Cognitive impairment screening method and device based on xr eye tracker and near-infrared brain imaging

Abstract

The application discloses a cognitive disorder screening method and device based on an XR eye tracker and near-infrared brain imaging, and the device comprises an extended reality scene generation module, an eye tracker data acquisition module, a near-infrared brain imaging module, a data analysis processing module and a cognitive disorder screening module. The extended reality scene generation module is used for generating and displaying a virtual reality scene and an augmented reality scene containing a cognitive task. The eye tracker data acquisition module is used for collecting eye movement data information sets of a user in real time when the user completes the cognitive task in the virtual reality scene and the augmented reality scene. The near-infrared brain imaging module is used for collecting blood oxygen information sets of the cerebral cortex of the user in real time when the user completes the cognitive task in the virtual reality scene and the augmented reality scene. The cognitive disorder screening module is used for performing cognitive disorder screening processing on evaluation process information, the eye movement data information sets and the blood oxygen information sets of the cerebral cortex, and obtaining cognitive disorder information of the user.

Description

Cognitive Impairment Screening Methods and Devices Based on XR Eye Tracking and Near-Infrared Brain Imaging Technical Field

[0001] This invention relates to the fields of extended reality and brain cognition, and specifically to a method and apparatus for screening cognitive impairment based on XR eye trackers and near-infrared brain imaging. Background Technology

[0002] With the accelerating trend of an aging society, an increasing number of people are experiencing cognitive impairment. Cognitive impairment is a common neurological abnormality, including dementia and mild cognitive impairment. Early and accurate screening is crucial for the diagnosis and treatment of cognitive impairment. Currently, commonly used screening methods for cognitive impairment include neuropsychological tests and imaging examinations, but these methods have certain limitations, such as high subjectivity, complex operation, and insensitivity to early lesions.

[0003] How to achieve rapid, accurate, and simple early screening for cognitive impairment is an urgent problem that needs to be solved in the context of the current aging society. Summary of the Invention

[0004] This invention primarily addresses the problem of how to achieve rapid, accurate, and simple early screening for cognitive impairment. This invention discloses a method and device for screening cognitive impairment based on XR eye tracker and near-infrared brain imaging.

[0005] The purpose of this invention is to provide a method and apparatus for screening cognitive impairment based on extended reality eye trackers and brain-computer interfaces, so as to overcome the limitations of existing cognitive impairment screening methods and improve the accuracy, objectivity and efficiency of screening.

[0006] In a first aspect, this application discloses a cognitive impairment screening device based on XR eye tracker and near-infrared brain imaging, comprising: an extended reality scene generation module, an eye tracker data acquisition module, a near-infrared brain imaging module, a data analysis and processing module, and a cognitive impairment screening module.

[0007] The extended reality scene generation module is used to generate and display virtual reality scenes and augmented reality scenes that include cognitive tasks;

[0008] The eye-tracking data acquisition module is used to acquire in real time a set of eye-tracking data information when the user completes cognitive tasks in virtual reality and augmented reality scenarios; the set of eye-tracking data information includes a fixation point information sequence, a fixation time information sequence, and a saccade frequency information sequence; the fixation point information sequence includes fixation point location information; the fixation time information sequence includes fixation time information; the saccade frequency information sequence includes saccade frequency information.

[0009] The near-infrared brain imaging module is used to collect in real time the set of blood oxygen information of the cerebral cortex when the user completes cognitive tasks in virtual reality and augmented reality scenarios; the set of blood oxygen information of the cerebral cortex includes a sequence of blood oxygen information of the brain.

[0010] The data analysis and processing module is connected to the extended reality scene generation module, the eye tracker data acquisition module, the near-infrared brain imaging module, and the cognitive impairment screening module, respectively, and is used to analyze and process the acquired eye movement data information set and the blood oxygen information set of the cerebral cortex to extract the first abnormal feature signal.

[0011] The cognitive impairment screening module is used to perform cognitive impairment screening processing on assessment process information, eye movement data information set and cerebral cortex blood oxygenation information set to obtain the user's cognitive impairment information; the cognitive impairment information includes whether cognitive impairment exists and the degree of cognitive impairment.

[0012] The data analysis and processing module is used to analyze and process the collected eye movement data set and the cerebral cortex blood oxygenation information set, and extract the first abnormal feature signal, including:

[0013] The data analysis and processing module performs anomaly detection processing on the collected eye movement data information set to obtain a first anomaly detection value;

[0014] Determine whether the first abnormal discrimination value is greater than a preset first discrimination threshold to obtain a first abnormal feature signal; if the first abnormal feature signal is negative, send the collected eye movement data information set and the cerebral cortex blood oxygen information set to the cognitive impairment screening module.

[0015] If the first abnormal feature signal is yes, it is confirmed that the cognitive impairment information exists, and the degree of cognitive impairment information is determined as the first abnormal discrimination value.

[0016] The extended reality scene generation module includes a virtual reality scene submodule and an augmented reality scene submodule;

[0017] The virtual reality scene submodule is used to display virtual reality scenes containing cognitive tasks to users;

[0018] The augmented reality scene submodule is used to display augmented reality scenes that include cognitive tasks to users.

[0019] The cognitive impairment screening module is used to process eye movement data and cerebral cortex blood oxygenation data to obtain the user's cognitive impairment information, including:

[0020] The cognitive impairment screening module preprocesses the eye movement data set and the cerebral cortex blood oxygenation information set to obtain preprocessed eye movement data set and cerebral cortex blood oxygenation information set respectively.

[0021] The preprocessed eye movement data set and the cerebral cortex blood oxygenation data set are subjected to cognitive impairment assessment processing to obtain the user's cognitive impairment information.

[0022] A second aspect of this invention discloses a cognitive impairment screening method based on XR eye tracker and near-infrared brain imaging, implemented using the aforementioned cognitive impairment screening device based on XR eye tracker and near-infrared brain imaging, comprising:

[0023] S1, using the extended reality scene generation module, generate and display virtual reality scenes and augmented reality scenes containing cognitive tasks;

[0024] S2, using the eye tracker data acquisition module, a set of eye movement data information is collected in real time when the user completes cognitive tasks in virtual reality and augmented reality scenes;

[0025] S3, using the near-infrared brain imaging module, real-time collection of blood oxygen information of the cerebral cortex when the user completes cognitive tasks in virtual reality and augmented reality scenarios;

[0026] S4, using the data analysis and processing module, analyze and process the collected eye movement data information set and the blood oxygen information set of the cerebral cortex to extract the first abnormal feature signal;

[0027] S5. Using the cognitive impairment screening module, cognitive impairment screening is performed on the assessment process information, eye movement data information set, and cerebral cortex blood oxygenation information set to obtain the user's cognitive impairment information.

