Cognitive function screening method, system, device and medium based on eye movement test

Abstract

This invention provides a method, system, device, and medium for cognitive function screening based on eye movement testing. The method involves acquiring patient data to be screened; constructing a cognitive function screening dataset based on the patient data; dividing the cognitive function screening dataset into a training set and a test set according to a preset ratio; inputting the training set into multiple preset classifiers for k-fold cross-validation training; integrating the multiple preset classifiers using a soft voting method to obtain a cognitive function screening model; and inputting the test set into the cognitive function screening model for testing to obtain screening results. This method enables rapid quantitative analysis based on eye movement data from multiple diseases and symptoms, achieving intelligent screening of cognitive impairment, improving the efficiency and accuracy of cognitive impairment screening, enhancing the diagnostic capabilities of doctors with insufficient experience, providing effective protection for early detection and treatment of diseases, and further reducing the burden on medical work and long-term social burden.

Description

Technical Field

[0001] This invention relates to the field of cognitive function screening technology for middle-aged and elderly people, and in particular to a cognitive function screening method, system, computer equipment, and storage medium based on eye movement testing. Background Art

[0002] Cognitive impairment refers to damage to one or more of the following functions: memory, language, visuospatial, executive, calculation, and comprehension / judgment. It can be classified into mild cognitive impairment and dementia based on the severity of the impairment. Mild cognitive impairment is an intermediate state between normal aging and dementia, where patients' daily activities are not significantly affected. Dementia, on the other hand, involves impairment in two or more cognitive domains, leading to a significant decline in daily or social abilities and imposing a considerable economic burden on the nation and society. However, traditional diagnostic methods for cognitive impairment include basic clinical symptoms, neuropsychological assessment, neuroimaging (CT), MRI, and SPECT, as well as laboratory tests (including complete blood count and blood biochemistry). These procedures are complex and cumbersome, and patients' awareness and cooperation with diagnosis are low, making early detection and timely targeted treatment and prevention of cognitive impairment a challenge.

[0003] Most existing intelligent screening methods for cognitive function assessment rely on the results of various diagnostic and treatment instruments and doctors' diagnostic experience. Although there are some studies on cognitive function screening based on machine learning, they are limited to learning from patients' physiological test data and Montreal Cognitive Assessment Scale data. Few scholars have considered the complex nonlinear relationships between various factors in cognitive function assessment and have combined various diagnostic and assessment data with eye movement data collected during cognitive testing to construct an eye movement dataset including patients with different cognitive impairments for machine screening analysis of cognitive function, in order to find new, more sensitive and convenient screening methods for cognitive impairment diseases. Summary of the Invention

[0004] The purpose of this invention is to provide a cognitive function screening method based on eye movement testing. By combining multimodal examination data with eye movement data collected from cognitive function testing to construct a cognitive function screening dataset, artificial intelligence technology is used to extract and analyze features from the cognitive function screening dataset to explore the association between abnormal eye movements and cognitive dysfunction diseases, provide cognitive function screening results, realize intelligent screening of cognitive dysfunction, improve the efficiency of cognitive dysfunction screening, enhance the diagnostic and treatment capabilities of doctors with insufficient clinical experience, and provide effective protection for the early detection and early treatment of diseases.

[0005] To achieve the above objectives, it is necessary to provide a cognitive function screening method, system, computer device, and storage medium based on eye-tracking testing to address the aforementioned technical problems.

[0006] In a first aspect, embodiments of the present invention provide a cognitive function screening method based on eye-tracking testing, the method comprising the following steps:

[0007] Acquire patient data to be screened; the patient data to be screened includes demographic characteristics, medical record data, Montreal Cognitive Assessment Scale results, and eye movement test data.

[0008] Based on the patient data to be screened, a cognitive function screening dataset is constructed, and the cognitive function screening dataset is divided into a training set and a test set according to a preset ratio;

[0009] The training set is input into multiple preset classifiers for k-fold cross-validation training, and the multiple preset classifiers are integrated by soft voting to obtain a cognitive function screening model.

