] Alternatively, successful statistical learning at a location deemed blind by clinical perimetry indicates that statistical learning is performed by higher visual or cognitive regions and does not rely upon VI as a substrate or is able occur as a result of the plastic visual system reacting to insult. These results warrant significant further investigation, as numerous additional possibilities regarding the development of alternative visual pathways need to be assessed.

[0068] Given the strong connection between statistical learning and development of different sensory and informational systems, these outcomes regarding statistical learning in this patient population, which is a well-established and relatively easy to utilize model, serve as a proxy for investigations into more challenging patient populations, such as pediatric cortical visual impairment.

[0069] Current models of visual development assume typical visual processing and thus are not appropriate nor accurate when discussing patients with visual impairments acquired congenitally or early in life. This study, and its future expansions, are critically informative for modifying these existing models to better account for both the degradedprocessing of the injured visual system and the plasticity that helps overcome these impairments.

[0070] The study accomplishes two primary goals: the development of a novel statistical learning paradigm, and the generation of preliminary data establishing the efficacy of such a task in cortically blind fields.

[0071] Further studies include training in naive CB patients, targeting the untrained blind field rather than the retrained regions assessed here. These efforts initially focus on testing the ability of the naive visual deficit to perform statistical learning in the same manner as training performed in the current study. These studies are expanded to assess viability for long-term home training at multiple blind field locations to induce greater amounts of visual improvement.

[0072] These studies further enhance the task through the addition of attentional precues, which have previously been effective in CB training. While statistical learning does not rely upon attention, it can be modulated by attentional deployment and thus may benefit from these manipulations in the same way as the previous perceptual learning tasks.

[0073] The results of such a study elucidate why daily passive viewing of rich visual stimuli do not automatically result in statistical learning for CB patients, in addition to inducing stronger learning effects.

[0074] An additional application is for use in pediatric cortical visual impairment. This is accomplished by embedding statistical information in videos that can be passively viewed by pediatric participants, inducing visual improvement without the need for focused task performance.

[0075] As traditional perceptual learning tasks can prove difficult to complete in pediatric patients, this intervention type provides substantial clinical and scientific benefit for this patient population.EXAMPLE 3: Statistical learning for improving visual performance within the perimetric deficit

[0076] A multisite retrospective review collects medical histories, including brain imaging and visual perimetry, from pediatric patients with CVI. Preliminary analyses from two sites have already yielded data for -100 patients. Here, through additional sites, this study increases this number to 250 (30-70 patients / site for four sites), enhancing the ability to draw robust conclusions from the datasets assessed. This increased patient pool also boosts the statistical power and capture a wider range of patients from more diverse backgrounds and etiologies. This study restricts the population of interest to those with visual field deficitsfrom post-geniculate cortical injury. Within this cohort, the study then documents primary etiologies and comorbidities that may impact vision. Perimetric visual fields are used to establish the natural history of visual deficit progression for each etiology, and to identify biomarkers that predict spontaneous changes observed in CVI - critical for interpreting training effects below. These analyses utilize methodologies for quantifying visual defects in adult occipital stroke patients, with preliminary data showing they can be readily applied to CVI perimetry.

[0077] Test the efficacy of statistical learning for improving visual performance within the perimetric deficit. A subset of CVI patients (n=10 / site, 8-17 yrs old) with visual field deficits are recruited from pediatric clinics. A novel statistical learning training paradigm is employed, with high contrast stimuli placed in the visual deficit, drifting in one of eight base directions. Patients are not told that motion directions are grouped in triplets, with particular combinations of directions having high (78%) or low (22%) prevalence. After training at home for 3 months, participants are retested in the lab. They are asked to rate their familiarity with various triplets, including samples of the high and low prevalence sets combined with ones never shown.

[0078] Correctly rating triplets by their presentation prevalence indicate successful implicit learning, while improved reaction time and discrimination performance for highly prevalent stimuli indicate successful explicit learning. Perimetry conducted pre- and posttraining are utilized to assess if learning reduces the perimetric deficit. Patient age, condition severity, and etiology are then used to identify which patient factors best correlate with greater benefit from such training. Training-induced outcomes are compared against the natural history developed above, ensuring any changes to perimetry and performance are attributable to the intervention.