[0028] The process of analyzing and processing the collected eye movement data set and the cerebral cortex blood oxygenation information set to extract the first abnormal feature signal includes:

[0029] S41, perform anomaly detection processing on the collected eye movement data information set to obtain the first anomaly detection value;

[0030] S42, determine whether the first abnormal discrimination value is greater than the preset first discrimination threshold, and obtain the first abnormal feature signal; if the first abnormal feature signal is negative, send the collected eye movement data information set and the blood oxygen information set of the cerebral cortex to the cognitive impairment screening module;

[0031] If the first abnormal feature signal is yes, it is confirmed that the cognitive impairment information exists, and the degree of cognitive impairment information is determined as the first abnormal discrimination value.

[0032] The cognitive impairment screening process, which involves analyzing assessment process information, eye-tracking data sets, and cerebral cortex blood oxygenation data sets, yields the user's cognitive impairment information, including:

[0033] S51, the eye movement data information set and the cerebral cortex blood oxygen information set are preprocessed respectively to obtain the preprocessed eye movement data information set and the cerebral cortex blood oxygen information set;

[0034] S52, cognitive impairment assessment processing is performed on the assessment process information, the preprocessed eye movement data information set, and the cerebral cortex blood oxygenation information set to obtain the user's cognitive impairment information.

[0035] The preprocessing operation includes:

[0036] Normalize each set of information to obtain the corresponding normalized set of information;

[0037] Data cleaning is performed on each normalized information set to obtain the corresponding cleaned data set.

[0038] Pattern recognition and discrimination processing is performed on each cleaned dataset to obtain the corresponding preprocessed information set.

[0039] The cognitive impairment assessment is performed on the assessment process information, the pre-processed eye movement data set, and the cerebral cortex blood oxygenation data set to obtain the user's cognitive impairment information, including:

[0040] S521, Establish a conditional probability association model;

[0041] S522, Substitute the historical evaluation dataset into the conditional probability association model, and train the conditional probability association model to obtain the trained conditional probability association model.

[0042] S523, using the trained conditional probability association model, the assessment process information, the preprocessed eye movement data information set, and the blood oxygenation information set of the cerebral cortex are processed to obtain cognitive impairment assessment result information;

[0043] S524, determine whether the cognitive impairment assessment result information is greater than a preset second discrimination threshold, and obtain a second abnormal feature signal; if the second abnormal feature signal is negative, confirm that the cognitive impairment information does not contain cognitive impairment;

[0044] If the second abnormal feature signal is yes, it is confirmed that the cognitive impairment information is cognitive impairment and the cognitive impairment assessment information is determined to be cognitive impairment assessment result information.

[0045] S525, using the cognitive impairment assessment information and the information on the existence of cognitive impairment, construct the user's cognitive impairment information.

[0046] The establishment of the conditional probability association model includes:

[0047] S5211, Establish a network of relationships among influencing factors;

[0048] S5212, using the evaluation process information, the preprocessed eye movement data information set, and the blood oxygenation information set of the cerebral cortex, conditional probability parameters are assigned to the influencing factor correlation network;

[0049] S5213 assigns evidence values ​​to the nodes of influencing factors in the influencing factor association network, thus obtaining a conditional probability association model.

[0050] The beneficial effects of this invention are as follows:

[0051] This invention extends reality display technology and eye-tracking and near-infrared brain imaging technology to the field of cognitive impairment screening, enabling rapid and accurate screening of cognitive impairment.

[0052] This invention can achieve a perfect fusion of virtual and real-world scenes through mixed reality technology, creating an environment that is attached to reality and dynamically reconfigurable, increasing the user's immersive experience and thus improving the credibility of test results.

[0053] This invention uses an eye tracker to monitor the user's eye movement trajectory in real time and obtain information such as visual attention; it also uses near-infrared brain imaging technology to directly detect blood activity in the brain, accurately reflecting the state of cognitive function from a physiological perspective.

[0054] This invention uses a conditional probability association model to process assessment process information, pre-processed eye movement data, and cerebral cortex blood oxygenation information to assess cognitive impairment, thereby obtaining the user's cognitive impairment information. This achieves integrated processing at both the assessment data level and the signal acquisition level. By utilizing assessment process information to provide feedback on the assessment results, the accuracy and comprehensiveness of the cognitive impairment screening model are improved, and the applicable scenarios of the model are expanded.

[0055] This invention utilizes a data analysis and processing module to first screen eye-tracking data in advance and introduces difference calculation processing to obtain the first anomaly discrimination value, thereby reducing the amount of data to be screened and improving screening efficiency. Attached Figure Description

[0056] Figure 1 is a diagram showing the composition of the device of the present invention;

[0057] Figure 2 is a flowchart of the implementation of the method of the present invention. Detailed Implementation

[0058] To better understand the content of this invention, an embodiment is provided here.

[0059] Figure 1 is a diagram of the device of the present invention; Figure 2 is a flowchart of the implementation of the method of the present invention.

[0060] In a first aspect, this application discloses a cognitive impairment screening device based on XR eye tracker and near-infrared brain imaging, comprising: an extended reality scene generation module, an eye tracker data acquisition module, a near-infrared brain imaging module, a data analysis and processing module, and a cognitive impairment screening module.

[0061] The extended reality scene generation module is used to generate and display virtual reality scenes and augmented reality scenes that include cognitive tasks; the cognitive tasks include object recognition, spatial navigation, memory testing, etc.

[0062] The eye tracker data acquisition module is used to collect a set of eye movement data information of the user in an extended reality scenario in real time. The set of eye movement data information includes a fixation point information sequence, a fixation time information sequence, and a saccade frequency information sequence. The fixation point information sequence includes fixation point location information; the fixation time information sequence includes fixation time information; and the saccade frequency information sequence includes saccade frequency information.

[0063] The near-infrared brain imaging module is used to acquire a set of blood oxygenation information from the cerebral cortex when the user completes a cognitive task. This set of cerebral cortex blood oxygenation information includes a sequence of brain blood oxygenation information.

[0064] The data analysis and processing module is connected to the extended reality scene generation module, the eye tracker data acquisition module, the near-infrared brain imaging module, and the cognitive impairment screening module, respectively, and is used to analyze and process the acquired eye movement data information set to extract the first abnormal feature signal.

[0065] The cognitive impairment screening module is used to process process assessment information, eye movement data, and cerebral cortex blood oxygenation information to obtain the user's cognitive impairment information; the cognitive impairment information includes whether cognitive impairment exists and the degree of cognitive impairment.

[0066] The near-infrared brain imaging module can be implemented using a near-infrared brain imaging device, the extended reality scene generation module can be implemented using an XR device, and the eye tracker data acquisition module can be implemented using an eye tracker.