[0010] The test set is input into the cognitive function screening model for testing, and the screening results are obtained.

[0011] Furthermore, the eye movement test data includes eye movement coordinate sequences from different patients during cognitive function tests.

[0012] Furthermore, the eye movement test data is obtained using an infrared eye-tracking test method.

[0013] Furthermore, the infrared eye-tracking testing method includes the following steps:

[0014] A near-infrared camera is pre-positioned between two near-infrared point light sources, and a display is positioned directly above the near-infrared camera.

[0015] When the patient begins the cognitive function test, the near-infrared point light source is lit up, and the near-infrared camera continuously captures human eye test images during the cognitive function test. The patient's left eye movement trajectory and right eye movement trajectory are continuously recorded on the display. The human eye test images are images that simultaneously include the left and right eyes.

[0016] Threshold segmentation is performed on the test images of each human eye to obtain the corresponding left eye pupil region, left eye corneal reflective point region, right eye pupil region, and right eye corneal reflective point region. Based on the left eye pupil region, left eye corneal reflective point region, right eye pupil region, and right eye corneal reflective point region, the corresponding center coordinates of the left eye pupil, left eye corneal reflective point, right eye pupil, and right eye corneal reflective point are obtained.

[0017] Based on the coordinates of the center of the left pupil and the center of the corneal reflection point of the left eye in the test images of each person's eyes, the corresponding left pupil corneal vector is obtained, and based on the coordinates of the center of the right pupil and the center of the corneal reflection point of the right eye, the corresponding right pupil corneal vector is obtained.

[0018] Based on the left eye pupil-corneal vector, the right eye pupil-corneal vector, and the corresponding left eye plane position and right eye plane position within the left eye movement trajectory from the test images of each person's eyes, the corresponding left eye movement coordinates and right eye movement coordinates are obtained through a preset position calibration method.

[0019] Based on the left and right eye movement coordinates of all human eye test images, the corresponding eye movement coordinate sequence is obtained.

[0020] Furthermore, the step of constructing a cognitive function screening dataset based on the patient data to be screened includes:

[0021] The patient data to be screened is subjected to quality screening according to preset standards to obtain preprocessed patient data.

[0022] Based on the demographic characteristics, medical records, and Montreal Cognitive Assessment Scale results of the preprocessed patient data, a baseline cognitive function assessment result was obtained.

[0023] The eye movement test data are labeled and classified using the benchmark cognitive function assessment results to obtain the cognitive function screening dataset.

[0024] Furthermore, the step of performing quality screening on the patient data to be screened according to preset standards to obtain preprocessed patient data includes:

[0025] Determine whether there are missing items in the corresponding data of each patient in the patient data to be screened. If there are, determine whether the proportion of all missing items of the corresponding patient exceeds a preset missing threshold.

[0026] If the proportion of all missing items for a patient exceeds the preset missing threshold, all data for that patient will be deleted. Otherwise, each missing item will be filled in according to the preset rules.

[0027] Furthermore, the preset classifier includes a random forest model, a gradient boosting machine model, an XGBoost model, and a support vector machine model;

[0028] The steps of inputting the training set into multiple preset classifiers for k-fold cross-validation training, and integrating the multiple preset classifiers using a soft voting method to obtain a cognitive function screening model include:

[0029] The training set is input into the random forest model for training to obtain the first prediction model;

[0030] The training set is input into the gradient booster model for training to obtain the second prediction model;

[0031] The training set is input into the XGBoost model for training to obtain the third prediction model;

[0032] The training set is input into the support vector machine model for training to obtain the fourth prediction model;

[0033] The first prediction model, the second prediction model, the third prediction model, and the fourth prediction model are integrated using the soft voting method to obtain the cognitive function screening model.

