Disease risk prediction method and device, equipment and storage medium

Abstract

The invention provides a disease risk prediction method and device, equipment and a storage medium, and the method comprises the steps: guiding a subject to complete each eye test through virtual reality equipment, collecting the eye activity data needed by the prediction of the disease risk in the test process, and carrying out the prediction of the disease risk; eye movement characteristics of a subject can be objectively, quantitatively and comprehensively recorded, and the accuracy of a disease risk prediction data source is ensured; on the other hand, through parameter values of various preset eye movement characteristic parameters extracted from the eye movement data and additional basic physical examination information of the subject, a disease risk prediction result of each neurodegenerative disease corresponding to the subject is output by means of a multi-task learning model; therefore, the disease risks of various different neurodegenerative diseases corresponding to the subject can be integrally and comprehensively predicted from multiple angles, and the accuracy of various disease risk prediction results is effectively improved.

Description

Technical Field

[0001] The present application relates to the technical field of disease risk warning, and specifically, to a disease risk prediction method, device, equipment and storage medium. Background Art

[0002] Neurodegenerative diseases such as Alzheimer's disease and Parkinson's disease are characterized by a long course of disease and mild early symptoms (i.e., usually no symptoms that seriously affect daily life). However, slight changes may occur in the eye, the organ closest to the brain. Therefore, for neurodegenerative diseases, the risk of the target subject suffering from a neurodegenerative disease can be predicted by detecting the subject's eye activity.

[0003] At present, the existing traditional methods mainly use relevant testing equipment to directly conduct various eye tests on subjects, and record the eye activities of subjects in each eye test based on paper-pencil scales or subjective observations, so as to determine the specific neurodegenerative disease associated with each single eye test with abnormal test results as the specific prediction result of the subject's disease risk based on the test results of each single eye test. However, the manifestation changes caused by different types of neurodegenerative diseases are often comprehensive and intrinsically related. Therefore, the above method of predicting disease risk only based on the test results of each single eye test fails to comprehensively predict the subject's disease risk from multiple angles, resulting in inaccurate disease risk prediction results. Summary of the invention

[0004] In view of this, the purpose of the present application is to provide a disease risk prediction method, device, equipment and storage medium. On the one hand, the subject is guided to complete each eye test through a virtual reality device, and the eye activity data required for disease risk prediction is collected during the test, which is conducive to objectively and quantitatively recording the subject's eye movement characteristics and ensuring the accuracy of the source of disease risk prediction data; on the other hand, by extracting the parameter values ​​of a variety of preset eye movement feature parameters from the above-mentioned eye activity data and the additional basic physical examination information of the subject, the disease risk prediction results of each neurodegenerative disease corresponding to the subject are output with the help of a multi-task learning model, so that the disease risks of a variety of different neurodegenerative diseases corresponding to the subject can be comprehensively predicted from multiple angles, effectively improving the accuracy of various disease risk prediction results.

[0005] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings.

[0006] In a first aspect, an embodiment of the present application provides a disease risk prediction method, the disease risk prediction method comprising:

[0007] In the process of collecting the eye movement data of the subject through the virtual reality device, visual guidance information for guiding the subject to perform each eye movement test is displayed on the graphical user interface of the virtual reality device, and the test result of each eye movement test is used as the collected eye movement data;

[0008] According to the parameter definition of each preset eye movement feature parameter, extracting the parameter value of each preset eye movement feature parameter from the eye activity data;

[0009] The parameter value of each of the preset eye movement feature parameters and the basic physical examination information of the subject are input into a pre-trained multi-task learning model, and the disease risk prediction result of each neurodegenerative disease corresponding to the subject is output through each sub-task learning model included in the multi-task learning model; wherein each of the sub-task learning models is used to predict the disease risk of a neurodegenerative disease corresponding to the subject.

[0010] In a second aspect, an embodiment of the present application provides a disease risk prediction device, the disease risk prediction device comprising:

[0011] A visual guidance module, used for displaying visual guidance information for guiding the subject to perform each eye movement test on a graphical user interface of the virtual reality device during the process of collecting the subject's eye movement data through the virtual reality device, and using the test result of each eye movement test as the collected eye movement data;

[0012] A data acquisition module, configured to extract a parameter value of each preset eye movement feature parameter from the eye activity data according to a parameter definition of each preset eye movement feature parameter;

[0013] A risk prediction module is used to input the parameter value of each of the preset eye movement feature parameters and the basic physical examination information of the subject into a pre-trained multi-task learning model, and output the disease risk prediction result of each neurodegenerative disease corresponding to the subject through each sub-task learning model included in the multi-task learning model; wherein each of the sub-task learning models is used to predict the disease risk of a neurodegenerative disease corresponding to the subject.

[0014] In a third aspect, an embodiment of the present application provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned disease risk prediction method when executing the computer program.

[0015] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned disease risk prediction method are executed.

[0016] The technical solution provided by the embodiments of the present application may have the following beneficial effects:

[0017] The embodiments of the present application provide a disease risk prediction method, apparatus, device and storage medium. On the one hand, a subject is guided to complete each eye test through a virtual reality device, and eye activity data required for disease risk prediction is collected during the test, which is conducive to objectively and quantitatively recording the subject's eye movement characteristics and ensuring the accuracy of the source of disease risk prediction data. On the other hand, the parameter values ​​of multiple preset eye movement feature parameters extracted from the above-mentioned eye activity data and the additional basic physical examination information of the subject are used to output the disease risk prediction results of each neurodegenerative disease corresponding to the subject with the help of a multi-task learning model, so that the disease risks of multiple different neurodegenerative diseases corresponding to the subject can be comprehensively predicted from multiple angles, effectively improving the accuracy of various disease risk prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A schematic diagram of a process for predicting disease risk provided in an embodiment of the present application is shown;

[0020] Figure 2a A schematic flow chart of a method for displaying visual guidance information for guiding a subject to perform a pupil light stimulation response test on a graphical user interface provided by an embodiment of the present application is shown;

[0021] Figure 2b A data schematic diagram of eye activity data that can be collected during a pupil light stimulation response test provided by an embodiment of the present application is shown;

[0022] Figure 3 A flow chart of a method for displaying visual guidance information for guiding a subject to perform a forward saccade test on a graphical user interface provided by an embodiment of the present application is shown;

[0023] Figure 4A flow chart of a method for displaying visual guidance information for guiding a subject to perform a reverse saccade test on a graphical user interface provided by an embodiment of the present application is shown;

[0024] Figure 5 A flow chart of a method for displaying visual guidance information for guiding a subject to perform a counting task test on a graphical user interface provided by an embodiment of the present application is shown;

[0025] Figure 6 A flow chart of a method for performing model training on each of the subtask learning models included in a multi-task learning model provided by an embodiment of the present application is shown;

[0026] Figure 7 A schematic diagram of the structure of a disease risk prediction device provided in an embodiment of the present application is shown;

[0027] Figure 8 A schematic diagram of the structure of an electronic device 800 provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of explanation and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn in real proportion. The flowchart used in this application shows the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowchart can be implemented out of sequence, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart under the guidance of the content of the present application, or remove one or more operations from the flowchart.