[0067] The data analysis and processing module is used to analyze and process the collected eye-tracking data information set, and extract the first abnormal feature signal, including:

[0068] The data analysis and processing module performs anomaly detection processing on the collected eye movement data information set to obtain a first anomaly detection value;

[0069] Determine whether the first abnormal discrimination value is greater than a preset first discrimination threshold to obtain a first abnormal feature signal; if the first abnormal feature signal is negative, send the collected eye movement data information set and the cerebral cortex blood oxygen information set to the cognitive impairment screening module.

[0070] If the first abnormal feature signal is yes, it is confirmed that the cognitive impairment information is present, and the degree of cognitive impairment information is determined as the first abnormal discrimination value;

[0071] The extended reality scene generation module includes a virtual reality scene submodule and an augmented reality scene submodule;

[0072] The virtual reality scene submodule is used to display virtual reality scenes containing cognitive tasks to users;

[0073] The augmented reality scene submodule is used to display augmented reality scenes containing cognitive tasks to users;

[0074] The cognitive impairment screening module is used to process process assessment information, eye movement data, and cerebral cortex blood oxygenation information to obtain the user's cognitive impairment information, including:

[0075] The eye movement data set and the cerebral cortex blood oxygenation information set are preprocessed to obtain the preprocessed eye movement data set and the cerebral cortex blood oxygenation information set, respectively.

[0076] The cognitive impairment information of the user is obtained by performing cognitive impairment assessment on the process evaluation information, the preprocessed eye movement data information set, and the blood oxygenation information set of the cerebral cortex;

[0077] The preprocessing operation includes:

[0078] Normalize each set of information to obtain the corresponding normalized set of information;

[0079] Data cleaning is performed on each normalized information set to obtain the corresponding cleaned data set.

[0080] Pattern recognition and discrimination processing is performed on each cleaned dataset to obtain the corresponding preprocessed information set;

[0081] The cognitive impairment assessment is performed on the process evaluation information, the preprocessed eye movement data set, and the cerebral cortex blood oxygenation information set to obtain the user's cognitive impairment information, including:

[0082] Establish a conditional probability association model;

[0083] The historical evaluation dataset is substituted into the conditional probability association model, and the conditional probability association model is trained to obtain the trained conditional probability association model.

[0084] Using the trained conditional probability association model, the process evaluation information, the preprocessed eye movement data set, and the blood oxygenation information set of the cerebral cortex are processed to obtain cognitive impairment evaluation results.

[0085] Determine whether the cognitive impairment assessment result is greater than a preset second discrimination threshold to obtain a second abnormal feature signal; if the second abnormal feature signal is negative, confirm that the cognitive impairment information does not indicate the absence of cognitive impairment.

[0086] If the second abnormal feature signal is yes, it is confirmed that the cognitive impairment information is cognitive impairment and the cognitive impairment assessment information is determined to be cognitive impairment assessment result information.

[0087] Using the cognitive impairment assessment information and the information on the presence or absence of cognitive impairment, the user's cognitive impairment information is constructed.

[0088] The establishment of the conditional probability association model includes:

[0089] Establish a network of relationships among influencing factors;

[0090] Using the assessment process information, the preprocessed eye movement data set, and the cerebral cortex blood oxygenation information set, conditional probability parameters are assigned to the influencing factor correlation network;

[0091] By assigning evidence values ​​to the nodes of influencing factors in the influencing factor association network, a conditional probability association model is obtained.

[0092] The assessment process information includes the following factors: whether classic methods are used, whether mature tools are used, whether supervised samples are used, whether the assessment data characteristics are suitable, whether the assessment business characteristics are suitable, whether the analysis of the assessment purpose is sufficient, whether the analysis of the assessment object is accurate, whether the assessment subject has full control, whether the assessment results are verified, whether the indicators are complete, whether the indicators are independent, whether the indicators are consistent, whether the indicators are objective, the credibility of the assessment process, and the result values ​​and probability values ​​of all the above factors related to the technical level of the personnel.

[0093] The assessment process information can be obtained by prior assessment of cognitive impairment screening methods.

[0094] The establishment of the influencing factor relationship network specifically includes using a hierarchical Bayesian network method, introducing one root node: assessment credibility; and then introducing seven intermediate nodes: assessment process credibility, assessment method credibility, primary data product quality, intermediate data product quality, final data product quality, assessment indicator construction quality, and quantitative indicator credibility.

[0095] Based on knowledge and information from the business domain, data domain, and evaluation domain, a network of influencing factors is established, consisting of 20 influencing factor nodes and 7 intermediate nodes, to establish the root node.

[0096] A network of relationships among influencing factors is constructed using root nodes, intermediate nodes, and influencing factor nodes. These 20 influencing factor nodes include: whether classical methods are used; whether mature tools are used; whether supervised samples are used; whether the data is adapted to the characteristics of the assessment data; whether the business characteristics of the assessment are adapted to the characteristics of the assessment; whether the analysis of the assessment purpose is sufficient; whether the analysis of the assessment object is accurate; whether the assessment subject has full control; whether the assessment results are verified; whether the indicators are complete; whether the indicators are independent; whether the indicators are consistent; whether the indicators are objective; the credibility of the assessment process; the technical level of the personnel; the mean and variance ratio of the preprocessed eye-tracking data set; the mean and variance ratio of the preprocessed cortical blood oxygenation data set; whether data cleaning is performed; and the median and variance ratio of the preprocessed eye-tracking data set and the preprocessed cortical blood oxygenation data set.

[0097] In the network of influencing factors, the credibility of the evaluation method is jointly determined by whether a classic method is used, whether a mature tool is used, whether a supervised sample is used, whether it is adapted to the characteristics of the evaluation data, and whether it is adapted to the characteristics of the evaluation business.

[0098] The credibility of the assessment process is determined by whether the analysis of the assessment purpose is sufficient, whether the analysis of the assessment object is accurate, whether the assessment entity has full control over the entire process, and whether the assessment results are verified.

[0099] The quality of the evaluation indicators is determined by six factors: whether the indicators are complete, whether they are independent, whether they are consistent, whether they are objective, the credibility of the evaluation process, and the technical level of the personnel.

[0100] The reliability of quantitative indicators is determined by two factors: the quality of indicator construction and the quality of the final data product. The quality of the primary data product is jointly determined by two factors: the mean and variance ratio of the preprocessed eye-tracking data set and the mean and variance ratio of the preprocessed cerebral cortex blood oxygenation data set.

[0101] The quality of intermediate data products is determined by three factors: the median and variance ratio of the eye-tracking data set (whether or not data cleaning and preprocessing have been performed), and the median and variance ratio of the preprocessed cerebral cortex blood oxygenation data set.

[0102] The quality of the final data product is determined by two factors: the quality of the intermediate data product and the technical level of the personnel.