[0079] Statistical learning paradigms have traditionally relied upon visual form discrimination, audition, and language processing, with limited investigation into the statistical learning of visual motion. The few prior studies containing a motion paradigm have presented motion by moving the stimulus through multiple locations, inducing retinal motion, rather than using stationary stimuli that contain a motion signal. Prior training studies in adults with stroke-induced blindness found that stimulation of discrete training locations placed just inside the blind field border was most effective at restoring visual performance.

[0080] Such a design requires participants to fixate centrally while stimuli are repeatedly presented at a singular location within the deficit. As such, the study utilizes a similar approach by developing a task that is localized to a single location and uses large,high-contrast stimuli with a motion signal to elicit optimal performance in CVI (FIG. 4). While patients are instructed to complete the presented discrimination task at the target location, the trial order are structured to create statistically linked motion direction triplets (i.e. sequential motion patterns).

[0081] Participants will not be informed that these patterns exist until the assessment task (FIG. 4C), though they are informed that a follow-up assessment is conducted. Embedding a statistical pattern within a discrimination task allows us to measure discrimination ability and stimulus processing (conscious or unconscious) at the target blindfield location.

[0082] While the majority of statistical learning studies have been performed in pediatric populations and indeed have been proposed to explain normal visual development in visually intact children, the unique features of the CVI visual system will require significant modification to traditional paradigms. As such, this establishes the efficacy and viability of this novel paradigm in CVI patients, in addition to providing new evidence for the features of statistical learning in the motion domain. While the success of the study allows the study to draw stronger conclusions about the impact of the training on the CVI deficit, it is not required to determine if statistical learning is possible in CVI participants, allowing this aim to be conducted simultaneously and independently of the study.

[0083] Subject Recruitment: Ten CVI participants per site are recruited at UR and Stanford sites from pediatric, neurology, and ophthalmology clinics. Participants are aged 8- 17 years old, with equal proportions of males and females. They are free of ocular (e.g., glaucoma or macular degeneration), visual (e.g., best-corrected acuity worse than 20 / 40), or neurological conditions (e.g., neglect) that may interfere with performing the learning task. The high rate of ocular comorbidities in CVI will necessitate careful screening to identify qualified participants. Recruitment is performed following pre-screening and after obtaining parental consent and affirmative assent of the participating child, as specified by the Institutional Review Boards at UR and Stanford.

[0084] Study design and methods: Baseline assessments: participants first undergo HVFs to precisely map the boundaries of their visual deficit. Next, to identify suitable blindfield locations for training, the study maps motion discrimination performance across the blind field. A high-frequency monitor (360hz refresh rate) with a calibrated, uniformly backlit LED / LCD hybrid display panel ensures precise stimulus timing. Responses are measured using an Access Controller (Sony Interactive Entertainment) joystick. Gazecontingent conditions are used with an Eyelink Portable Duo eyetracker (SR. Research),requiring participants to fixate within a 1 deg window around a central target before and during stimulus presentation. Head-free eye-tracking will allow for optimal patient comfort. Training locations for this aim are chosen based on ability to perform coarse (left vs right) motion discriminations using a random dot stimulus?, while still meeting a classification of blindness according to composite, interpolated perimetry. As such, all testing locations contain a maximum luminance sensitivity <10dB, a criterion for impairment established by the Social Security Administration.

[0085] In preferred embodiments, the exact number of displays is a ratio. The numbers presented (5x and 2-5x) are calculated from the number of total trials. In various embodiments, there are more or less trials, with more or less sets of stimuli, and the exact number can change.

[0086] In a preferred embodiment, the null set is zero presentations during training and is only ever presented in the subsequent analysis task.

[0087] This determination is made by combining monocularly collected 24-2 HVFs between both eyes and interpolating between testing points in a manner developed to generate high -resolution maps of combined-eye visual fields and identical to the analyses used in the study. Fields are required to have high reliability, defined as fixation losses, false positives, and false negatives all below 20%. Preliminary analysis of fields in the study suggests these criteria is easily achieved by CVI and CB patients. Failure to produce reliable fields or to perform motion discriminations will result in removal from the study. Once rapid mapping finds a suitable target training location, participants be taught to perform the Statistical Learning task. They are presented with a single moving cow-like pattern generated from non-oriented noise, drifting in one of eight directions (FIG. 4A) at the selected training location.

