Devices, equipment, storage media and systems for screening and rehabilitation of patients with cognitive impairment

Abstract

This application relates to a screening and rehabilitation device, equipment, storage medium, and system for patients with cognitive impairment. The device includes: a first screening module for screening based on the target subject's AD8 scale score; terminating the screening when the score is greater than or equal to a preset threshold, thus identifying the target subject as a candidate AD patient; a second screening module for screening candidate AD patients and candidate VaD patients based on the target subject's scores in relevant tests; a third screening module for acquiring brain medical images for further screening; a training module for outputting training programs executable on a mobile terminal for each candidate AD patient and candidate VaD patient; and an adjustment module for acquiring the target subject's training performance and / or physiological parameters, and adjusting the difficulty of the current program, changing the training program, or terminating the training. The technical solution of this application offers fast and low-cost confirmation of AD or VaD.

Description

Technical Field

[0001] This application relates to the field of smart healthcare technology, and in particular to a screening and rehabilitation device, equipment, storage medium and system for patients with cognitive impairment. Background Technology

[0002] Alzheimer's disease (AD) is a degenerative disease of the central nervous system characterized by progressive cognitive impairment and behavioral damage, affecting memory, thinking, language, and daily functions. Vascular dementia (VaD) is a cognitive impairment syndrome caused by cerebrovascular diseases, including brain damage resulting from various cerebrovascular diseases such as cerebral infarction, cerebral hemorrhage, and cerebral arteriosclerosis. The main differences between the two are: In terms of pathogenesis, AD is a neurodegenerative disease related to β-amyloid protein deposition and neurofibrillary tangles in the brain; VaD is caused by brain damage due to cerebrovascular diseases. Regarding the characteristics of cognitive impairment, AD initially presents primarily with recent memory impairment, with cognitive function progressively declining; VaD initially presents primarily with executive function impairment, with cognitive impairment occurring suddenly or worsening in a stepwise manner. AD patients may experience personality changes, behavioral abnormalities, and other psychiatric symptoms; VaD patients often have focal neurological signs and gait abnormalities. In clinical differentiation, distinguishing between AD and VaD requires combining clinical features, imaging findings, biomarkers, and disease course characteristics.

[0003] In existing technologies, the verification of AD or VaD is time-consuming and costly. Summary of the Invention

[0004] In view of this, this application provides a screening and rehabilitation device, equipment, storage medium and system for patients with cognitive impairment in order to solve at least one of the problems existing in the prior art.

[0005] To achieve the above objectives, the technical solution of this application is implemented as follows:

[0006] In a first aspect, embodiments of this application provide a screening and rehabilitation device for patients with cognitive impairment, applicable to Alzheimer's disease (AD) and vascular dementia (VaD), the device comprising:

[0007] The first screening module is used to screen based on the target subject's AD8 scale score; if the score is greater than or equal to a preset threshold, the screening is terminated and the target subject is determined to be a candidate AD patient; otherwise, the screening continues.

[0008] The second screening module is used to screen out candidate AD patients based on the target subject's test scores in hippocampal related memory ability tests, and is also used to screen out candidate VaD patients based on the target subject's test scores in non-memory cognitive tests.

[0009] The third screening module is used to obtain brain medical images of target subjects who are both candidate AD patients and candidate VaD patients for further screening.

[0010] The training module is used to output training programs that can be executed on mobile terminals for candidate AD patients and candidate VaD patients, respectively.

[0011] An adjustment module is used to acquire the training performance of the target object and / or the physiological parameters of the target object, and adjust the training program according to the training performance and / or physiological parameters to adjust the difficulty of the current project, change the training project, or terminate the training.

[0012] In an optional implementation, the first screening module is further configured to:

[0013] Output an electronic version of the AD8 scale that can be executed on a mobile device;

[0014] Obtain the input responses to the AD8 scale to obtain the rating score.

[0015] In an optional implementation, the second screening module is further configured to:

[0016] Output a first test program for testing the hippocampal memory-related abilities of the target subject and a second test program for testing the non-memory-related cognitive abilities of the target subject;

[0017] The test scores of the first test program and the second test program are obtained; the first test program includes a similar image recognition program; the second test program includes a line-connecting program and a cerebrovascular risk questionnaire program.