[0103] The credibility of the final root node assessment is jointly determined by the credibility of the assessment process, the credibility of the assessment method, the quality of the primary data product, the quality of the intermediate data product, the quality of the final data product, the quality of the assessment indicator construction, and the credibility of the quantitative indicators.

[0104] Based on the above-mentioned determining relationships, a network of relationships among influencing factors is constructed.

[0105] In the network of influencing factors, the values ​​of the 20 influencing factor nodes are determined based on the assessment process information, the preprocessed eye movement data information set, and the blood oxygenation information set of the cerebral cortex. Then, the assessment reliability value of one root node is calculated. The assessment reliability value is used to represent the cognitive impairment assessment result information.

[0106] The step of assigning conditional probability parameters to the influencing factor correlation network using the assessment process information, the preprocessed eye movement data set, and the cerebral cortex blood oxygenation information set includes:

[0107] The aforementioned network of relationships between influencing factors is defined as a Boolean network, with all nodes taking values ​​of True or False, representing whether the negative aspects of the corresponding influencing factor events occur or not.

[0108] For any non-leaf node in the network, i.e., one root node and seven intermediate nodes, establish a conditional probability dependency relationship between the non-leaf node and all its corresponding parent nodes, i.e., P(non-leaf node|all parent nodes), expressed in the form of a conditional probability expression or the LeakyNoisyor function, and assign values ​​to the conditional probability parameters. When the number of parent nodes is less than or equal to two, the conditional probability expression is used directly; when the number of parent nodes is greater than two, the Leaky Noisyor function noisyor() is used. The parent node corresponding to the non-leaf node is the node one level above the non-leaf node.

[0109] The specific expression of the Leaky Noisyor function noisyor() is as follows:

[0110] Let X be any node in a Boolean Bayesian network, and let its N parent nodes form a set. n is the parent node index. Let X be the nth parent node, and let... express All nodes in the array are set to False. express Nodes in the middle Except for True, all other nodes are set to False. express Let the indices of all nodes that are True be set to:

[0111]

[0112] Represents all parent nodes of X The conditional probability that X is True when all values ​​are False. Represents all parent nodes of X Nodes in the middle The conditional probability that X is True when all nodes except True are False; when the number of parent nodes is less than or equal to 2, let X and its parent node set be... The conditional probability expression is:

[0113]

[0114] Accordingly, when the number of parent nodes is greater than 2, the expression for the conditional probability dependency between non-leaf nodes and all parent nodes adopts the Leaky Noisyor function noisyor(), the specific expression of which is:

[0115]

[0116] The function takes as input all parent nodes of X. Value pairs and p0.

[0117] Based on the probability values ​​in the evaluation process information, the preprocessed eye movement data information set, and the blood oxygenation information set of the cerebral cortex, the conditional probability values ​​of non-lobular nodes and all their corresponding parent nodes are calculated.

[0118] Assigning evidence values ​​to the nodes of influencing factors in the influencing factor correlation network to obtain a conditional probability correlation model includes:

[0119] Evidence values ​​are assigned to 20 nodes in the influencing factor association network. Each evidence value represents the probability of an influencing factor occurring, i.e., the probability that the influencing factor node is True. For any influencing factor node Y, the evidence is divided into hard evidence and soft evidence. Hard evidence has a value of 1 or 0, i.e., P(Y=True)=1 or P(Y=True)=0, which indicates whether the influencing factor event has occurred or not. Soft evidence has a value α∈(0,1), i.e., 0<P(Y=True)=α<1, which describes the objective or subjective probability of the influencing factor event occurring, or the proportion of the range of the influencing factor event occurring in the overall space or time of cognitive impairment screening. The evidence from all influencing factor nodes constitutes the network evidence set Evidence.

[0120] Calculate the conditional probability P(A=False|Evidence) of the network root node taking the value False when the network evidence set Evidence is known, where A represents the root node. Use the conditional probability P(A=False|Evidence) as information on the degree of cognitive impairment.

[0121] The conditional probability P(A=False|Evidence) of the network root node being False when the network evidence set Evidence is known can be obtained using Bayesian network modeling software.

[0122] The step of substituting the historical evaluation dataset into the conditional probability association model and training the conditional probability association model to obtain the trained conditional probability association model includes:

[0123] The historical assessment dataset includes a collection of data and a corresponding set of cognitive impairment assessment values; the collection of data includes the collected data; the collected data includes historical assessment process information, historical eye movement data information, and historical blood oxygenation information; the set of cognitive impairment assessment values ​​includes the cognitive impairment assessment value corresponding to each collected data, which is also the label value corresponding to each collected data.

[0124] The historical assessment process information includes historical data on the results and probability values ​​of all the above factors, including whether classic methods were used, whether mature tools were used, whether supervised samples were used, whether the assessment data characteristics were appropriate, whether the assessment business characteristics were appropriate, whether the analysis of the assessment purpose was sufficient, whether the analysis of the assessment object was accurate, whether the assessment entity controlled the entire process, whether the assessment results were verified, whether the indicators were complete, whether the indicators were independent, whether the indicators were consistent, whether the indicators were objective, the credibility of the assessment process, and the technical level of the personnel.

[0125] Initialize the number of training iterations;

[0126] The collected data from the historical evaluation dataset is used as input data and input into the conditional probability association model;

[0127] The input data is processed using the conditional probability association model to obtain the predicted value;

[0128] The difference between the predicted value and the label information corresponding to the input data is calculated to obtain the difference value.

[0129] Determine whether the difference value satisfies the convergence condition to obtain the first determination result;

[0130] When the first judgment result is negative, it is determined whether the training iteration count value is equal to the training count threshold to obtain the second judgment result;

[0131] When the second judgment result is negative, the model training state is determined to be that the termination training condition is not met.

[0132] When the second judgment result is yes, it is determined that the model training state meets the termination training condition;

[0133] When the first judgment result is yes, it is determined that the model training state meets the termination training condition;

[0134] When the training state of the model does not meet the termination training condition, the parameter update model is used to update the parameters of the conditional probability association model, the training iteration count is increased by 1, and the collected data in the historical evaluation dataset is used as input data to input the conditional probability association model.

[0135] When the training state of the model meets the termination training condition, the training process of the conditional probability association model is completed, and the trained conditional probability association model is obtained.

[0136] The difference value satisfies the convergence condition when it is less than a preset convergence threshold; the difference value does not satisfy the convergence condition when it is not less than a preset convergence threshold.

[0137] The difference calculation process can be implemented using a loss function.

[0138] The loss function can be the cross-entropy loss function.