[0029] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0030] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0031] Here, the disease risk prediction method provided in the embodiment of the present application can be applied to a target electronic device, wherein a multi-task learning model for disease risk prediction is deployed on the target electronic device, and data communication can be performed between the target electronic device and the virtual reality device, so that the target electronic device can send relevant control instructions to the virtual reality device to control the virtual reality device to guide the subject to complete each eye test, and collect eye activity data required for disease risk prediction during the test.

[0032] It should be noted that the above-mentioned target electronic device can be a terminal device or a server; wherein the above-mentioned terminal device can be a fixed terminal device (such as a personal computer) or a mobile terminal device (such as a mobile phone, a tablet computer); the above-mentioned server can be an independent physical server, or it can be a server cluster or a distributed system, a cloud server, etc. composed of multiple physical servers; the embodiments of the present application do not impose any limitation on the specific device type to which the above-mentioned target electronic device belongs.

[0033] To facilitate understanding of the embodiments of the present application, a disease risk prediction method, device, equipment and storage medium provided in the embodiments of the present application are introduced in detail below.

[0034] Reference Figure 1 As shown, Figure 1 A flow chart of a disease risk prediction method provided in an embodiment of the present application is shown, wherein the disease risk prediction method includes steps S101-S103; specifically:

[0035] S101, in the process of collecting eye movement data of a subject through a virtual reality device, visual guidance information for guiding the subject to perform each eye movement test is displayed on a graphical user interface of the virtual reality device, and the test result of each eye movement test is used as the collected eye movement data.

[0036] Here, the above-mentioned virtual reality device can be a wearable device worn by the subject; wherein, the virtual reality device only needs to ensure that it can at least display the visual guidance information corresponding to each of the above-mentioned eye movement tests in a VR (Virtual Reality) manner, and is equipped with an infrared video module to collect the eye activity data of the subject; for example, the above-mentioned virtual reality device can specifically be a VR helmet equipped with an eye movement capture module, or it can be VR glasses with eye movement capture function, etc.; the specific device type and device model of the above-mentioned virtual reality device are not limited in any way in the embodiments of the present application.

[0037] Specifically, in the embodiment of the present application, the virtual reality device can guide the subject to perform an eye movement test; wherein the eye movement test includes but is not limited to at least one of the following:

[0038] (1) Pupil light response test: The pupil light response test is mainly used to test the various activity characteristics of the pupil when the subject is stimulated by strong light.

[0039] (2) Forward glance test; wherein the forward glance test is mainly used to test the eye function of the subject when the subject performs eye movement passively; for example, the forward glance test can be used to test the change in the subject's eye gaze direction from the center position of the above-mentioned graphical user interface to the surroundings (e.g., specific positions in four different directions, such as up, down, left, and right, whose distance from the center position on the graphical user interface is a preset distance threshold).

[0040] (3) Reverse saccade test; wherein the reverse saccade test is mainly used to test the eye function status reflected by the subject when actively performing eye movements; for example, the reverse saccade test can be used to test the change in the subject's eye gaze direction from the center position of the above-mentioned graphical user interface to the surroundings (such as specific position points in four different directions such as up, down, left and right on the graphical user interface whose distance from the center position is a preset distance threshold) in the opposite direction.

[0041] (4) Counting task test; the counting task test is mainly used to test the eye function of the subjects when they are exposed to various visual stimuli (i.e., the effects of various visual stimuli on the eye movements of the subjects); for example, the counting task test can test the changes in the eye gaze position and gaze duration of the subjects when they observe multiple target balls that appear randomly (i.e., the positions and numbers of the balls appear randomly) on the above-mentioned graphical user interface.

[0042] It should be noted that, in actual applications, the specific type of eye movement test that the subject needs to be guided to perform can be selected according to actual test requirements; the embodiment of the present application does not impose any mandatory restrictions on the specific type of eye movement test that can be performed in step S101.

[0043] S102, extracting a parameter value of each preset eye movement feature parameter from the eye activity data according to a parameter definition of each preset eye movement feature parameter.

[0044] Here, different types of eye movement tests are associated with different preset eye movement feature parameters; among them, the eye movement tests performed in step S101 include the pupil light stimulation response test, the forward saccade test, the reverse saccade test, and the counting task test as examples, which are described in detail as follows:

[0045] Specifically, for the above-mentioned pupil light stimulation response test, the preset eye movement feature parameters associated with the pupil light stimulation response test may be a plurality of eye movement feature parameters related to the pupil changes of the subject; for example, in the pupil light stimulation response test, the average pupil size in the absence of light stimulation (that is, the baseline pupil size, which can be calculated based on the collected pupil diameter), the pupil contraction latency when stimulated by strong light (that is, the time interval from the appearance of strong light to the contraction of the subject's pupil), the maximum pupil contraction speed, the maximum contraction value, the specific time required for the pupil to return to stability, the pupil size after stabilization, the difference between the stabilized pupil size and the above-mentioned baseline pupil size, etc.

[0046] Specifically, for the above-mentioned forward scanning test, the preset eye movement feature parameters associated with the forward scanning test can be a plurality of eye movement feature parameters related to the subject's eye movement; for example, in the forward scanning test, the subject's line of sight (i.e., the direction of eye gaze) scans from the center position of the above-mentioned graphical user interface to any specific position point on the surrounding areas (e.g., specific position points in four different directions such as up, down, left and right on the graphical user interface whose distance from the center position is a preset distance threshold), including the average scanning speed, average scanning completion time, etc.