[0088] Responses are measured using an Access Controller. Participants are instructed that their reaction times are measured, in addition to task performance, to avoid confounds in response time analysis due to lags in participant responses. Efforts are made to promote participant engagement by presenting a running score during training, using a cowshaped stimulus, and using “mooing” noises for correct or incorrect feedback. Stimulus parameters including size (5 deg diameter), speed (10 deg / s), and contrast (100%) are designed for optimal detection by the cortically injured visual system. For the Statistical Learning task, each High-Probability stimulus triplet is presented 25 times, randomly interleaved with 7 presentations for each of the Low-Probability triplets, for a total of 128triplet presentations or 384 total trials (300 trials of High probability plus 84 trials of Low probability; FIG. 4B).

[0089] This trial count was designed to imitate previously successful motion learning studies 18, in which triplets were presented with 24 repetitions, as well as approximating the number of trials previously used in successful interventions for post-stroke adults (300 trials per location). Stimuli presentations are pseudo-random, with care taken to ensure unintended statistical formations do not occur between or across triplet presentations. Total motion signal across each triplet is identical to avoid biasing triplet familiarity with signal strength. Auditory cues during stimulus presentation in the blind field will also be used to ensure participants are aware of stimulus onset during the task. Following a Statistical Learning session, participants will perform an Assessment task (FIG. 4C), in which triplets are presented at a location mirror-symmetric to where training was conducted [i.e., those exposed at (5,5) are assessed at (-5,5)]. This is to ensure any failure to recognize a triplet is not due to a lack of awareness during testing. Triplets will appear in two intervals, pairing High- and Low-Probability, High and Null-Probability, or Low and Null-Probability, resulting in a total of 72 comparisons. To account for potential lapse rate, each comparison is conducted twice, for a total of 144 trials. Should inconsistent performance be observed, additional repetitions are performed.

[0090] As in the training task, trial order is identical across participants, who are asked to indicate which presentation interval contained a more familiar triplet, with no feedback provided until completion of the assessment. Fixation will remain enforced during assessment, and participants are queried as to their subjective awareness of the triplet presentation and whether that awareness had any bearing on their performance during assessment.

[0091] Home-training intervention and post-training assessment: after the in-lab baseline visit, participants are provided a laptop with a program customized to train their target blind-field location at home. This training is identical to the Statistical Learning session performed in-lab (FIG. 3 A). Participants will train at home for 3 months, five days per week, then return to lab for follow-up testing identical to the baseline visit. This will include a repeat of the Statistical Learning task under gaze-contingent conditions, assessment task, and perimetry.

[0092] Analyses and Basic Predictions: Statistical learning is often used in pediatric populations, and statistical learning of motion specifically is known to occur in adult, visually intact controls; thus, the study expects that the proposed task design will induce both explicitand implicit learning in CVI patients. Motion discrimination performance is expected to be near 100% correct in the visually intact hemifield, as discriminations of high contrast targets with long durations are exceptionally easy for the intact visual system. Likewise, statistical learning is expected to occur at the intact field location during both visits. Initial motion discriminations within the CVI deficit are likely to be significantly impaired, but susceptible to training.

[0093] The study assesses explicit learning by comparing participant reaction time when discriminating the third piece of a given triplet vs all other discriminations. If the participant has gained explicit knowledge of the triplet, they are cued to the likely final presentation after viewing the first two-thirds of the triplet. This will result in reduced reaction time when responding to this final stimulus.

[0094] Performance for the end of a triplet are compared against uncorrelated pairs of motion in which the same final direction is utilized, to ensure differences in response time are not a result of response pattern or ease (i.e., responding faster for cartesian vs oblique directions). Implicit knowledge of the statistics behind stimulus presentation are estimated from the assessment task (FIG. 3C), represented by enhanced familiarity with High- Probability triplets vs Low or Null-Probability triplets. Statistical learning is known to be a fast process, with observable effects expected after a single training session. Transfer of improvements to visual perimetry are unlikely to arise as quickly but are detectable after three-months of home training.EXAMPLE 4: Early statistical learning outcomes

[0095] FIG. 5A is an example for learning task and stimulus. A high contrast drifting pattern was presented for 500ms moving in one of eight directions within a visual deficit (grey region), with participants indicating the perceived direction of motion with a joystick. Training consisted of one session of 384 trials. In FIG. 5B, stimuli were presented one at a time, grouped in triplets of high (HP) or low (LP) presentation probability. HP triplets were shown 25 times each, while LP were shown 7 times. Four triplets of each category were shown in total. Participants were not informed of groupings during training. A null set of triplets are never shown until a subsequent assessment task. FIG. 5C shows the results of a study in which eight participants with cortical blindness were trained on the learning task. Responses were considered correct if within 67.5° (i.e + or - one direction) of the presented direction. This allowed for response error in joystick usage and raised chance performance to 37.5%. Five participants performed meaningfully above this threshold (white dots), while three remained near chance (grey dots). For above-chance participants, performance wasbetter for HP trials than LP trials, indicating successful learning (p=0.009). FIG. 5D shows that reaction time also decreased over time for the five above-chance performers (r2=0.11, p<0.001), though with no difference between high and low-probability trials.