[0034] Secondly, embodiments of the present invention provide a cognitive function screening system based on eye-tracking testing, the system comprising:

[0035] The data acquisition module is used to acquire data of patients to be screened; the data of patients to be screened includes demographic characteristics, medical record data, Montreal Cognitive Assessment Scale results, and eye movement test data;

[0036] The preprocessing module is used to construct a cognitive function screening dataset based on the patient data to be screened, and to divide the cognitive function screening dataset into a training set and a test set according to a preset ratio;

[0037] The model building module is used to input the training set into multiple preset classifiers for k-fold cross-validation training, and to integrate the multiple preset classifiers through soft voting to obtain a cognitive function screening model.

[0038] The result generation module is used to input the test set into the cognitive function screening model for testing and obtain the screening results.

[0039] Thirdly, embodiments of the present invention also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0040] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0041] This application provides a cognitive function screening method, system, computer device, and storage medium based on eye-tracking testing. The method acquires patient data to be screened, including demographic characteristics, medical records, Montreal Cognitive Assessment Scale results, and eye movement test data. Based on the preprocessed patient data, a cognitive function screening dataset is constructed. This dataset is then divided into training and testing sets according to a preset ratio. The training set is input into multiple preset classifiers for k-fold cross-validation training. The multiple preset classifiers are then integrated using a soft voting method to obtain a cognitive function screening model. The testing set is then input into the cognitive function screening model for testing to obtain the screening results. Compared with existing technologies, this eye-tracking testing-based cognitive function screening method combines multimodal examination data with eye movement data collected from cognitive function tests to construct a cognitive function screening dataset. Rapid quantitative analysis based on eye movement data enables intelligent screening of cognitive impairment, improving screening efficiency, enhancing the diagnostic capabilities of doctors with insufficient experience, providing effective guarantees for early detection and treatment of diseases, and further reducing the burden on medical work and long-term social burden. Attached Figure Description

[0042] Figure 1 This is a schematic diagram illustrating the application scenario of the cognitive function screening method based on eye movement testing in this embodiment of the invention;

[0043] Figure 2 This is a flowchart illustrating the cognitive function screening method based on eye movement testing in an embodiment of the present invention.

[0044] Figure 3 This is a schematic diagram of the structure of the cognitive function screening system based on eye movement testing in an embodiment of the present invention;

[0045] Figure 4 This is an internal structural diagram of the computer device in an embodiment of the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and beneficial effects of this application clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Obviously, the embodiments described below are only part of the embodiments of the present invention and are used to illustrate the present invention, but are not intended to limit the scope of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0047] The cognitive function screening method based on eye movement testing provided by this invention can be applied to, for example... Figure 1The terminal or server shown. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be a standalone server or a server cluster consisting of multiple servers. The server can screen patient data based on demographic characteristics, cranial MRI, disease symptoms, diagnostic test data, Montreal Cognitive Assessment Scale results, and eye movement test data obtained from patients with different cognitive functions. Based on this, it constructs a cognitive function screening dataset based on eye movement data. The cognitive function screening method based on eye movement tests provided in this invention is used to complete the cognitive function screening of different patients. The final screening results are then applied to other learning tasks on the server or transmitted to the terminal for user reception.

[0048] In one embodiment, such as Figure 2 As shown, a cognitive function screening method based on eye-tracking testing is provided, including the following steps:

[0049] S11. Obtain patient data to be screened; the patient data to be screened includes demographic characteristics, medical record data, Montreal Cognitive Assessment Scale results, and eye movement test data.

[0050] The medical record data includes cranial magnetic resonance imaging, disease symptoms, and diagnostic test data. The demographic characteristics, medical record data, Montreal Cognitive Assessment Scale results, and medical record data are mainly used to label different cognitive assessment results of eye movement test data. All of these can be obtained through existing methods, and no specific restrictions are imposed here. It should be noted that the specific diseases involved in the patients to be tested include cerebrovascular diseases, neurodegenerative diseases, radiation encephalopathy, peripheral neuropathy, and tumors, which may lead to cognitive impairment. The corresponding patient data to be screened can be understood as including relevant data of patients to be screened with multiple diseases at the same time, ensuring the comprehensiveness of the analysis data and thus effectively improving the accuracy of subsequent machine learning.