[0047] Specifically, for the above-mentioned reverse saccade test, the preset eye movement feature parameters associated with the reverse saccade test can also be multiple eye movement feature parameters related to the subject's eye movement; for example, in the reverse saccade test, the subject's line of sight (that is, the direction of eye gaze) scans from the center position of the above-mentioned graphical user interface to any specific position point on the surrounding areas (such as specific position points in four different directions such as up, down, left and right on the graphical user interface whose distance from the center position is a preset distance threshold) in the opposite direction, and the average reverse saccade completion time, reverse saccade accuracy, etc.

[0048] Specifically, for the above-mentioned counting task test, the preset eye movement feature parameters associated with the counting task test may be a plurality of eye movement feature parameters related to the subject's gaze habits; for example, in the counting task test, for each time a plurality of target balls appear randomly, the average difference between the number of target balls observed by the subject and the actual number of balls that appear, the average frequency of regular contraction and expansion of pupil diameter, the number of the subject's gaze points (i.e., the gaze duration for a position point on the graphical user interface that exceeds a certain preset gaze time threshold can be recorded as a gaze point), the specific gaze duration corresponding to each gaze point, etc.

[0049] S103, inputting the parameter value of each of the preset eye movement feature parameters and the basic physical examination information of the subject into a pre-trained multi-task learning model, and outputting the disease risk prediction result of each neurodegenerative disease corresponding to the subject through each subtask learning model included in the multi-task learning model.

[0050] Here, each of the subtask learning models is used to predict the disease risk of a neurodegenerative disease corresponding to the subject; that is, the above-mentioned multi-task learning model is composed of multiple subtask learning models, and each subtask learning model corresponds to a neurodegenerative disease; wherein, the specific model number of subtask learning models and the specific disease types of neurodegenerative diseases that can be predicted by the above-mentioned multi-task learning model are not subject to any mandatory limitations in the embodiments of the present application.

[0051] For example, the multi-task learning model includes four sub-task learning models, and neurodegenerative diseases include Parkinson's disease, Alzheimer's disease, Huntington's disease, and anterior temporal dementia. If the first sub-task learning model corresponds to Parkinson's disease, the second sub-task learning model corresponds to Alzheimer's disease, the third sub-task learning model corresponds to Huntington's disease, and the fourth sub-task learning model corresponds to anterior temporal dementia, then the multi-task learning model can output the first disease risk prediction result (i.e., the probability of suffering from Parkinson's disease), the second disease risk prediction result (i.e., the probability of suffering from Alzheimer's disease), the third disease risk prediction result (i.e., the probability of suffering from Huntington's disease), and the fourth disease risk prediction result (i.e., the probability of suffering from anterior temporal dementia) corresponding to the subject.

[0052] In the embodiment of the present application, as an optional embodiment, the above step S103 may be performed according to the method described in the following steps a1-a2, specifically:

[0053] Step a1: for each of the subtask learning models, determine the parameter information of multiple target parameters associated with the subtask learning model and the parameter information of shared parameters commonly associated with all subtask learning models from the parameter values ​​of all preset eye movement feature parameters input and the basic physical examination information of the subject.

[0054] Here, for each subtask learning model, the model input data corresponding to the subtask learning model mainly consists of the above-mentioned multiple target parameters and the above-mentioned shared parameters; wherein, the above-mentioned multiple target parameters are determined according to the disease characteristics of the target neurodegenerative disease corresponding to the subtask learning model (that is, the above-mentioned multiple target parameters represent specific parameters only related to the target neurodegenerative disease corresponding to the subtask learning model); the above-mentioned shared parameters are determined according to the common characteristics of each of the neurodegenerative diseases (that is, the above-mentioned shared parameters represent common parameters related to the neurodegenerative diseases corresponding to each subtask learning model); and whether it is the above-mentioned multiple target parameters or the above-mentioned shared parameters, their data sources come from all the input data input into the multi-task learning model (that is, the parameter values ​​of all preset eye movement feature parameters input and the basic physical examination information of the subject).

[0055] Specifically, the above-mentioned shared parameters may include: common features such as the subject's age, gender, vision, etc. that can be obtained from the subject's basic physical examination information; in addition, the shared parameters may also include some preset feature parameters among all preset feature parameters (such as the baseline pupil size obtained based on the pupil light stimulation response test in the above-mentioned step S102, etc.).

[0056] Specifically, the above-mentioned multiple target parameters can be determined according to the disease characteristics of the target neurodegenerative disease actually corresponding to the subtask learning model; for example, if the target neurodegenerative disease is intrinsically related to the change in pupil diameter when exposed to light, gaze habits, and eye movement, then the above-mentioned multiple target parameters corresponding to the subtask learning model may include: multiple first parameters from pupil light stimulation response tests (used to reflect the characteristics of the pupil diameter change when the subject is exposed to light), multiple second parameters from forward saccade tests and reverse saccade tests (used to reflect the characteristics of the subject's eye movement), and multiple third parameters from counting task tests (used to reflect the characteristics of the subject's gaze habits).

[0057] Step a2: input parameter information of the multiple target parameters and parameter information of the shared parameters into the subtask learning model, and output the disease risk prediction result of the target neurodegenerative disease corresponding to the subject through the subtask learning model.

[0058] Here, in the multi-task learning model, an output layer can be configured for each sub-task learning model, wherein, in each output layer, each sub-task learning model can use the softmax activation function to output the probability of the target neurodegenerative disease corresponding to itself (that is, the disease risk prediction result of the target neurodegenerative disease corresponding to the subject); at this time, the characteristics of the softmax activation function can map the specific value range of the output value of each output layer (that is, the probability of each neurodegenerative disease) to between 0 and 1, and constrain the sum of the output values ​​of each output layer to be 1 (equivalent to the sum of the probability of each neurodegenerative disease being 1), so as to be suitable for the output of multiple result probabilities (that is, the disease risk prediction results output by each sub-task learning model) required by the multi-task learning model.

[0059] The following uses the target electronic device as an example to illustrate each step in the disease risk prediction method provided in the embodiment of the present application:

[0060] Regarding the specific implementation process of the above steps S101-S102, when the eye movement test in step S101 is a pupil light stimulation response test, Figure 2a A flow chart of a method for displaying visual guidance information for guiding a subject to perform a pupil light stimulation response test on a graphical user interface provided by an embodiment of the present application is shown, such as Figure 2a As shown, when executing step S101, the method includes steps S201-S203, specifically:

[0061] S201, in response to the start of the pupil light stimulation response test, displaying a black background image for a first preset duration on the graphical user interface through the virtual reality device.

[0062] Here, when conducting a pupil light stimulation response test on a subject, first, a black background image of a first preset duration needs to be presented to the subject on a graphical user interface through a virtual reality device so that the subject's pupil can adapt to the current light conditions.