[0018] In an optional implementation, the second screening module is further configured to:

[0019] For target subjects who are both candidate AD patients and candidate VaD patients, if the test score of the cerebrovascular risk questionnaire is less than a preset score threshold, the screening weight of the online procedure is increased.

[0020] In one alternative embodiment, the brain medical images include brain magnetic resonance images and / or brain CT images;

[0021] The third screening module is also used for:

[0022] The hippocampal volume or white matter lesion area can be determined by analyzing the acquired brain MRI and / or brain CT images.

[0023] The target individuals are further screened based on the hippocampal volume or the area of ​​white matter lesions.

[0024] In one alternative implementation, the training module includes:

[0025] A first training module is configured to output at least one of the following first-type training programs for the candidate AD patients:

[0026] Virtual scene memory training program;

[0027] Similar face memory training program;

[0028] Training program for identifying differences between similar animals;

[0029] Family photo album reconstruction training program;

[0030] The second training module is used to output at least one of the following second-type training programs for the candidate VaD patients:

[0031] Gait correction training program;

[0032] Virtual multitasking execution training program;

[0033] Rapid information processing training program;

[0034] Training program for motor coordination and decision-making speed.

[0035] In an optional implementation, the gait correction training procedure is used to:

[0036] Obtain the gait data of the target object;

[0037] Based on the gait data, abnormal features were identified;

[0038] Based on the aforementioned abnormal characteristics, a gamified training task is output.

[0039] In an optional implementation, acquiring the gait data of the target object includes:

[0040] Use the mobile device's camera to capture walking video of the target object;

[0041] Based on the walking video, the skeletal key points of the target object are identified to obtain the gait data of the target object; the gait data includes at least the toe-to-ground clearance and the heel-to-ground angle;

[0042] The step of identifying abnormal features based on the gait data includes:

[0043] Based on the toe-off-ground clearance and heel-to-ground angle, abnormal features are identified, and a fall risk assessment value for the target object is generated.

[0044] In an optional implementation, the step of outputting a gamified training task based on the anomaly features includes:

[0045] A sequence of virtual footprints synchronized with the real-time gait of the target object is overlaid on the screen of the mobile terminal; wherein the stride length, occurrence sequence and horizontal offset of the virtual footprint sequence are set after analysis based on the abnormal characteristics, so as to guide the target object to adjust its gait pattern during walking.

[0046] In one optional implementation, the training performance of the target object includes: the training record of the target object in the training program, and / or the facial expressions of the target object.

[0047] In an optional implementation, the adjustment module is further configured to:

[0048] Obtain physiological parameter values ​​measured by the wearable device of the target object;

[0049] Based on the physiological parameter values, adjust the difficulty of the current project, change the training program, or terminate the training.

[0050] In an optional implementation, the adjustment module is further configured to:

[0051] Obtain environmental information about the environment in which the target object is located;

[0052] Adjust the difficulty of the current project, change the training program, or terminate the training based on the environmental information.

[0053] In an optional implementation, the adjustment module is further configured to:

[0054] Acquire training records for multiple target objects and input them into a pre-defined reinforcement learning model for learning;

[0055] When the change in the reward function of the reinforcement learning model is less than the convergence threshold, the iterative training of the reinforcement learning model is stopped; the reinforcement learning model is used to adjust the generation rules of the training items based on the input training records.

[0056] Secondly, embodiments of this application provide a computing device, the computing device comprising: a storage component, a communication bus, and a processing component, wherein:

[0057] The storage component is used to store the operating program of the screening and rehabilitation device for patients with cognitive impairment;

[0058] The communication bus is used to enable communication between the storage component and the processing component;

[0059] The processing unit is used to execute the operating program of the cognitive impairment patient screening and rehabilitation device to realize the operation of each module in any of the cognitive impairment patient screening and rehabilitation devices as described above.

[0060] Thirdly, embodiments of this application provide a computer-readable storage medium on which an executable program is stored.

[0061] When the executable program is executed by the processor, it enables the operation of each module in any of the cognitive impairment patient screening and rehabilitation devices described above.

[0062] Fourthly, embodiments of this application provide a screening and rehabilitation system for patients with cognitive impairment, comprising:

[0063] The computing devices described above;

[0064] The first client is used to collect various screening data of the target object and send them to the computing device;

[0065] The second client is used to export data from the computing device for offline treatment of the target subject.