[0051] Eye movement test data includes eye movement coordinate sequences of different patients during cognitive function tests, which can be obtained in principle through different infrared eye movement testing methods. To ensure the accuracy of the eye movement test data, and thus the precision of subsequent cognitive function screening results, this embodiment preferably uses the following infrared eye movement testing method to obtain the corresponding eye movement test data for different patients, specifically including the following steps:

[0052] A near-infrared camera is pre-positioned between two near-infrared point light sources, and a display is positioned directly above the near-infrared camera. The near-infrared camera is used to capture images including the human eye, and the two LED near-infrared point light sources provide illumination for the capture, forming a reflected virtual image on the cornea of ​​the human eye, referred to as a "corneal reflection point." In this embodiment, the two near-infrared light sources are located on both sides of the near-infrared camera. To improve the clarity of the captured images and the stability of the data, the patient's head position is fixed during testing, with their eyes facing the display.

[0053] When the patient begins the cognitive function test, the near-infrared point light source is illuminated, and the near-infrared camera continuously captures images of the patient's eyes during the cognitive function test. The patient's left and right eye movement trajectories are continuously recorded on the display. The human eye test images are images that simultaneously include both the left and right eyes. The human eye test images are a set of images that can reflect the continuous changes in the human eye, and different images have corresponding left and right eye movement trajectories recorded on the display. This allows for real-time analysis and calculation of the eye movement coordinates by simultaneously using the human eye test images and the left and right eye movement trajectories.

[0054] Threshold segmentation was performed on each human eye test image to obtain the corresponding left eye pupil region, left eye corneal reflective point region, right eye pupil region, and right eye corneal reflective point region. Based on these regions, the corresponding coordinates of the left eye pupil center, left eye corneal reflective point center, right eye pupil center, and right eye corneal reflective point center were obtained. Notably, in the human eye test images, the pupil grayscale is low (below 50), while the corneal reflective point grayscale is high (above 200). Based on the grayscale difference between the pupil and the corneal reflective point, threshold segmentation is performed on the human eye test image to extract the regions where the left eye pupil, the left eye corneal reflective point, the right eye pupil, and the right eye corneal reflective point are located. Based on the extracted regions, the corresponding coordinates of the center of the left eye pupil, the center of the left eye corneal reflective point, the center of the right eye pupil, and the center of the right eye corneal reflective point are obtained. The center coordinates of the left eye corneal reflective point are the average of the coordinates of the two corneal reflective points in the left eye, and the center coordinates of the right eye corneal reflective point are the average of the coordinates of the two corneal reflective points in the right eye.

[0055] Based on the coordinates of the center of the left pupil and the center of the left corneal reflective point in the test images of each person's eyes, the corresponding left pupil corneal vector is obtained, and based on the coordinates of the center of the right pupil and the center of the right corneal reflective point, the corresponding right pupil corneal vector is obtained; where the left pupil corneal vector is a vector starting from the center of the left corneal reflective point and ending at the center of the left pupil; the right pupil corneal vector is a vector starting from the center of the right corneal reflective point and ending at the center of the right pupil.

[0056] Based on the left and right pupil-corneal vectors of each human eye test image, and the corresponding planar positions of the left and right eyeballs within the left and right eye movement trajectories, the corresponding left and right eye movement coordinates are obtained through a preset position calibration method. Specifically, there is a mapping relationship between the left pupil-corneal vector of the human eye test image and the left eyeball planar position within the left eye movement trajectory on the display, and between the right pupil-corneal vector of the human eye test image and the right eye planar position within the right eye movement trajectory on the display. The corresponding mapping function can be obtained through multiple position calibration points. In this embodiment, preferably, nine preset positions on the display—center, left, right, top, bottom, upper left, upper right, lower left, and lower right—are calibrated. The left eye's pupil-corneal vector and the corresponding coordinates for each preset position are substituted into a system of mapping function equations to solve for the left eye calibration mapping function coefficients, thus obtaining the left eye calibration function. Similarly, the right eye's pupil-corneal vector and the corresponding coordinates for each preset position are substituted into the system of mapping function equations to solve for the right eye calibration mapping function coefficients, thus obtaining the right eye calibration function. The specific process of obtaining the left and right eye movement coordinates is illustrated in the following example:

[0057] Let x s and y s These are the x and y coordinates, respectively, of the left eye's movement coordinate system on the monitor screen, in cm; x e and y e These represent the horizontal and vertical components of the left eye pupil-corneal vector, respectively, in pixels. The corresponding mapping function is:

[0058]

[0059] Among them, a0, a1, a2, a3, a4, a5, b0, b1, b2, b3, b4, and b5 are the coefficients of the mapping function, and these 12 values ​​are all unknown before calibration. The calibration process is the process of solving these 12 unknowns. The coordinates (x, y, y) of the aforementioned 9 calibration points (preset positions) on the display plane are... s1 ,y s1 ), (x s2 ,y s2 ), (x s3 ,y s3 ), (x s4 ,y s4 ), (x s5 ,y s5 ), (x s6 ,y s6 ), (x s7 ,y s7 ), (x s8 ,y s8 ) and (xs9 ,y s9 The following is known. When the eye looks at these 9 calibration points (preset positions), the corneal vector of the left eye pupil can be calculated according to the previous steps (x...). e1 ,y e1 ), (x e2 ,y e2 ), (x e3 ,y e3 ), (x e4 ,y e4 ), (x e5 ,y e5 ), (x e6 ,y e6 ), (x e7 ,y e7 ), (x e8 ,y e8 ) and (x e9 ,y e9 Substituting the values ​​into the mapping function, we obtain the following system of 18 equations:

[0060]

[0061]

[0062] The above system of equations is an overdetermined system with more equations than unknown variables. Solving it using the least squares method yields the mapping function coefficients a0, a1, a2, a3, a4, a5, b0, b1, b2, b3, b4, and b5, thus obtaining the left eye calibration function. Similarly, the right eye calibration function can be obtained, completing the entire calibration process. These functions are used to substitute the left and right pupil-corneal vectors corresponding to the real-time acquired human eye test images during eye movement testing into the corresponding calibration functions, allowing for the real-time calculation of the left and right eye movement coordinates on the display plane.

[0063] Based on the left and right eye movement coordinates of all eye test images, the corresponding eye movement coordinate sequence is obtained. Specifically, the series of left and right eye movement coordinates obtained from the corresponding eye test images during each patient's cognitive test can be directly used as the corresponding eye movement coordinate sequence for subsequent machine learning. However, to ensure the comprehensiveness and accuracy of the cognitive assessment results, the eye movement coordinate sequences corresponding to the patient's left and right eyes are merged by averaging to obtain the final eye movement coordinate sequence (binocular eye movement coordinate sequence).

[0064] S12. Based on the patient data to be screened, construct a cognitive function screening dataset, and divide the cognitive function screening dataset into a training set and a test set according to a preset ratio;

[0065] The cognitive function screening dataset can be understood as the dataset constructed based on the data of the patients to be screened, including the following steps:

[0066] The patient data to be screened is subjected to quality screening according to preset standards to obtain preprocessed patient data; wherein, the step of performing quality screening on the patient data to be screened according to preset standards to obtain preprocessed patient data includes:

[0067] Determine whether there are missing items in the corresponding data of each patient in the patient data to be screened. If there are, determine whether the proportion of all missing items of the corresponding patient exceeds a preset missing threshold.

[0068] If the percentage of missing items for a patient exceeds a preset threshold, all data for that patient is deleted. Otherwise, missing items are filled according to preset rules. For example, if the percentage of missing items exceeds 20%, all data for that patient is deleted automatically. If the percentage is less than 20%, different methods are used to fill in the missing values ​​depending on the type of missing data: the mode is used to fill in the missing discrete data, and the median is used to fill in the missing continuous data.