[0063] It should be noted that the specific value of the above-mentioned first preset time length can be determined according to the actual test requirements of the pupil light stimulation response test. For example, it can be 30 seconds or 35 seconds, etc. The embodiment of the present application does not impose any limitation on the specific value of the above-mentioned first preset time length.

[0064] S202, in response to the first preset time duration being completed, flashing white light on the subject via the virtual reality device.

[0065] Here, based on the implementation of the above-mentioned step S201, when the pupil of the subject has adapted to the current light, and the timing of the target electronic device responding to the first preset time period ends, the virtual reality device can be controlled to flash white light on the subject according to the preset white light flashing time period; that is, the pupil of the subject is stimulated by light by flashing white light on the subject.

[0066] As an example, taking the preset white light flashing duration as 1 second and the above-mentioned first preset duration as 30 seconds as an example, after the 30-second timing ends, the target electronic device can control the virtual reality device to flash white light for 1 second on the subject within the 31st second.

[0067] It should be noted that the specific flashing duration of the white light flashing (that is, the above-mentioned preset white light flashing duration) and the white light intensity used when the white light flashes can also be determined according to the actual test requirements of the pupil light stimulation response test, and the embodiments of the present application do not impose any limitations on this.

[0068] S203, in response to the end of the flashing of the white light, displaying a black background image of a second preset duration on the graphical user interface through the virtual reality device.

[0069] Here, taking the second preset time length of 59 seconds as an example, after the white light flashing ends, the target electronic device can control the virtual reality device in an implementation similar to the above step S201 to continue to restore the display of the black background image for 59 seconds on the graphical user interface; at this time, after the timing of the second preset time length ends, it can be confirmed that the pupil light stimulation response test of the subject is completed.

[0070] Specifically, as an optional embodiment, the target electronic device responds to the expiration of the timing of the second preset time period, and can also display relevant prompt information for prompting the test content of the next eye movement test on the graphical user interface through the virtual reality device, or play the above-mentioned relevant prompt information to the subject in the form of voice playback, so that the subject can more easily understand the visual guidance information displayed on the graphical user interface during the next eye movement test.

[0071] Here, after the pupil light stimulation response test is performed on the subject according to the method described in the above steps S201-S203, Figure 2b FIG. 1 is a data schematic diagram showing eye activity data that can be collected during a pupil light stimulation response test provided by an embodiment of the present application, such as Figure 2bAs shown, the horizontal axis represents the test time of the pupil light stimulation response test recorded, and the vertical axis represents the amplitude corresponding to the pupil contraction of the subject during the pupil light stimulation response test (since it is the contraction amplitude, the value of the vertical axis is a negative number, and the unit can be the pupil contraction change ratio expressed in %); wherein, Figure 2b In the figure, the 0 moment of the horizontal axis is used to mark the time when the subject's pupil is detected to contract (ie, the amplitude of the vertical axis changes significantly).

[0072] Specifically, Figure 2b Taking the eye movement data that can be collected during the pupil light stimulation response test as an example, when executing step S102, the preset eye movement feature parameters that can be extracted from this part of the eye movement data may include at least one of the following:

[0073] 1. In the pupil light stimulation response test, the average pupil size when there is no light stimulation (i.e., the baseline pupil size); for example, taking the first preset time length in the above step S201 as 30 seconds, the average pupil size of the subject can be collected in the second half of the first preset time length (e.g., the 28th second to the 30th second) when the subject has adapted to the current light conditions.

[0074] 2. The latency of pupil contraction when stimulated by strong light (i.e., the time interval between the onset of strong light and the contraction of the subject's pupil); for example, Figure 2b As shown in the figure, pupil contraction latency ① indicates that the timing starts from the moment when the strong light appears until the pupil of the subject is detected to contract (i.e., corresponding to Figure 2b The 0 moment mark on the horizontal axis) corresponds to the length of time between them.

[0075] 3. The maximum contraction speed of the pupil; for example, Figure 2b As shown, the slope of each tangent direction of the curve can be calculated, wherein the maximum calculated slope is the maximum contraction speed of the pupil②.

[0076] 4. The maximum contraction value of the pupil; for example, Figure 2b As shown, since the ordinate represents the amplitude corresponding to the pupil contraction of the subject during the pupil light stimulation response test, where the negative sign is used to indicate that the amplitude belongs to the contraction amplitude, the maximum pupil contraction value ③ is the minimum value of the ordinate (also equivalent to the maximum absolute value of the ordinate).

[0077] 5. The specific time required for the pupil to return to stability; for example, Figure 2b As shown in the figure, the specific time required for the pupil to return to stability④ is the time from the time when the subject's pupil is detected to contract, which corresponds to Figure 2bThe 0 time mark on the horizontal axis) is used to detect that the subject's pupil has returned to a stable state and the timing ends.

[0078] It should be noted that there is no unique specific condition for determining whether the pupil has recovered stability. For example, Figure 2b If the pupil contraction amplitude represented by the middle vertical coordinate is less than or equal to the preset amplitude threshold, it is determined that the subject's pupil has been detected to have returned to stability; or if the difference between the subject's pupil size and the previously detected baseline pupil size is less than or equal to the preset detection error, it is determined that the subject's pupil has been detected to have returned to stability.

[0079] 6. Pupil size after the pupil has stabilized; for example, Figure 2b As shown, from the time when the subject's pupil is detected to be stable until the end of the pupil light stimulation response test, the average value of the pupil contraction amplitude represented by the vertical coordinate during this period of time is calculated, and then the above average value is superimposed on the baseline pupil size to obtain the pupil size after the pupil is stable⑤.

[0080] 7. The difference between the pupil size after stabilization and the above-mentioned reference pupil size. For example, after calculating the pupil size after the pupil is stabilized ⑤, it is only necessary to calculate the difference between the pupil size after the pupil is stabilized ⑤ and the above-mentioned reference pupil size to obtain the difference between the pupil size after stabilization and the above-mentioned reference pupil size ⑥. At this time, if Figure 2b As shown, the difference between the stabilized pupil size and the above-mentioned reference pupil size⑥ can also be calculated directly by using the difference in pupil contraction amplitude relative to the reference pupil after the pupil regains stability.

[0081] 8. The preset characteristic parameters may also include area parameters under the change curve within a certain preset test time; for example, Figure 2b As shown, the area parameter ⑦ can represent the area auc20 under the change curve of 20 seconds (that is, the above preset test time is 20 seconds).