[0139] The parameter update model is as follows:

[0140]

[0141] θ←θ+v;

[0142] In the formula, Let v be the difference value calculated for the i-th collected data in the historical evaluation dataset, θ be the parameter update value, η be the initial parameter learning rate, and α be the momentum angle parameter, where 0 ≤ α ≤ π / 4. This indicates taking the partial derivative with respect to the variable θ;

[0143] The parameters to be updated in the conditional probability association model are the conditional probability values ​​between each node;

[0144] The normalization process maps data from different value ranges to a specified value range, and its mathematical expression is:

[0145]

[0146] Where t is a specific data value in the dataset, t max t represents the maximum value of all data in the dataset. min t is the minimum value of all data in the dataset. 归一化 The value of data t after normalization;

[0147] The data cleaning process includes filling in missing values, smoothing noisy data, and smoothing or deleting outliers. Smoothing noisy data involves first identifying the noisy data, and then smoothing it based on the data preceding and following it. The noisy data refers to values ​​whose values ​​are less than the sensor's detection sensitivity for the observed data, or greater than the sensor's measurement upper limit for the observed data. Outlier identification can be performed using a Kalman filter. The filling values ​​for missing values ​​can be determined by averaging the measurements within a certain sampling interval before and after the missing value.

[0148] The pattern recognition and discrimination process includes:

[0149] For each data attribute in the cleaned dataset, the data collection information is used as the known independent variable, and the data value is used as the known dependent variable. An approximation curve is constructed using the known independent and dependent variables. A function approximation method is used to fit the approximation curve to obtain the best uniform approximation polynomial f(Ix) for the data attribute. The known independent variable is then processed using the best uniform approximation polynomial f(Ix) to obtain an approximate dependent variable. The absolute value of the difference between the approximate dependent variable and the corresponding known dependent variable is determined to be greater than a set second regression threshold. If it is greater than the second regression threshold, the data is deleted from the cleaned dataset; if it is less than or equal to the second regression threshold, the data is not processed.

[0150] The data from all the execution feature pre-analysis of the cleaned data set are fused to obtain the corresponding pre-processed information set;

[0151] The curve fitting of the curve to be approximated using the function approximation method can employ the best uniform linear approximation method. The best uniform approximation polynomial f(Ix) is expressed as:

[0152] f(Ix)=α P1 (Ix) P1 +α P1-1 (Ix) P1-1 +…+α2(Ix) 2 +α1(Ix)+α0,

[0153] Where P1 is the order of the best uniform approximation polynomial f(Ix), α0, α1, α2, ..., α P1 The coefficients of the best uniform approximation polynomial f(Ix);

[0154] The process of performing anomaly detection on the acquired eye-tracking data set yields a first anomaly detection value, including...

[0155] An eye-tracking data matrix is ​​constructed using the aforementioned eye-tracking data set; the row vectors of the eye-tracking data matrix are sequences of fixation point information, fixation time information, and saccade frequency information.

[0156] A standard eye-tracking data sequence is obtained; the row vectors of the standard eye-tracking data sequence are the standard fixation point information sequence, the standard fixation time information sequence, and the standard saccade frequency information sequence.

[0157] The difference calculation process is performed on the standard eye-tracking data sequence and the eye-tracking data matrix to obtain the first anomaly discrimination value;

[0158] The difference calculation includes:

[0159] The difference matrix is ​​obtained by subtracting the standard eye-tracking data sequence and the eye-tracking data matrix.

[0160] Perform cross-correlation calculations on the difference matrix to obtain the cross-correlation matrix R; i,j Let r represent the element in the i-th row and j-th position of the cross-correlation matrix R. i,j The calculation expression is:

[0161] r i,j =E(e i e j ),

[0162] Where E() represents the mean, e i This represents the i-th row of the difference matrix;

[0163] For the cross-correlation matrix R, normalization is performed on each row vector to obtain a normalized correlation value sequence; the expression for the normalization calculation is:

[0164]

[0165] in, Let N represent the j-th value of the i-th normalized correlation value sequence, where N is the number of elements contained in the normalized correlation value sequence.

[0166] Logarithmic summation is performed on each normalized correlation value sequence to obtain the uncertainty value of each normalized correlation value sequence;

[0167] The calculation expression for the logarithmic summation process is as follows:

[0168]

[0169] Among them, H i Let be the uncertainty value of the i-th normalized correlation value sequence;

[0170] Angle normalization is performed on the uncertainty value of each normalized correlation value sequence to obtain the characteristic value of each normalized correlation value sequence;

[0171] The calculation expression for the angle normalization process is:

[0172]

[0173] in, Let be the eigenvalue of the i-th normalized correlation value sequence, and M be the total number of normalized correlation value sequences;

[0174] By using the eigenvalues ​​of all sequences, the mean of the row vectors of the difference matrix is ​​weighted and summed to obtain the first anomaly detection value.

[0175] XR (Extended Reality) technology includes AR (Augmented Reality), VR (Virtual Reality), and MR (Mixed Reality), which uses hardware devices combined with various technologies to merge virtual content with real-world scenes.

[0176] A second aspect of this application discloses a cognitive impairment screening method based on XR eye tracker and near-infrared brain imaging, implemented using the aforementioned cognitive impairment screening device based on XR eye tracker and near-infrared brain imaging, comprising:

[0177] S1, using the extended reality scene generation module, generate and display virtual reality scenes and augmented reality scenes containing cognitive tasks;

[0178] S2, using the eye tracker data acquisition module, a set of eye movement data information is collected in real time when the user completes cognitive tasks in virtual reality and augmented reality scenes;

[0179] S3, using the near-infrared brain imaging module, real-time collection of blood oxygen information of the cerebral cortex when the user completes cognitive tasks in virtual reality and augmented reality scenarios;

[0180] S4, using the data analysis and processing module, analyze and process the collected eye movement data information set and the blood oxygen information set of the cerebral cortex to extract the first abnormal feature signal;

[0181] S5. Using the cognitive impairment screening module, cognitive impairment screening is performed on the assessment process information, eye movement data information set, and cerebral cortex blood oxygenation information set to obtain the user's cognitive impairment information.

[0182] The process of analyzing and processing the collected eye movement data set and the cerebral cortex blood oxygenation information set to extract the first abnormal feature signal includes:

[0183] S41, perform anomaly detection processing on the collected eye movement data information set to obtain the first anomaly detection value;

[0184] S42, determine whether the first abnormal discrimination value is greater than the preset first discrimination threshold, and obtain the first abnormal feature signal; if the first abnormal feature signal is negative, send the collected eye movement data information set and the blood oxygen information set of the cerebral cortex to the cognitive impairment screening module;

[0185] If the first abnormal feature signal is yes, it is confirmed that the cognitive impairment information exists, and the degree of cognitive impairment information is determined as the first abnormal discrimination value.