[0066] The cognitive impairment screening and rehabilitation device, equipment, storage medium, and system provided in this application embodiment achieve stratified screening through a first screening module and a second screening module, which is time-efficient and low-cost. The first screening module can be terminated early, further reducing screening time. A third screening module enables further screening of patients who are simultaneously candidate AD and candidate VaD, making the screening more accurate. A training module creates a closed loop between screening and rehabilitation, allowing the screening data to be used for rehabilitation training and improving rehabilitation outcomes. Adjusting the modules can achieve more effective training or a better user experience. Therefore, the cognitive impairment screening and rehabilitation device, equipment, storage medium, and system provided in this application embodiment can solve the problems of long confirmation times and high costs associated with AD or VaD.

[0067] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0068] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0069] Figure 1 This is a schematic diagram of the structure of the cognitive impairment patient screening and rehabilitation device provided in the embodiments of this application;

[0070] Figure 2 A schematic diagram of the brain medical image processing process in the cognitive impairment patient screening and rehabilitation device provided in the embodiments of this application;

[0071] Figure 3 A flowchart illustrating the learning process of the reinforcement learning model in the cognitive impairment patient screening and rehabilitation device provided in this application embodiment;

[0072] Figure 4 A schematic flowchart illustrating the execution process of the cognitive impairment patient screening and rehabilitation device provided in the embodiments of this application;

[0073] Figure 5 A detailed flowchart illustrating the execution process of the cognitive impairment patient screening and rehabilitation device provided in this application embodiment;

[0074] Figure 6 This is a schematic diagram of the structure of a computing device provided in an embodiment of this application;

[0075] Figure 7 This is a schematic diagram of a cognitive impairment patient screening and rehabilitation system provided in an embodiment of this application.

[0076] Explanation of reference numerals in the attached figures:

[0077] 10. Screening and rehabilitation device; 11. First screening module; 12. Second screening module; 13. Third screening module; 14. Training module; 15. Adjustment module; 50. Computing device; 51. Storage component; 52. Communication bus; 53. Processing component; 54. Input device; 55. Output device; 56. External communication interface; 611. Screening module; 612. Rehabilitation module; 62. First client; 63. Second client. Detailed Implementation

[0078] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the specific embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present application and to fully convey the scope of the disclosure of the present application to those skilled in the art.

[0079] The following description provides numerous specific details to offer a more thorough understanding of this application. However, it will be apparent to those skilled in the art that this application can be practiced without one or more of these details. In other instances, to avoid confusion with this application, some technical features well-known in the art have not been described; that is, not all features of actual embodiments are described herein, nor are well-known functions and structures described in detail.

[0080] To fully understand this application, detailed steps and structures will be presented in the following description to illustrate the technical solution of this application. Preferred embodiments of this application are described in detail below; however, in addition to these detailed descriptions, this application may have other implementation methods.

[0081] The inventors of this application discovered during the research and development process that current screening and rehabilitation devices and methods for patients with cognitive impairment have the following problems:

[0082] Time-consuming: Current testing methods, such as the Mini-Mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA), are characterized by high level of specialization, complex operation, and long processing time, making it difficult to achieve large-scale screening.

[0083] Inaccurate: Single-method approaches (such as questionnaires alone) are prone to misdiagnosing mixed-type dementia.

[0084] High costs: Specialized imaging equipment, such as positron emission tomography (PET) and cerebrospinal fluid testing equipment, is difficult to make available at the grassroots level. Blood biomarkers require specialized laboratory support, which may not be covered by primary community healthcare.

[0085] Therefore, the inventors of this application have conducted extensive research and development and proposed the following technical solution.

[0086] Example 1

[0087] This application provides a screening and rehabilitation device for patients with cognitive impairment (hereinafter referred to as screening and rehabilitation device 10 or device), see reference Figure 1 The device includes:

[0088] The first screening module 11 is used to screen based on the target subject's AD8 scale score; if the score is greater than or equal to a preset threshold, the screening is terminated and the target subject is determined to be a candidate AD patient; otherwise, the screening continues.

[0089] The second screening module 12 is used to screen out candidate AD patients based on the target subject's test score in a hippocampal memory ability test, and is also used to screen out candidate VaD patients based on the target subject's test score in a non-memory cognitive test.