[0082] Regarding the specific implementation process of the above steps S101-S102, when the eye movement test in step S101 is a forward saccade test, Figure 3 A flow chart of a method for displaying visual guidance information for guiding a subject to perform a forward saccade test on a graphical user interface provided by an embodiment of the present application is shown, such as Figure 3 As shown, when executing step S101, the method includes step S301, specifically:

[0083] S301, in response to the start of the forward scanning test, a target bright spot that moves according to a preset movement rule is displayed on the graphical user interface through the virtual reality device; wherein, during the test process of the forward scanning test, the target bright spot is used to guide the subject's line of sight to follow the target bright spot to move.

[0084] Here, when performing a forward scanning test, the background image displayed on the graphical user interface can be a pure black background image or a dark background image with a color darkness lower than a preset threshold, as long as the background image can be visually significantly different from the target bright spot.

[0085] Specifically, in the embodiment of the present application, when executing step S301, the target bright spot may perform a single movement according to the following preset movement rule:

[0086] The target bright spot first appears at the center of the graphical user interface and stays for a first reference time (e.g., 1 second), then randomly flashes to any specific position point on the graphical user interface (any specific position point in four different directions of up, down, left, and right, whose distance from the center position is a preset distance threshold), and stays at the specific position point for a second reference time (e.g., 1.2 seconds) before disappearing, and then, after a third reference time (e.g., after disappearing for 2 seconds), it re-displays at the above-mentioned center position.

[0087] It should be noted that in order to ensure the accuracy of the parameter values ​​of the preset eye movement characteristic parameters associated with the above-mentioned forward scanning test, during the actual forward scanning test, the target bright spot can be controlled to move repeatedly according to the above-mentioned preset movement rules, so as to calculate the average values ​​such as the average scanning speed and average scanning completion time of the subjects in each single movement process as the parameter value of each preset eye movement characteristic parameter.

[0088] Specifically, the intersection points between the corresponding directional rays and the graphical user interface can be obtained from four directions, namely, top, bottom, left, and right, in such a way that the angle between the directional rays and the horizontal line where the center position is located is a preset angle, and the intersection points between the corresponding directional rays and the graphical user interface can be used as specific positions where the target bright spot can randomly flash on the graphical user interface.

[0089] It should be noted that the embodiments of the present application do not impose any limitation on the specific size of the target bright spot; for example, the specific size of the target bright spot may be an apparent diameter of 1.5°, or an apparent diameter of 1.4°, etc.

[0090] Specifically, when executing step S102, the preset eye movement feature parameters that can be extracted from the forward saccade test process may include at least one of the following:

[0091] 1. The average scanning speed of the subject's line of sight from the center of the above-mentioned graphical user interface to any specific position point on the surrounding area; for example, taking 20 repeated random flashes as an example, from the appearance of the target bright spot, the scanning speed is recorded each time the subject's gaze direction moves to the specific position point where the target bright spot is located. Based on the recorded 20 scanning speeds, the average scanning speed can be calculated.

[0092] 2. Average saccadic eye movement latency; for example, still taking 20 repeated random flashes as an example, start timing each time the target bright spot appears, and end timing when the subject's eye movement is detected, and the recorded saccadic eye movement latency can be obtained; at this time, based on the recorded 20 saccadic eye movement latencies, the average saccadic eye movement latency can be calculated.

[0093] 3. Average scanning task completion time; for example, still taking 20 repeated random flashes as an example, the detection starts with each target bright spot appearing, and ends when the target bright spot disappears. During each detection, if it is detected that the subject's eyeballs move to the specific position where the target bright spot is located / the direction of the eyeball movement is consistent with the visual field direction where the target bright spot is located, it is determined that the subject has successfully completed a scanning task. Therefore, based on the average completion time of the successfully completed scanning task recorded during the detection of 20 random flashes, the average scanning task completion time can be obtained.

[0094] 4. Completion accuracy of the scanning task; for example, still taking 20 repeated random flashes as an example, the detection starts with each appearance of the target bright spot and ends when the target bright spot disappears. During each detection, if it is detected that the subject's eyeballs move to the specific position where the target bright spot is located / the direction of the eyeball movement is consistent with the visual field direction where the target bright spot is located, it is determined that the subject has successfully completed a scanning task. Therefore, based on the proportion of the number of successful completions of the scanning task recorded during the 20 random flash detections in the total number of 20 tests, the completion accuracy of the scanning task can be obtained.

[0095] Regarding the specific implementation process of the above steps S101-S102, when the eye movement test in step S101 is a reverse saccade test, Figure 4 A flow chart of a method for displaying visual guidance information for guiding a subject to perform a reverse saccade test on a graphical user interface provided by an embodiment of the present application is shown, such as Figure 4 As shown, when executing step S101, the method includes step S401, specifically:

[0096] S401, in response to the start of the reverse saccade test, a target bright spot that moves according to a preset movement rule is displayed on the graphical user interface through the virtual reality device; wherein, during the test process of the reverse saccade test, the target bright spot is used to guide the subject's line of sight to move in the opposite direction of the movement of the target bright spot.

[0097] Specifically, step S401 is the same as the specific implementation method of the above-mentioned step S301. It only needs to adapt the criterion for the subject to successfully complete the scanning task from "detecting that the subject's eyes move to the specific position where the target bright spot is located / the direction of the eye movement is consistent with the direction of the field of vision where the target bright spot is located" to "detecting that the subject's eyes move to the opposite position of the specific position where the target bright spot is located / the direction of the eye movement is opposite to the direction of the field of vision where the target bright spot is located". The repetitive parts will not be repeated here.

[0098] Regarding the specific implementation process of the above steps S101-S102, when the eye movement test in step S101 is a counting task test, Figure 5 A flow chart of a method for displaying visual guidance information for guiding a subject to perform a counting task test on a graphical user interface provided by an embodiment of the present application is shown, such as Figure 5 As shown, when executing step S101, the method includes step S501, specifically:

[0099] S501, in response to the start of the counting task test, visual guidance information for guiding the subject to perform a single counting task test is displayed on the graphical user interface through the virtual reality device according to the number of repeated tests corresponding to the counting task test.

[0100] Specifically, as an optional embodiment, the visual guidance information for guiding the subject to perform a single counting task test can be displayed on the graphical user interface by the method described in the following steps b1-b2:

[0101] Step b1, in response to the start of a single counting task test, displaying a plurality of randomly distributed red icons and a plurality of green icons on the graphical user interface through the virtual reality device.