[0186] The cognitive impairment screening process, which involves analyzing assessment process information, eye-tracking data sets, and cerebral cortex blood oxygenation data sets, yields the user's cognitive impairment information, including:

[0187] S51, the eye movement data information set and the cerebral cortex blood oxygen information set are preprocessed respectively to obtain the preprocessed eye movement data information set and the cerebral cortex blood oxygen information set;

[0188] S52, cognitive impairment assessment processing is performed on the assessment process information, the preprocessed eye movement data information set, and the cerebral cortex blood oxygenation information set to obtain the user's cognitive impairment information.

[0189] The preprocessing operation includes:

[0190] Normalize each set of information to obtain the corresponding normalized set of information;

[0191] Data cleaning is performed on each normalized information set to obtain the corresponding cleaned data set.

[0192] Pattern recognition and discrimination processing is performed on each cleaned dataset to obtain the corresponding preprocessed information set.

[0193] The cognitive impairment assessment is performed on the assessment process information, the pre-processed eye movement data set, and the cerebral cortex blood oxygenation data set to obtain the user's cognitive impairment information, including:

[0194] S521, Establish a conditional probability association model;

[0195] S522, Substitute the historical evaluation dataset into the conditional probability association model, and train the conditional probability association model to obtain the trained conditional probability association model.

[0196] S523, using the trained conditional probability association model, the assessment process information, the preprocessed eye movement data information set, and the blood oxygenation information set of the cerebral cortex are processed to obtain cognitive impairment assessment result information;

[0197] S524, determine whether the cognitive impairment assessment result information is greater than a preset second discrimination threshold, and obtain a second abnormal feature signal; if the second abnormal feature signal is negative, confirm that the cognitive impairment information does not contain cognitive impairment;

[0198] If the second abnormal feature signal is yes, it is confirmed that the cognitive impairment information is cognitive impairment and the cognitive impairment assessment information is determined to be cognitive impairment assessment result information.

[0199] S525, using the cognitive impairment assessment information and the information on the existence of cognitive impairment, construct the user's cognitive impairment information.

[0200] The establishment of the conditional probability association model includes:

[0201] S5211, Establish a network of relationships among influencing factors;

[0202] S5212, using the evaluation process information, the preprocessed eye movement data information set, and the blood oxygenation information set of the cerebral cortex, conditional probability parameters are assigned to the influencing factor correlation network;

[0203] S5213 assigns evidence values ​​to the nodes of influencing factors in the influencing factor association network, thus obtaining a conditional probability association model.

[0204] The assessment process information includes the following factors: whether classic methods are used, whether mature tools are used, whether supervised samples are used, whether the assessment data characteristics are suitable, whether the assessment business characteristics are suitable, whether the analysis of the assessment purpose is sufficient, whether the analysis of the assessment object is accurate, whether the assessment subject has full control, whether the assessment results are verified, whether the indicators are complete, whether the indicators are independent, whether the indicators are consistent, whether the indicators are objective, the credibility of the assessment process, and the result values ​​and probability values ​​of all the above factors related to the technical level of the personnel.

[0205] The assessment process information can be obtained by prior assessment of cognitive impairment screening methods.

[0206] The establishment of the influencing factor relationship network specifically includes using a hierarchical Bayesian network method, introducing one root node: assessment credibility; and then introducing seven intermediate nodes: assessment process credibility, assessment method credibility, primary data product quality, intermediate data product quality, final data product quality, assessment indicator construction quality, and quantitative indicator credibility.

[0207] Based on knowledge and information from the business domain, data domain, and evaluation domain, a network of influencing factors is established, consisting of 20 influencing factor nodes and 7 intermediate nodes, to establish the root node.

[0208] A network of relationships among influencing factors is constructed using root nodes, intermediate nodes, and influencing factor nodes. These 20 influencing factor nodes include: whether classical methods are used; whether mature tools are used; whether supervised samples are used; whether the data is adapted to the characteristics of the assessment data; whether the business characteristics of the assessment are adapted to the characteristics of the assessment; whether the analysis of the assessment purpose is sufficient; whether the analysis of the assessment object is accurate; whether the assessment subject has full control; whether the assessment results are verified; whether the indicators are complete; whether the indicators are independent; whether the indicators are consistent; whether the indicators are objective; the credibility of the assessment process; the technical level of the personnel; the mean and variance ratio of the preprocessed eye-tracking data set; the mean and variance ratio of the preprocessed cortical blood oxygenation data set; whether data cleaning is performed; and the median and variance ratio of the preprocessed eye-tracking data set and the preprocessed cortical blood oxygenation data set.

[0209] In the network of influencing factors, the credibility of the evaluation method is jointly determined by whether a classic method is used, whether a mature tool is used, whether a supervised sample is used, whether it is adapted to the characteristics of the evaluation data, and whether it is adapted to the characteristics of the evaluation business.

[0210] The credibility of the assessment process is determined by whether the analysis of the assessment purpose is sufficient, whether the analysis of the assessment object is accurate, whether the assessment entity has full control over the entire process, and whether the assessment results are verified.

[0211] The quality of the evaluation indicators is determined by six factors: whether the indicators are complete, whether they are independent, whether they are consistent, whether they are objective, the credibility of the evaluation process, and the technical level of the personnel.

[0212] The reliability of quantitative indicators is determined by two factors: the quality of indicator construction and the quality of the final data product. The quality of the primary data product is jointly determined by two factors: the mean and variance ratio of the preprocessed eye-tracking data set and the mean and variance ratio of the preprocessed cerebral cortex blood oxygenation data set.

[0213] The quality of intermediate data products is determined by three factors: the median and variance ratio of the eye-tracking data set (whether or not data cleaning and preprocessing have been performed), and the median and variance ratio of the preprocessed cerebral cortex blood oxygenation data set.

[0214] The quality of the final data product is determined by two factors: the quality of the intermediate data product and the technical level of the personnel.

[0215] The credibility of the final root node assessment is jointly determined by the credibility of the assessment process, the credibility of the assessment method, the quality of the primary data product, the quality of the intermediate data product, the quality of the final data product, the quality of the assessment indicator construction, and the credibility of the quantitative indicators.

[0216] Based on the above-mentioned determining relationships, a network of relationships among influencing factors is constructed.

[0217] In the network of influencing factors, the values ​​of the 20 influencing factor nodes are determined based on the assessment process information, the preprocessed eye movement data information set, and the blood oxygenation information set of the cerebral cortex. Then, the assessment reliability value of one root node is calculated. The assessment reliability value is used to represent the cognitive impairment assessment result information.

[0218] The step of assigning conditional probability parameters to the influencing factor correlation network using the assessment process information, the preprocessed eye movement data set, and the cerebral cortex blood oxygenation information set includes:

[0219] The aforementioned network of relationships between influencing factors is defined as a Boolean network, with all nodes taking values ​​of True or False, representing whether the negative aspects of the corresponding influencing factor events occur or not.