[0090] The third screening module 13 is used to obtain brain medical images of target subjects who are both candidate AD patients and candidate VaD patients for further screening.

[0091] Training module 14 is used to output training programs that can be executed on mobile terminals for candidate AD patients and candidate VaD patients, respectively.

[0092] The adjustment module 15 is used to obtain the training performance of the target object and / or the physiological parameters of the target object, and adjust the training program according to the training performance and / or physiological parameters to adjust the difficulty of the current project, change the training project, or terminate the training.

[0093] Specifically, the target audience can be patients with cognitive impairment or suspected patients with cognitive impairment. In the following description, the target audience can be collectively referred to as patients or users, and those skilled in the art can make the distinction based on the specific circumstances.

[0094] Specifically, cognitive impairment can include Alzheimer's disease (AD), VaD, and mixed dementia. It is understood that cognitive impairment can also be caused by other factors, and the rehabilitation device of this application embodiment can also provide some rehabilitation assistance for other cognitive impairments. For a more precise description, those screened by the screening and rehabilitation device 10 of this application are referred to as candidate AD patients or candidate VaD patients, but for the sake of brevity, they are simply referred to as AD patients or VaD patients.

[0095] In the first screening module 11, the AD8 scale is short for the Ascertain Dementia 8 (AD8) questionnaire. The AD8 scale has high sensitivity; for example, the sensitivity of the AD8 scale can be as high as 84%, which helps to reduce false negatives.

[0096] Specifically, the AD8 scale contains eight questions about cognitive decline. Participants simply need to select "yes" or "no." A "yes" answer increases the score, with higher scores indicating a higher risk. More specifically, AD8 covers cognitive domains such as memory, orientation, executive function, and language ability. The AD8 scale is simple and easy to use, suitable for non-professionals, especially those with lower levels of education or those unable to cooperate with complex tests. However, completing the AD8 scale also has the drawback of being time-consuming; for example, it may take more than half an hour.

[0097] In this embodiment, to improve efficiency and save time, a preset assessment threshold of 2 points is set. That is, if the assessment score is ≥2 points, a high risk of AD is directly determined, and the current AD8 scale assessment is terminated. If the score is <2 points, the current AD8 scale assessment can continue, and subsequent screening can proceed. Specifically, the AD8 scale can achieve rapid initial screening, taking less than 20 seconds.

[0098] In the second screening module 12, the hippocampal related memory ability test is used to test the target subject's episodic memory ability and pattern separation ability. This ability is related to the human hippocampus and can be used to screen AD patients.

[0099] Non-memory cognitive tests can be used to assess a target individual's non-memory cognitive abilities, such as executive function. VaD patients typically exhibit more severe impairment in executive function compared to less severe impairment in memory. Therefore, executive function tests can be used to screen for VaD patients.

[0100] Understandably, high-risk AD patients can be directly screened through the first screening module 11. For those with lower risk or suspected risk, screening can be conducted together through the first screening module 11 and the second screening module 12. The second screening module uses a hippocampal-related memory ability test to screen AD patients. VaD patients can be excluded in the first screening module 11 and then screened through a non-memory cognitive test in the second screening module 12.

[0101] In the third screening module 13, there are relatively few target subjects who simultaneously belong to both candidate AD and candidate VaD patients, making misjudgment easier. Therefore, further screening using brain medical images can make the screening more accurate. The results of the third screening module 13 are threefold: First, the subject is classified as an AD patient, indicating a possible deviation in the results of the first and second screening modules (this is possible due to testing errors). Second, the subject is classified as a VaD patient, indicating a possible deviation in the results of the second screening module. Third, the subject is simultaneously classified as both a candidate AD and a candidate VaD patient, meaning one target subject has both conditions, although this is relatively rare. This indicates that the results of the first and second screening modules were correct.

[0102] Specifically, brain medical images can be uploaded by the target subject or their relatives through photos, or a URL of the brain medical image can be provided, which the third screening module 13 accesses to obtain the image. The third screening module 13 can also be used to perform image analysis on the acquired brain medical images to obtain information about the lesions in the target subject's brain.