[0102] Step b2: in response to the timing of the third preset time period ending, displaying a dark background image of a fourth preset time period on the graphical user interface through the virtual reality device.

[0103] Here, the display time of the multiple red icons and the multiple green icons on the graphical user interface is the third preset time; for example, if the third preset time is 1 second and the fourth preset time is 3 seconds, then in each single counting task test process, the multiple red icons and the multiple green icons can be randomly displayed on the graphical user interface for 1 second, and then the dark background image is displayed for 3 seconds, and then the above test process is repeated, and the multiple red icons and the multiple green icons are randomly displayed on the graphical user interface for 1 second again.

[0104] It should be noted that during each single counting task test, the display quantity and display position of the red icons and green icons are random and do not require any restrictions; and the specific values ​​of the above-mentioned repeated test times, the third preset time length, and the fourth preset time length can also be set according to actual test requirements, and the embodiments of the present application do not impose any limitations on this.

[0105] Regarding the specific implementation process of the above step S103, Figure 6 A flow chart of a method for training each subtask learning model included in a multi-task learning model provided by an embodiment of the present application is shown, such as Figure 6 As shown, the method includes steps S601-S604, specifically:

[0106] S601 , for a plurality of pre-selected sample objects, respectively collecting parameter values ​​of all preset eye movement feature parameters of each of the sample objects and basic physical examination information of each of the sample objects as sample data corresponding to each of the sample objects.

[0107] Here, the plurality of sample objects include: patients with each of the neurodegenerative diseases and healthy subjects who do not suffer from any neurodegenerative disease; each of the sample objects carries a sample label for marking the disease type of the sample object.

[0108] For example, taking the multi-task learning model as an example, in which four sub-task learning models are included and neurodegenerative diseases include Parkinson's disease, Alzheimer's disease, Huntington's disease, and anterior temporal dementia, multiple Parkinson's disease patients, multiple Alzheimer's disease patients, multiple Huntington's disease patients, multiple anterior temporal dementia patients, and multiple healthy subjects (i.e., those without Parkinson's disease, Alzheimer's disease, Huntington's disease, and anterior temporal dementia) can be selected in advance as the multiple sample objects; wherein, for healthy subjects, the sample labels carried by the healthy subjects are used to mark that they do not suffer from any neurodegenerative diseases.

[0109] It should be noted that the embodiments of the present application do not impose any restrictions on the specific number of sample objects of each type, and it is only necessary to ensure that the sample data can meet the data requirements for model training.

[0110] S602: For each of the subtask learning models, determine parameter information of multiple candidate parameters associated with the subtask learning model and parameter information of shared parameters commonly associated with all subtask learning models from sample data corresponding to each of the sample objects.

[0111] Here, the multiple candidate parameters are determined according to the disease characteristics of the target neurodegenerative disease corresponding to the subtask learning model; and the shared parameters are determined according to the common characteristics of each of the neurodegenerative diseases.

[0112] S603, inputting parameter information of the multiple candidate parameters and parameter information of the shared parameters into the subtask learning model, and outputting a disease risk prediction result of the target neurodegenerative disease corresponding to each of the sample objects through the subtask learning model.

[0113] Specifically, the specific implementation of the above steps S602-S603 is the same as the implementation of steps a1-a2 in the above step S103, and the repeated parts are not repeated here.

[0114] S604, calculating the loss between the sample label carried by each of the sample objects and the disease risk prediction result of the target neurodegenerative disease corresponding to each of the sample objects, and adjusting the corresponding weight coefficients of each of the candidate parameters and the shared parameters in the subtask learning model according to the loss calculation result until the subtask learning model reaches convergence.

[0115] Here, the sample label carried by each of the sample objects is equivalent to the actual result of the disease risk corresponding to the sample object itself (i.e., whether the disease is present and the specific type of disease). Therefore, by calculating the loss between the sample label carried by each sample object and the disease risk prediction result of the target neurodegenerative disease corresponding to each sample object, the weight coefficient corresponding to each of the candidate parameters and the shared parameters in the subtask learning model can be adjusted until the subtask learning model converges, and then the training process of the subtask learning model is determined to be ended.

[0116] It should be noted that the specific calculation method of the above loss is not unique. For example, the cross entropy loss function can be used for calculation, or other classification prediction loss functions can be used for calculation. This embodiment of the application does not impose any limitation on this.

[0117] Based on the above-mentioned disease risk prediction method provided in the embodiment of the present application, on the one hand, the subject is guided to complete each eye test through a virtual reality device, and the eye movement data required for disease risk prediction is collected during the test, which is conducive to objectively and quantitatively recording the subject's eye movement characteristics and ensuring the accuracy of the source of disease risk prediction data; on the other hand, by extracting the parameter values ​​of a plurality of preset eye movement feature parameters from the above-mentioned eye activity data and the additional basic physical examination information of the subject, the disease risk prediction results of each neurodegenerative disease corresponding to the subject are output with the help of a multi-task learning model, so that the disease risks of a plurality of different neurodegenerative diseases corresponding to the subject can be comprehensively predicted from multiple angles as a whole, effectively improving the accuracy of various disease risk prediction results.

[0118] Based on the same inventive concept, the present application also provides a disease risk prediction device corresponding to the above-mentioned disease risk prediction method. Since the principle of solving the problem by the disease risk prediction device in the embodiment of the present application is similar to the above-mentioned disease risk prediction method in the embodiment of the present application, the implementation of the disease risk prediction device can refer to the implementation of the above-mentioned disease risk prediction method, and the repeated parts will not be repeated.

[0119] Reference Figure 7 As shown, Figure 7 A schematic diagram of the structure of a disease risk prediction device provided in an embodiment of the present application is shown, wherein the disease risk prediction device comprises:

[0120] A visual guidance module 701 is used for displaying visual guidance information for guiding the subject to perform each eye movement test on a graphical user interface of the virtual reality device during the process of collecting the subject's eye movement data through the virtual reality device, and using the test result of each eye movement test as the collected eye movement data;

[0121] The data acquisition module 702 is used to extract the parameter value of each preset eye movement feature parameter from the eye activity data according to the parameter definition of each preset eye movement feature parameter;

[0122] The risk prediction module 703 is used to input the parameter value of each of the preset eye movement feature parameters and the basic physical examination information of the subject into a pre-trained multi-task learning model, and output the disease risk prediction result of each neurodegenerative disease corresponding to the subject through each sub-task learning model included in the multi-task learning model; wherein each of the sub-task learning models is used to predict the disease risk of a neurodegenerative disease corresponding to the subject.