[0220] For any non-leaf node in the network, i.e., one root node and seven intermediate nodes, establish a conditional probability dependency relationship between the non-leaf node and all its corresponding parent nodes, i.e., P(non-leaf node|all parent nodes), expressed in the form of a conditional probability expression or the LeakyNoisyor function, and assign values ​​to the conditional probability parameters. When the number of parent nodes is less than or equal to two, the conditional probability expression is used directly; when the number of parent nodes is greater than two, the Leaky Noisyor function noisyor() is used. The parent node corresponding to the non-leaf node is the node one level above the non-leaf node.

[0221] The specific expression of the Leaky Noisyor function noisyor() is as follows:

[0222] Let X be any node in a Boolean Bayesian network, and let its N parent nodes form a set. n is the parent node index. Let X be the nth parent node, and let... express All nodes in the array are set to False. express Nodes in the middle Except for True, all other nodes are set to False. express Let the indices of all nodes that are True be set to:

[0223]

[0224] Represents all parent nodes of X The conditional probability that X is True when all values ​​are False. Represents all parent nodes of X Nodes in the middle The conditional probability that X is True when all nodes except True are False; when the number of parent nodes is less than or equal to 2, let the conditional probability expression of X and its parent node set X be:

[0225]

[0226] Accordingly, when the number of parent nodes is greater than 2, the expression for the conditional probability dependency between non-leaf nodes and all parent nodes adopts the Leaky Noisyor function noisyor(), the specific expression of which is:

[0227]

[0228] The input of this function is the value pairs of all the parent nodes of X and p0.

[0229] According to the probability values in the evaluation process information, the preprocessed set of eye movement data information, and the set of blood oxygen information of the cerebral cortex, calculate the conditional probability values of the non-leaf nodes and their corresponding parent nodes;

[0230] Assign evidence values to the influencing factor nodes in the influencing factor association relationship network to obtain a conditional probability association model, including:

[0231] Assign evidence values to 20 influencing factor nodes in the influencing factor association relationship network. The evidence value is the probability of the occurrence of the influencing factor, that is, the probability that the influencing factor node takes True; for any influencing factor node Y, its evidence is divided into two types: hard evidence and soft evidence. The hard evidence takes values of 1 or 0, that is, P(Y = True) = 1 or P(Y = True) = 0, which is used to represent whether the influencing factor event occurs or not. The soft evidence takes values of α ∈ (0, 1), that is, <0<P(Y = True) = α<1>, which is used to express the objective or subjective probability of the occurrence of the influencing factor event, or to express the proportion of the range of the occurrence of the influencing factor event in the overall space or time of cognitive impairment screening; the evidence of all influencing factor nodes constitutes the network evidence set Evidence;

[0232] Find the conditional probability P(A = False|Evidence) that the network root node takes False when the network evidence set Evidence is known, where A represents the root node, and use the conditional probability P(A = False|Evidence) as the cognitive impairment degree information.

[0233] The conditional probability P(A = False|Evidence) that the network root node takes False when the network evidence set Evidence is known can be obtained by using Bayesian network modeling software.

[0234] Substitute the historical evaluation data set into the conditional probability association model, and perform training processing on the conditional probability association model to obtain a trained conditional probability association model, including:

[0235] The historical evaluation data set includes a collection data set and a corresponding cognitive impairment evaluation value set; the collection data set includes collection data; the collection data includes historical evaluation process information, historical eye movement data information, and historical blood oxygen information; the cognitive impairment evaluation value set includes cognitive impairment evaluation values corresponding to each collection data, which are also the label values corresponding to each collection data;

[0236] The historical assessment process information includes historical data on the results and probability values ​​of all the above factors, including whether classic methods were used, whether mature tools were used, whether supervised samples were used, whether the assessment data characteristics were appropriate, whether the assessment business characteristics were appropriate, whether the analysis of the assessment purpose was sufficient, whether the analysis of the assessment object was accurate, whether the assessment entity controlled the entire process, whether the assessment results were verified, whether the indicators were complete, whether the indicators were independent, whether the indicators were consistent, whether the indicators were objective, the credibility of the assessment process, and the technical level of the personnel.

[0237] Initialize the number of training iterations;

[0238] The collected data from the historical evaluation dataset is used as input data and input into the conditional probability association model;

[0239] The input data is processed using the conditional probability association model to obtain the predicted value;

[0240] The difference between the predicted value and the label information corresponding to the input data is calculated to obtain the difference value.

[0241] Determine whether the difference value satisfies the convergence condition to obtain the first determination result;

[0242] When the first judgment result is negative, it is determined whether the training iteration count value is equal to the training count threshold to obtain the second judgment result;

[0243] When the second judgment result is negative, the model training state is determined to be that the termination training condition is not met.

[0244] When the second judgment result is yes, it is determined that the model training state meets the termination training condition;

[0245] When the first judgment result is yes, it is determined that the model training state meets the termination training condition;

[0246] When the training state of the model does not meet the termination training condition, the parameter update model is used to update the parameters of the conditional probability association model, the training iteration count is increased by 1, and the collected data in the historical evaluation dataset is used as input data to input the conditional probability association model.

[0247] When the training state of the model meets the termination training condition, the training process of the conditional probability association model is completed, and the trained conditional probability association model is obtained.

[0248] The difference value satisfies the convergence condition when it is less than a preset convergence threshold; the difference value does not satisfy the convergence condition when it is not less than a preset convergence threshold.

[0249] The difference calculation process can be implemented using a loss function.

[0250] The loss function can be the cross-entropy loss function.

[0251] The parameter update model is as follows:

[0252]

[0253] θ←θ+v;

[0254] In the formula, Let v be the difference value calculated for the i-th collected data in the historical evaluation dataset, θ be the parameter update value, η be the initial parameter learning rate, and α be the momentum angle parameter, where 0 ≤ α ≤ π / 4. This indicates taking the partial derivative with respect to the variable θ;

[0255] The parameters to be updated in the conditional probability association model are the conditional probability values ​​between each node;

[0256] The normalization process maps data from different value ranges to a specified value range, and its mathematical expression is:

[0257]

[0258] Where t is a specific data value in the dataset, t max t represents the maximum value of all data in the dataset. min t is the minimum value of all data in the dataset. 归一化 The value of data t after normalization;

[0259] The data cleaning process includes filling in missing values, smoothing noisy data, and smoothing or deleting outliers. Smoothing noisy data involves first identifying the noisy data, and then smoothing it based on the data preceding and following it. The noisy data refers to values ​​whose values ​​are less than the sensor's detection sensitivity for the observed data, or greater than the sensor's measurement upper limit for the observed data. Outlier identification can be performed using a Kalman filter. The filling values ​​for missing values ​​can be determined by averaging the measurements within a certain sampling interval before and after the missing value.