[0069] Based on the demographic characteristics, medical records, and Montreal Cognitive Assessment Scale results of the preprocessed patient data, a baseline cognitive function assessment result is obtained. This baseline cognitive function assessment result is a relatively professional cognitive function assessment that integrates information such as demographic characteristics, medical records, and Montreal Cognitive Assessment Scale results. For example, it can be obtained by multiple neurologists based on the Montreal Cognitive Assessment Scale results, demographic characteristics, and medical records.

[0070] The eye movement test data are labeled and classified using the benchmark cognitive function assessment results to obtain the cognitive function screening dataset. Before labeling and classification, the eye movement test data also needs to undergo quality screening. This can be done by establishing eye movement data quality standards through ophthalmologists or other means, uniformly screening the collected eye movement test data, and then matching it with the obtained benchmark cognitive function assessment results to construct a cognitive function screening dataset that includes eye movement test data and corresponding labels, for subsequent cognitive function screening classification learning.

[0071] S13. Input the training set into multiple preset classifiers for k-fold cross-validation training, and integrate the multiple preset classifiers through soft voting to obtain a cognitive function screening model.

[0072] The number and type of preset classifiers can be selected according to actual application needs. To address the problem of single-model learning and training easily getting trapped in local optima, resulting in poor generalization ability, this embodiment selects four machine learning algorithms—random forest, gradient boosting machine, XGBoost, and support vector machine—as base classifiers. This ensures effective screening of cognitive function while further improving the classification accuracy and generalization ability of the cognitive function screening model. Specifically, the steps of inputting the training set into multiple preset classifiers for k-fold cross-validation training and integrating the multiple preset classifiers using a soft voting method to obtain the cognitive function screening model include:

[0073] The training set is input into the random forest model for training to obtain the first prediction model;

[0074] The training set is input into the gradient booster model for training to obtain the second prediction model;

[0075] The training set is input into the XGBoost model for training to obtain the third prediction model;

[0076] The training set is input into the support vector machine model for training to obtain the fourth prediction model;

[0077] The first, second, third, and fourth prediction models are integrated using the soft voting method to obtain the cognitive function screening model. The soft voting method involves assigning corresponding weights to the first, second, third, fourth, and fifth prediction models and integrating them to obtain the cognitive function screening model. Specifically, the final screening result of this cognitive function screening model is the one with the highest probability among the weighted average of the various cognitive function assessment results predicted by each predictor classifier.

[0078] S14. Input the test set into the cognitive function screening model for testing to obtain the screening results.

[0079] This application embodiment collects eye movement test data from patients with different diseases and symptoms using near-infrared eye movement testing. Based on the baseline cognitive function assessment results obtained through a comprehensive evaluation of corresponding demographic characteristics, medical records, and the Montreal Cognitive Assessment Scale, the data is labeled and classified to construct a cognitive function screening dataset. Multiple preset classifiers are used to extract and analyze features from this dataset. A cognitive function screening model for intelligent screening of cognitive function is obtained through k-fold cross-validation training. This model reasonably and effectively correlates abnormal eye movements with cognitive impairment diseases, thereby providing accurate and effective cognitive function screening results. Based on intelligent screening of cognitive impairment, this not only improves the efficiency of cognitive impairment screening but also effectively enhances its accuracy, thereby improving the diagnostic capabilities of doctors with insufficient experience. It also provides effective protection for the early detection and treatment of cognitive diseases, further reducing the burden on medical work and long-term social burden.

[0080] It should be noted that although the steps in the flowchart above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order requirement for the execution of these steps, and they can be executed in other orders.

[0081] In one embodiment, such as Figure 3 As shown, a cognitive function screening system based on eye movement testing is provided, the system comprising:

[0082] Data acquisition module 1 is used to acquire data of patients to be screened; the data of patients to be screened includes demographic characteristics, medical record data, Montreal Cognitive Assessment Scale assessment results and eye movement test data;

[0083] Preprocessing module 2 is used to construct a cognitive function screening dataset based on the patient data to be screened, and to divide the cognitive function screening dataset into a training set and a test set according to a preset ratio;

[0084] Model building module 3 is used to input the training set into multiple preset classifiers for k-fold cross-validation training, and integrate the multiple preset classifiers through soft voting to obtain a cognitive function screening model;

[0085] Result generation module 4 is used to input the test set into the cognitive function screening model for testing and obtain screening results.