[0123] In an optional implementation, when the eye movement test is a pupil light stimulation reaction test, the visual guidance module 701 is used to display visual guidance information for guiding the subject to perform the pupil light stimulation reaction test on the graphical user interface by the following method:

[0124] In response to the start of the pupil light stimulation response test, displaying a black background image for a first preset time length on the graphical user interface through the virtual reality device;

[0125] In response to the first preset time duration being completed, flashing white light on the subject through the virtual reality device;

[0126] In response to the end of the white light flashing, a black background image of a second preset duration is displayed on the graphical user interface through the virtual reality device.

[0127] In an optional implementation, when the eye movement test is a forward saccade test, the visual guidance module 701 is used to display visual guidance information for guiding the subject to perform the forward saccade test on the graphical user interface by the following method:

[0128] In response to the start of the forward scanning test, a target bright spot moving according to a preset movement rule is displayed on the graphical user interface through the virtual reality device; wherein, during the test process of the forward scanning test, the target bright spot is used to guide the subject's line of sight to follow the target bright spot to move.

[0129] In an optional implementation, when the eye movement test is a reverse saccade test, the visual guidance module 701 is used to display visual guidance information for guiding the subject to perform the reverse saccade test on the graphical user interface by the following method:

[0130] In response to the start of the reverse saccade test, a target bright spot moving according to a preset movement rule is displayed on the graphical user interface through the virtual reality device; wherein, during the test process of the reverse saccade test, the target bright spot is used to guide the subject's line of sight to move in the opposite direction of the movement of the target bright spot.

[0131] In an optional implementation, when the eye movement test is a counting task test, the visual guidance module 701 is used to display visual guidance information for guiding the subject to perform the counting task test on the graphical user interface by the following method:

[0132] In response to the start of the counting task test, displaying visual guidance information for guiding the subject to perform a single counting task test on the graphical user interface through the virtual reality device according to the number of repeated tests corresponding to the counting task test;

[0133] Wherein, the visual guidance information for guiding the subject to perform a single counting task test is displayed on the graphical user interface by the following method:

[0134] In response to the start of the single counting task test, a plurality of randomly distributed red icons and a plurality of green icons are displayed on the graphical user interface through the virtual reality device; wherein the display duration of the plurality of red icons and the plurality of green icons on the graphical user interface is a third preset duration;

[0135] In response to the timing of the third preset duration ending, a dark background image of a fourth preset duration is displayed on the graphical user interface through the virtual reality device.

[0136] In an optional embodiment, when outputting the disease risk prediction result of each neurodegenerative disease corresponding to the subject through each subtask learning model included in the multi-task learning model, the risk prediction module 703 is used to:

[0137] For each of the subtask learning models, parameter information of a plurality of target parameters associated with the subtask learning model and parameter information of shared parameters commonly associated with all subtask learning models are determined from parameter values ​​of all preset eye movement feature parameters input and basic physical examination information of the subject; wherein the plurality of target parameters are determined according to disease characteristics of the target neurodegenerative disease corresponding to the subtask learning model; and the shared parameters are determined according to the common characteristics of each of the neurodegenerative diseases;

[0138] The parameter information of the multiple target parameters and the parameter information of the shared parameters are input into the subtask learning model, and the disease risk prediction result of the target neurodegenerative disease corresponding to the subject is output through the subtask learning model.

[0139] In an optional implementation, the risk prediction module 703 is used to train each of the subtask learning models included in the multi-task learning model by the following method:

[0140] For a plurality of pre-selected sample objects, parameter values ​​of all preset eye movement feature parameters of each sample object and basic physical examination information of each sample object are collected respectively as sample data corresponding to each sample object; wherein the plurality of sample objects include: patients with each of the neurodegenerative diseases and healthy subjects who do not suffer from any neurodegenerative diseases; each sample object carries a sample label for marking the disease type of the sample object;

[0141] For each of the subtask learning models, parameter information of a plurality of candidate parameters associated with the subtask learning model and parameter information of a shared parameter commonly associated with all the subtask learning models are determined from sample data corresponding to each of the sample objects; wherein the plurality of candidate parameters are determined according to disease characteristics of the target neurodegenerative disease corresponding to the subtask learning model; and the shared parameter is determined according to the common characteristics of each of the neurodegenerative diseases;

[0142] Inputting parameter information of the plurality of candidate parameters and parameter information of the shared parameters into the subtask learning model, and outputting a disease risk prediction result of the target neurodegenerative disease corresponding to each of the sample objects through the subtask learning model;

[0143] The loss between the sample label carried by each of the sample objects and the disease risk prediction result of the target neurodegenerative disease corresponding to each of the sample objects is calculated, and according to the loss calculation result, the weight coefficient corresponding to each of the candidate parameters and the shared parameters in the subtask learning model is adjusted until the subtask learning model reaches convergence.

[0144] Based on the same inventive concept, the present application also provides an electronic device corresponding to the above-mentioned disease risk prediction method. Since the principle of solving the problem by the electronic device in the embodiment of the present application is similar to the above-mentioned disease risk prediction method in the embodiment of the present application, the implementation of the electronic device can refer to the implementation of the above-mentioned disease risk prediction method, and the repeated parts will not be repeated.

[0145] like Figure 8 As shown, an embodiment of the present application provides an electronic device 800 for executing the disease risk prediction method in the present application, the device includes a memory 801, a processor 802, and a computer program stored in the memory 801 and executable on the processor 802, wherein the memory 801 and the processor 802 communicate via a bus, and the processor 802 implements the steps of the disease risk prediction method when executing the computer program.

[0146] Specifically, the above-mentioned memory 801 and processor 802 can be general-purpose memory and processor, which are not specifically limited here. When the processor 802 runs the computer program stored in the memory 801, the above-mentioned disease risk prediction method can be executed.

[0147] Corresponding to the disease risk prediction method in the present application, an embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned disease risk prediction method are executed.

[0148] Specifically, the storage medium can be a general storage medium, such as a mobile disk, a hard disk, etc. When the computer program on the storage medium is run, the above-mentioned disease risk prediction method can be executed.

[0149] In the embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. The system embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of systems or units, which can be electrical, mechanical or other forms.

[0150] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0151] In addition, each functional unit in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0152] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0153] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.