[0096] While various embodiments have been described above, it should be understood that such disclosures have been presented by way of example only and are not limiting. Thus, the breadth and scope of the subject compositions and methods should not be limited by any of the above-described exemplary embodiments but should be defined only in accordance with the following claims and their equivalents.

[0097] The above description is for the purpose of teaching the person of ordinary skill in the art how to practice the present invention, and it is not intended to detail all those obvious modifications and variations of it which will become apparent to the skilled worker upon reading the description. It is intended, however, that all such obvious modifications and variations be included within the scope of the present invention, which is defined by the following claims. The claims are intended to cover the components and steps in any sequence which is effective to meet the objectives there intended, unless the context specifically indicates the contrary.

Claims

WHAT IS CLAIMED IS:

1. A method for evaluating or improving a visual system of a subject, comprising: displaying multiple visual stimuli to the subject over a time period, wherein each visual stimulus comprises a motion signal; and receiving subsequent inputs from the subject indicating the subject's perception of a direction of motion of the motion signal in each visual stimulus and the subject’s reaction time for gaining the perception.

2. The method of claim 1, wherein the motion signal moves in one of eight directions.

3. The method of claim 1 or 2, wherein the multiple visual stimuli are grouped based on statistical learning principle.

4. The method of claim 3, wherein the multiple visual stimuli are displayed in groups of 3, 4, 5, 6, 7, 8 or 9 visual stimuli per group, with an interval time period between two adjacent visual stimuli and a separate time period between two adjacent groups.

5. The method of claim 4, wherein the interval time period is between 50-1000 ms.

6. The method of claim 5, wherein the interval time period is 200 ms.

7. The method of any one of claims 4-6, wherein the separate time period is between 100-2000 ms.

8. The method of any one of claims 4-6, wherein the separate time period is 500 ms.

9. The method of any one of claims 3-8, wherein some groups are displayed greater than 5 times (High) during a test, and some groups are displayed 2-5 times (Low) during the test.

10. The method of claim 9, wherein the test includes stimuli groups that are displayed only one time during the test (Null), wherein the stimuli groups are displayed sequentially in pairs ofHigh / Low, High / Null or Low / Null.

11. The method of and one of claims 1-10, wherein the motion signal is a high contrast signal.

12. The method of claim 11, wherein the motion signal has a 100% contrast.

13. The method of and one of claims 1-10, wherein the motion signal is a low contrast signal.

14. The method of claim 13, wherein the motion signal has a 50% contrast.

15. The method of any one of claims 1-14, wherein the motion signal has a size of 2- 20 deg diameter.

16. The method of claim 15, wherein the motion signal has a size of 5 deg diameter.

17. The method of any one of claims 1-16, wherein the motion signal moves at a speed in the range of 2-30 deg / s.

18. The method of claim 17, wherein the motion signal moves at a speed of 10 deg / s.

19. The method of any one of claims 1-18, wherein the motion signal is located within the subject’s vision.

20. The method of any one of claims 1-19, wherein input from the subject is received from a joystick, numeric keypad, touchpad, or a keyboard.

21. The method of any one of claims 1-20, further comprising the step of providing an evaluation of the visual system of the subject based on inputs received from the subject.

22. The method of any one of claims 1-21, wherein the subject is an adult with cortical blindness.

23. The method of any one of claims 1-21, wherein the subject is a pediatric patient with cortical visual impairment.

24. A non-transitory computer readable medium having stored thereon computer- readable instructions that, when executed by a processor, cause the processor to execute a method for evaluating or improving a visual system of a subject, the instructions comprising: instructions for displaying multiple visual stimuli on a display device over a time period, wherein each visual stimulus comprises a motion signal; and instructions for receiving subsequent inputs from the subject indicating the subject's perception of a direction of motion of the motion signal in each visual stimulus and the subject’s reaction time for gaining the perception.

Images