[0103] In training module 14, targeted rehabilitation training can be provided separately for AD patients and VaD patients. Compared to methods without accurate screening, where AD patients and VaD patients are trained using the same approach, this embodiment of the application, by screening and distinguishing between AD patients and VaD patients, and then using different training programs for each, results in better training outcomes.

[0104] Furthermore, in the training module 14, the data obtained from the screening in the previous three screening modules can be directly used to form more targeted and personalized training programs, so that screening and rehabilitation form a closed loop and improve rehabilitation results.

[0105] Furthermore, the training program can be run on mobile devices, such as smartphones and tablets, for user convenience. Specifically, the training program can be a mobile application (APP) or a WeChat mini-program. The server for the training program can be located in the cloud, and the server can be equipped with an artificial intelligence (AI) engine.

[0106] In adjustment module 15, the difficulty of the current item in the training program can be adjusted based on the target subject's training performance. That is, the training program can automatically adjust the difficulty based on the target subject's training performance, without requiring manual adjustment (such as adjustments by a doctor). For example, it can start with a relatively low difficulty level, gradually increasing the difficulty based on training performance. Of course, it shouldn't be too easy at the beginning, otherwise the training effect will be poor. Therefore, a moderately low training difficulty can be used as the initial difficulty, rather than the lowest possible difficulty. If a special patient is encountered where the initial difficulty is too high and affects the training effect, the training difficulty can be automatically lowered.

[0107] Specifically, for patients with special circumstances, the training program can be changed or the training can be terminated to achieve more effective training results or a better user experience.

[0108] The cognitive impairment screening and rehabilitation device provided in this application embodiment achieves stratified screening through a first screening module 11 and a second screening module 12, which is time-efficient and low-cost. The first screening module 11 can be terminated early, further reducing screening time. A third screening module 13 further screens patients who are simultaneously candidate AD and candidate VaD, making the screening more accurate. A training module 14 creates a closed loop between screening and rehabilitation, allowing the screening data to be used for rehabilitation training and improving rehabilitation outcomes. Adjusting module 15 can achieve even more effective training or a better user experience.

[0109] In some embodiments of this application, the first screening module 11 is further configured to:

[0110] Output an electronic version of the AD8 scale that can be executed on a mobile device;

[0111] Obtain the input responses to the AD8 scale to obtain the rating score.

[0112] This allows for convenient and quick questionnaire completion of the AD8 scale. It's understandable that the electronic version can employ many more user-friendly testing methods, such as touchscreens and voice interaction.

[0113] In some embodiments of this application, the second screening module 12 is further configured to:

[0114] Output a first test program to test the hippocampal memory capacity of the target object and a second test program to test the executive capacity of the target object;

[0115] The test scores of the first test program and the second test program are obtained; the first test program includes a similar image recognition program; the second test program includes a line-connecting program and a cerebrovascular risk questionnaire program.

[0116] Understandingly, a testing program can also be called a testing game. It can output programs or games that test the cognitive abilities of a target subject, requiring the target subject to perform corresponding operations based on their cognitive abilities. The testing program includes a first testing program and a second testing program. Compared to the fixed problem output by AD8 mentioned above, the testing program adjusts the difficulty level of the test according to the target subject's operation to obtain more accurate test results; the training program in training module 14 below is similar. Compared to traditional rehabilitation training, such as the mechanical repetition of paper-and-pen tasks, patient compliance is poor (completion rate <50%). The completion rate in this embodiment is increased to 85%, and the training program in training module 14 below is similar.

[0117] Specifically, similar images in a similar image recognition program can include images with similar content, or images with the same content but different ages. The similar images can be up to five images of everyday objects, with each image displayed for up to three seconds.

[0118] More specifically, it can provide 10 images (3 old, 3 similar distractors, and 4 new), and the operation for the target object includes the option to click confirm, which can include "same", "similar", "new", etc.

[0119] The scoring rules for the test are: correctly rejecting similar distractors (15 points per item, 45 points in total) + correctly identifying old items (10 points per item, 30 points in total). Here, the higher the score, the lower the risk.

[0120] It should be noted that the similar image recognition program primarily tests the pattern separation ability of the target object, which is closely related to the function of the dentate gyrus of the hippocampus and is of great significance for the early diagnosis of AD. Understandably, the similar image recognition program can also be used to identify other patients with cognitive impairment. Specifically, the AD risk assessment criteria can be found in Table 1:

[0121] Table 1

[0122]

[0123] Specifically, the online procedure and the cerebrovascular risk questionnaire procedure can be:

[0124] The numbers 1-5 are randomly distributed, and the target objects are connected in order of size (time limit 15 seconds).