[0086] For specific limitations regarding an eye-tracking-based cognitive function screening system, please refer to the limitations of an eye-tracking-based cognitive function screening method described above, which will not be repeated here. Each module in the aforementioned eye-tracking-based cognitive function screening system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0087] Figure 4 An internal structural diagram of a computer device is shown in one embodiment. This computer device may specifically be a terminal or a server. Figure 4 As shown, the computer device includes a processor, memory, network interface, display, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a cognitive function screening method based on eye-tracking testing. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0088] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computing devices may include more or fewer components than those shown in the figure, or combine certain components, or have the same component arrangement.

[0089] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method described above.

[0090] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0091] In summary, the present invention provides a cognitive function screening method, system, computer device, and storage medium based on eye-tracking testing. The method acquires multi-disease, multi-symptom patient data for screening, including demographic characteristics, medical record data, Montreal Cognitive Assessment Scale results, and eye movement test data. Based on the preprocessed patient data, a cognitive function screening dataset is constructed. This dataset is then divided into training and testing sets according to a preset ratio. The training set is input into multiple preset classifiers, including random forest, gradient boosting, XGBoost, and support vector machine models, for k-fold cross-validation training. Finally, the multiple preset classifiers are integrated using a soft voting method. This technical solution involves obtaining a cognitive function screening model, inputting a test set into the model for testing, and obtaining screening results. It constructs a cognitive function screening dataset by combining multimodal examination data with eye movement data collected from cognitive function tests. Artificial intelligence technology is then used to extract and analyze features from this dataset, exploring the correlation between abnormal eye movements and cognitive impairment diseases. This provides cognitive function screening results, achieving intelligent screening for cognitive impairment. This not only improves the efficiency of cognitive impairment screening but also effectively enhances its accuracy, thereby improving the diagnostic capabilities of doctors with insufficient experience. This provides effective protection for early detection and treatment of diseases, further reducing the burden on medical work and long-term social burden.

[0092] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0093] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the technical principles of this invention, and these improvements and substitutions should also be considered within the scope of protection of this application. Therefore, the scope of protection of this patent application should be determined by the scope of the claims.

Claims

1. A cognitive function screening method based on eye-tracking testing, characterized in that, The method includes the following steps: Acquire patient data to be screened; the patient data to be screened includes demographic characteristics, medical record data, Montreal Cognitive Assessment Scale (MCA) results, and eye movement test data; the demographic characteristics, medical record data, and MCA results are used to label the eye movement test data with different cognitive assessment results; the eye movement test data includes eye movement coordinate sequences of different patients during cognitive function tests; Based on the patient data to be screened, a cognitive function screening dataset is constructed, and the cognitive function screening dataset is divided into a training set and a test set according to a preset ratio; the cognitive function screening dataset includes eye movement test data and corresponding classification labels from the patient data to be screened. The training set is input into multiple preset classifiers for k-fold cross-validation training, and the multiple preset classifiers are integrated by soft voting to obtain a cognitive function screening model. The test set is input into the cognitive function screening model for testing to obtain the screening results; The eye movement test data is obtained through an infrared eye-tracking test method, which includes: Based on the left and right pupil-corneal vectors and their corresponding planar positions within the left and right eye movement trajectories from the acquired eye test images, the corresponding left and right eye movement coordinates are obtained using a preset position calibration method. The eye test images are obtained by continuously capturing images of the left and right eyes during the cognitive function test using a near-infrared camera at the start of the test. The preset position calibration method is based on a mapping function between the pupil-corneal vectors of the eye test images and the planar positions of the eye movement trajectories. This mapping function is expressed as: Where, x s and y s These represent the x and y coordinates in the eye movement coordinate system, respectively; x e and y e ... Based on the left and right eye movement coordinates of all human eye test images, the corresponding eye movement coordinate sequence is obtained; each eye movement coordinate in the eye movement coordinate sequence is the average value of the left and right eye movement coordinates at the corresponding time.