[0154] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application. They should all be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A disease risk prediction method, characterized in that: The disease risk prediction method comprises: In the process of collecting the eye movement data of the subject through the virtual reality device, visual guidance information for guiding the subject to perform each eye movement test is displayed on the graphical user interface of the virtual reality device, and the test result of each eye movement test is used as the collected eye movement data; According to the parameter definition of each preset eye movement feature parameter, extracting the parameter value of each preset eye movement feature parameter from the eye activity data; The parameter value of each of the preset eye movement feature parameters and the basic physical examination information of the subject are input into a pre-trained multi-task learning model, and the disease risk prediction result of each neurodegenerative disease corresponding to the subject is output through each sub-task learning model included in the multi-task learning model; wherein each of the sub-task learning models is used to predict the disease risk of a neurodegenerative disease corresponding to the subject.

2. The disease risk prediction method according to claim 1, characterized in that: When the eye movement test is a pupil light stimulation response test, the visual guidance information for guiding the subject to perform the pupil light stimulation response test is displayed on the graphical user interface by the following method: In response to the start of the pupil light stimulation response test, displaying a black background image for a first preset time length on the graphical user interface through the virtual reality device; In response to the first preset time duration being completed, flashing white light on the subject through the virtual reality device; In response to the end of the white light flashing, a black background image of a second preset duration is displayed on the graphical user interface through the virtual reality device.

3. The disease risk prediction method according to claim 1, characterized in that: When the eye movement test is a forward saccade test, visual guidance information for guiding the subject to perform the forward saccade test is displayed on the graphical user interface by the following method: In response to the start of the forward scanning test, a target bright spot moving according to a preset movement rule is displayed on the graphical user interface through the virtual reality device; wherein, during the test process of the forward scanning test, the target bright spot is used to guide the subject's line of sight to follow the target bright spot to move.

4. The disease risk prediction method according to claim 1, characterized in that: When the eye movement test is a reverse saccade test, visual guidance information for guiding the subject to perform the reverse saccade test is displayed on the graphical user interface by the following method: In response to the start of the reverse saccade test, a target bright spot moving according to a preset movement rule is displayed on the graphical user interface through the virtual reality device; wherein, during the test process of the reverse saccade test, the target bright spot is used to guide the subject's line of sight to move in the opposite direction of the movement of the target bright spot.

5. The disease risk prediction method according to claim 1, characterized in that: When the eye movement test is a counting task test, the visual guidance information for guiding the subject to perform the counting task test is displayed on the graphical user interface by the following method: In response to the start of the counting task test, displaying visual guidance information for guiding the subject to perform a single counting task test on the graphical user interface according to the number of repeated tests corresponding to the counting task test through the virtual reality device; Wherein, the visual guidance information for guiding the subject to perform a single counting task test is displayed on the graphical user interface by the following method: In response to the start of the single counting task test, a plurality of randomly distributed red icons and a plurality of green icons are displayed on the graphical user interface through the virtual reality device; wherein the display duration of the plurality of red icons and the plurality of green icons on the graphical user interface is a third preset duration; In response to the timing of the third preset duration ending, a dark background image of a fourth preset duration is displayed on the graphical user interface through the virtual reality device.

6. The disease risk prediction method according to claim 1, characterized in that: The outputting of the disease risk prediction result of each neurodegenerative disease corresponding to the subject through each subtask learning model included in the multi-task learning model includes: For each of the subtask learning models, parameter information of a plurality of target parameters associated with the subtask learning model and parameter information of shared parameters commonly associated with all subtask learning models are determined from parameter values ​​of all preset eye movement feature parameters input and basic physical examination information of the subject; wherein the plurality of target parameters are determined according to disease characteristics of the target neurodegenerative disease corresponding to the subtask learning model; and the shared parameters are determined according to the common characteristics of each of the neurodegenerative diseases; The parameter information of the multiple target parameters and the parameter information of the shared parameters are input into the subtask learning model, and the disease risk prediction result of the target neurodegenerative disease corresponding to the subject is output through the subtask learning model.

7. The disease risk prediction method according to claim 1, characterized in that: Each of the subtask learning models included in the multi-task learning model is trained by the following method: For a plurality of pre-selected sample objects, parameter values ​​of all preset eye movement feature parameters of each sample object and basic physical examination information of each sample object are collected respectively as sample data corresponding to each sample object; wherein the plurality of sample objects include: patients with each of the neurodegenerative diseases and healthy subjects who do not suffer from any neurodegenerative diseases; each sample object carries a sample label for marking the disease type of the sample object; For each of the subtask learning models, parameter information of a plurality of candidate parameters associated with the subtask learning model and parameter information of a shared parameter commonly associated with all the subtask learning models are determined from sample data corresponding to each of the sample objects; wherein the plurality of candidate parameters are determined according to disease characteristics of the target neurodegenerative disease corresponding to the subtask learning model; and the shared parameter is determined according to the common characteristics of each of the neurodegenerative diseases; Inputting parameter information of the multiple candidate parameters and parameter information of the shared parameters into the subtask learning model, and outputting a disease risk prediction result of the target neurodegenerative disease corresponding to each of the sample objects through the subtask learning model; The loss between the sample label carried by each of the sample objects and the disease risk prediction result of the target neurodegenerative disease corresponding to each of the sample objects is calculated, and according to the loss calculation result, the weight coefficient corresponding to each of the candidate parameters and the shared parameters in the subtask learning model is adjusted until the subtask learning model reaches convergence.

8. A disease risk prediction device, characterized in that: The disease risk prediction device comprises: A visual guidance module, used for displaying visual guidance information for guiding the subject to perform each eye movement test on a graphical user interface of the virtual reality device during the process of collecting the subject's eye movement data through the virtual reality device, and using the test result of each eye movement test as the collected eye movement data; A data acquisition module, configured to extract a parameter value of each preset eye movement feature parameter from the eye activity data according to a parameter definition of each preset eye movement feature parameter; A risk prediction module is used to input the parameter value of each of the preset eye movement feature parameters and the basic physical examination information of the subject into a pre-trained multi-task learning model, and output the disease risk prediction result of each neurodegenerative disease corresponding to the subject through each sub-task learning model included in the multi-task learning model; wherein each of the sub-task learning models is used to predict the disease risk of a neurodegenerative disease corresponding to the subject.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of the disease risk prediction method as described in any one of claims 1 to 7 are performed.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the disease risk prediction method as described in any one of claims 1 to 7 are executed.

Images