[0125] Scoring: 30 points for time ≤ 8 seconds, 15 points for time 8-15 seconds, 0 points for timeout.

[0126] It should be noted that the online program mainly tests the target object's execution function and information processing speed, which is related to the function of the fronto-subcortical pathway and has high sensitivity for screening VaD.

[0127] Specifically, the trail-making procedure in this application embodiment is further simplified based on the existing trail-making procedure, making it more suitable for patients with cognitive impairment and accelerating the screening process. For example, the original trail-making procedure was the Trail Making Test (TMT), which was divided into two parts, A and B. The specific content of TMT-A was as follows:

[0128] Test material: A piece of paper with numbers 1 to 25 randomly distributed on it.

[0129] Test requirements: Participants need to connect the numbers in order using a pen.

[0130] Record metrics: time required to complete the test and number of errors. Completion time reflects the subject's information processing speed, while the number of errors may reflect the subject's attention maintenance ability and accuracy. This application's embodiment simplifies TMT-A to numbers 1-5, accelerating the testing process.

[0131] It should be noted that, due to the simplified design of the relevant procedures in this application, the testing time for both the first and second testing procedures can be less than 50 seconds. Combined with the rapid initial screening using the AD8 scale, the total testing time for the two screening modules can be less than 90 seconds. This makes it suitable for large-scale screening scenarios at the grassroots level, such as community health checkups. Compared to the shortcomings of existing screening processes that are too time-consuming (e.g., MMSE requires 10-15 minutes) and difficult to popularize in primary healthcare, the embodiments of this application significantly improve upon these limitations.

[0132] Furthermore, for accurate screening, a cerebrovascular risk questionnaire can be administered. This questionnaire may include an investigation of risk factors such as hypertension, diabetes, and a history of stroke. Each item is worth 10 points, with an initial maximum score of 30 points. 10 points are deducted for each risk factor selected.

[0133] By implementing online procedures and cerebrovascular risk questionnaires, the screening of VaD patients can achieve a specificity of >80%.

[0134] Specifically, the VaD risk assessment criteria can be found in Table 2:

[0135] Table 2

[0136]

[0137] Specifically, the above-mentioned connection procedure can also be replaced by a clock drawing program. The clock drawing program requires the target object to draw a clock on the screen and mark a specified time, such as "0:10". This program comprehensively assesses the target object's visuospatial abilities, executive functions, and abstract thinking abilities, and is one of the internationally recognized screening tools for patients with cognitive impairment.

[0138] AD and VaD risk levels are determined separately and are not added together in the total score.

[0139] In some embodiments of this application, the second screening module 12 is further configured to:

[0140] For target subjects who are both candidate AD patients and candidate VaD patients, if the test score of the cerebrovascular risk questionnaire is less than a preset score threshold, the screening weight of the online procedure is increased.

[0141] Being classified as both a candidate for Alzheimer's disease (AD) and a candidate for VaD indicates a possible mixed-type dementia. However, this is not entirely certain, as mixed-type dementia is easily misdiagnosed. Therefore, the weight of executive function tests, specifically the connecting procedure, can be increased based on the patient's cerebrovascular risk factors. A preset score threshold of 10 can be used.

[0142] For example, in a cerebrovascular risk questionnaire, if a patient has a history of hypertension, diabetes, and stroke, and all three items are checked, the score is 0. Therefore, the cerebrovascular risk is very high, increasing the weight of the functional test. The score of the online test is multiplied by a weight of 0.8. For example, if a patient's test time is 12 seconds (belonging to the 8-15 second range), they should score 15 points, but multiplying by the weight gives 12 points. This further increases the risk of being a candidate VaD patient. In other words, dynamic weighting improves the accuracy and speed of identification. The weighting can be automatically increased by the device, without manual setting.

[0143] In some embodiments of this application, the brain medical images include brain magnetic resonance images and / or brain CT images;

[0144] The third screening module 13 is also used for:

[0145] The hippocampal volume or white matter lesion area can be determined by analyzing the acquired brain MRI and / or brain CT images.