2. The cognitive function screening method based on eye-tracking testing as described in claim 1, characterized in that, The infrared eye-tracking testing method also includes the following steps: A near-infrared camera is pre-positioned between two near-infrared point light sources, and a display is positioned directly above the near-infrared camera. When the patient begins the cognitive function test, the near-infrared point light source is lit up, and the near-infrared camera continuously captures human eye test images during the cognitive function test. The patient's left eye movement trajectory and right eye movement trajectory are continuously recorded on the display. The human eye test images are images that simultaneously include the left and right eyes. Threshold segmentation is performed on the test images of each human eye to obtain the corresponding left eye pupil region, left eye corneal reflective point region, right eye pupil region, and right eye corneal reflective point region. Based on the left eye pupil region, left eye corneal reflective point region, right eye pupil region, and right eye corneal reflective point region, the corresponding center coordinates of the left eye pupil, left eye corneal reflective point, right eye pupil, and right eye corneal reflective point are obtained. Based on the coordinates of the center of the left pupil and the center of the corneal reflection point of the left eye in the test images of each person's eyes, the corresponding left pupil corneal vector is obtained, and based on the coordinates of the center of the right pupil and the center of the corneal reflection point of the right eye, the corresponding right pupil corneal vector is obtained.

3. The cognitive function screening method based on eye movement testing as described in claim 1, characterized in that, The step of constructing a cognitive function screening dataset based on the patient data to be screened includes: The patient data to be screened is subjected to quality screening according to preset standards to obtain preprocessed patient data. Based on the demographic characteristics, medical records, and Montreal Cognitive Assessment Scale results of the preprocessed patient data, a baseline cognitive function assessment result was obtained. The eye movement test data are labeled and classified using the benchmark cognitive function assessment results to obtain the cognitive function screening dataset.

4. The cognitive function screening method based on eye movement testing as described in claim 3, characterized in that, The step of performing quality screening on the patient data to be screened according to preset standards to obtain preprocessed patient data includes: Determine whether there are missing items in the corresponding data of each patient in the patient data to be screened. If there are, determine whether the proportion of all missing items of the corresponding patient exceeds a preset missing threshold. If the proportion of all missing items for a patient exceeds the preset missing threshold, all data for that patient will be deleted. Otherwise, each missing item will be filled in according to the preset rules.

5. The cognitive function screening method based on eye movement testing as described in claim 1, characterized in that, The preset classifiers include random forest model, gradient boosting machine model, XGBoost model and support vector machine model; The steps of inputting the training set into multiple preset classifiers for k-fold cross-validation training, and integrating the multiple preset classifiers using a soft voting method to obtain a cognitive function screening model include: The training set is input into the random forest model for training to obtain the first prediction model; The training set is input into the gradient booster model for training to obtain the second prediction model; The training set is input into the XGBoost model for training to obtain the third prediction model; The training set is input into the support vector machine model for training to obtain the fourth prediction model; The first prediction model, the second prediction model, the third prediction model, and the fourth prediction model are integrated using the soft voting method to obtain the cognitive function screening model.

6. A cognitive function screening system based on eye-tracking testing, characterized in that, The system employs the cognitive function screening method based on eye-tracking testing as described in claim 1, wherein the system comprises: The data acquisition module is used to acquire data of patients to be screened; the data of patients to be screened includes demographic characteristics, medical record data, Montreal Cognitive Assessment Scale results, and eye movement test data; The preprocessing module is used to construct a cognitive function screening dataset based on the patient data to be screened, and to divide the cognitive function screening dataset into a training set and a test set according to a preset ratio; The model building module is used to input the training set into multiple preset classifiers for k-fold cross-validation training, and to integrate the multiple preset classifiers through soft voting to obtain a cognitive function screening model. The result generation module is used to input the test set into the cognitive function screening model for testing and obtain the screening results.

7. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images