[0260] The pattern recognition and discrimination process includes:

[0261] For each data attribute in the cleaned dataset, the data collection information is used as the known independent variable, and the data value is used as the known dependent variable. An approximation curve is constructed using the known independent and dependent variables. A function approximation method is used to fit the approximation curve to obtain the best uniform approximation polynomial f(Ix) for the data attribute. The known independent variable is then processed using the best uniform approximation polynomial f(Ix) to obtain an approximate dependent variable. The absolute value of the difference between the approximate dependent variable and the corresponding known dependent variable is determined to be greater than a set second regression threshold. If it is greater than the second regression threshold, the data is deleted from the cleaned dataset; if it is less than or equal to the second regression threshold, the data is not processed.

[0262] The data from all the execution feature pre-analysis of the cleaned data set are fused to obtain the corresponding pre-processed information set;

[0263] The curve fitting of the curve to be approximated using the function approximation method can employ the best uniform linear approximation method. The best uniform approximation polynomial f(Ix) is expressed as:

[0264] f(Ix)=α P1 (Ix) P1 +α P1-1 (Ix) P1-1 +…+α2(Ix) 2 +α1(Ix)+α0,

[0265] Where P1 is the order of the best uniform approximation polynomial f(Ix), α0, α1, α2, ..., α P1 The coefficients of the best uniform approximation polynomial f(Ix);

[0266] The process of performing anomaly detection on the acquired eye-tracking data set yields a first anomaly detection value, including...

[0267] An eye-tracking data matrix is ​​constructed using the aforementioned eye-tracking data set; the row vectors of the eye-tracking data matrix are sequences of fixation point information, fixation time information, and saccade frequency information.

[0268] A standard eye-tracking data sequence is obtained; the row vectors of the standard eye-tracking data sequence are the standard fixation point information sequence, the standard fixation time information sequence, and the standard saccade frequency information sequence.

[0269] The difference calculation process is performed on the standard eye-tracking data sequence and the eye-tracking data matrix to obtain the first anomaly discrimination value;

[0270] The difference calculation includes:

[0271] The difference matrix is ​​obtained by subtracting the standard eye-tracking data sequence and the eye-tracking data matrix.

[0272] Perform cross-correlation calculations on the difference matrix to obtain the cross-correlation matrix R; i,j Let r represent the element in the i-th row and j-th position of the cross-correlation matrix R. i,j The calculation expression is:

[0273] r i,j =E(e i e j ),

[0274] Where E() represents the mean, e i This represents the i-th row of the difference matrix;

[0275] For the cross-correlation matrix R, normalization is performed on each row vector to obtain a normalized correlation value sequence; the expression for the normalization calculation is:

[0276]

[0277] in, Let N represent the j-th value of the i-th normalized correlation value sequence, where N is the number of elements contained in the normalized correlation value sequence.

[0278] Logarithmic summation is performed on each normalized correlation value sequence to obtain the uncertainty value of each normalized correlation value sequence;

[0279] The calculation expression for the logarithmic summation process is as follows:

[0280]

[0281] Among them, H i Let be the uncertainty value of the i-th normalized correlation value sequence;

[0282] Angle normalization is performed on the uncertainty value of each normalized correlation value sequence to obtain the characteristic value of each normalized correlation value sequence;

[0283] The calculation expression for the angle normalization process is:

[0284]

[0285] in, Let be the eigenvalue of the i-th normalized correlation value sequence, and M be the total number of normalized correlation value sequences.

[0286] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A cognitive impairment screening device based on XR eye tracker and near-infrared brain imaging, characterized in that, include: The system includes an extended reality scene generation module, an eye-tracking data acquisition module, a near-infrared brain imaging module, a data analysis and processing module, and a cognitive impairment screening module. The extended reality scene generation module generates and displays virtual reality and augmented reality scenes containing cognitive tasks. The eye-tracking data acquisition module collects real-time eye movement data sets of the user while performing cognitive tasks in the virtual reality and augmented reality scenes. The eye movement data sets include a fixation point information sequence, a fixation time information sequence, and a saccade frequency information sequence. The fixation point information sequence includes fixation point location information. The fixation time information sequence includes… The system includes fixation time information; the saccade frequency information sequence, including saccade frequency information; the near-infrared brain imaging module, used to acquire in real time the set of blood oxygen information of the cerebral cortex when the user completes cognitive tasks in virtual reality and augmented reality scenes; the set of blood oxygen information of the cerebral cortex includes a sequence of blood oxygen information of the brain; the data analysis and processing module, connected to the augmented reality scene generation module, the eye tracker data acquisition module, the near-infrared brain imaging module and the cognitive impairment screening module respectively, is used to analyze and process the acquired set of eye movement data and the set of blood oxygen information of the cerebral cortex, and extract the first abnormal feature signal; The cognitive impairment screening module is used to perform cognitive impairment screening processing on assessment process information, eye movement data information set and cerebral cortex blood oxygenation information set to obtain the user's cognitive impairment information; The cognitive impairment information includes whether cognitive impairment exists and the degree of cognitive impairment.

2. The cognitive impairment screening device based on XR eye tracker and near-infrared brain imaging as described in claim 1, characterized in that, The data analysis and processing module is used to analyze and process the collected eye movement data information set and the cerebral cortex blood oxygen information set to extract a first abnormal feature signal, including: the data analysis and processing module performs anomaly discrimination processing on the collected eye movement data information set to obtain a first abnormal discrimination value; and determines whether the first abnormal discrimination value is greater than a preset first discrimination threshold to obtain the first abnormal feature signal. If the first abnormal feature signal is negative, the collected eye movement data set and the cerebral cortex blood oxygenation information set are sent to the cognitive impairment screening module; if the first abnormal feature signal is positive, the cognitive impairment information is confirmed to be present, and the degree of cognitive impairment information is determined as the first abnormal discrimination value.

3. The cognitive impairment screening device based on XR eye tracker and near-infrared brain imaging as described in claim 2, characterized in that, The extended reality scene generation module includes a virtual reality scene submodule and an augmented reality scene submodule; the virtual reality scene submodule is used to display a virtual reality scene containing cognitive tasks to the user; the augmented reality scene submodule is used to display an augmented reality scene containing cognitive tasks to the user.

4. The cognitive impairment screening device based on XR eye tracker and near-infrared brain imaging as described in claim 2, characterized in that, The cognitive impairment screening module is used to perform cognitive impairment screening processing on the eye movement data set and the cerebral cortex blood oxygenation data set to obtain the user's cognitive impairment information. This includes: the cognitive impairment screening module preprocessing the eye movement data set and the cerebral cortex blood oxygenation data set to obtain preprocessed eye movement data sets and cerebral cortex blood oxygenation data sets; and performing cognitive impairment assessment processing on the preprocessed eye movement data sets and cerebral cortex blood oxygenation data sets to obtain the user's cognitive impairment information.

Images