[0146] The target individuals are further screened based on the hippocampal volume or the area of ​​white matter lesions.

[0147] Magnetic resonance imaging (MRI) or CT scans can provide a clearer and more comprehensive view of brain structure to confirm brain health.

[0148] Specifically, artificial intelligence can be used to analyze MRI or CT images to obtain hippocampal volume or white matter lesion area. For example, images can be analyzed and segmented using a lightweight mobile neural network version 3 (MobileNetV3).

[0149] Specifically, the magnetic resonance or CT images can be in Digital Imaging and Communications in Medicine (DICOM) format with a resolution ≥1 mm³. Preprocessing includes denoising (non-local mean filtering) and standardization, which refers to the normalization of the standard score (Z-score).

[0150] The third screening module 13 can be used for segmentation of magnetic resonance images or CT images. The segmentation process includes:

[0151] Using a lightweight MobileNetV3 model with an input layer size of 224×224×3 (RGB channels), the output is a binary mask of the hippocampal region. The hippocampal region binary mask is a 3D matrix (or stacked 2D slices), where each voxel has a value of 0 (background) or 1 (hippocampal region). Understandably, before obtaining the hippocampal region binary mask, it is necessary to obtain the hippocampal ROI (Region of Interest), and then extract the hippocampal boundaries from the ROI using segmentation algorithms (such as thresholding or deep learning) to achieve accurate separation of the hippocampus from neighboring structures, thus obtaining the hippocampal region binary mask.

[0152] Hippocampal volume calculation: Mask voxels number × voxel volume (e.g., 1 mm³), with the result displayed as a percentage (compared to a database of healthy individuals of the same age). The mask voxels number is the total number of voxels with a value of 1 in the statistical mask, representing the number of voxels occupied by the hippocampal region in the image. The voxel volume is the actual size of each voxel in physical space, determined by the image resolution. For example, if the image resolution is 1 mm × 1 mm × 1 mm (isotropic), then the volume of each voxel is 1 mm³. If the resolution is 0.5 mm × 0.5 mm × 2 mm, then the voxel volume is 0.5 × 0.5 × 2 = 0.5 mm³.

[0153] refer to Figure 2 The specific process of brain medical image processing can include: image input, preprocessing, hippocampal ROI extraction, hippocampal volume calculation, and result output.

[0154] Understandably, AD patients can be identified through hippocampal volume, while VaD patients can be identified through white matter lesion area. Therefore, obtaining hippocampal volume or white matter lesion area can significantly improve the accuracy of identification. By quantifying key indicators such as hippocampal volume and white matter lesion area, the accuracy rate of AD / VaD differentiation has increased from 78% to 89%. AI-automated interpretation compensates for the shortage of professional doctors, generating reports within 5 minutes; and it provides imaging evidence for rehabilitation effect assessment.

[0155] The MobileNetV3 lightweight model reduces image analysis costs by 90%, enabling primary care physicians to independently perform initial screening for Alzheimer's disease (AD) / VaD. It also allows for the understanding and acquisition of other structural features in the brain that can help confirm AD or VaD. Furthermore, it improves the sensitivity of disease progression monitoring (e.g., a 2% warning threshold for hippocampal atrophy rate) and allows for dynamic optimization of intervention plans.

[0156] In addition, when obtaining MRI images is difficult, fundus photographs can be used instead of MRI images of the brain. This is because changes in the retinal vessels in the fundus are closely related to VaD (vadilation of blood vessels), and therefore, fundus photographs can be used in place of MRI images of the brain.

[0157] Finally, the test scores from the first and second test programs are obtained and then combined for a final judgment. During the final judgment, the scores need to be adjusted according to age and education level.

[0158] ≥70 years old: The AD total score threshold is reduced by 5 points.

[0159] For students with primary school education or below: the online test time threshold is relaxed by 20%.

[0160] Next, with reference to specific embodiments, the cognitive impairment patient screening and rehabilitation device provided in this application will be further described.

[0161] Case A (suspected AD):

[0162] AD module: 15 points for similar interference items + 20 points for old items = 35 points (high risk).

[0163] VaD module: 30 points for connectivity test + 30 points for risk factors = 60 points (low risk).

[0164] Determination: High risk for Alzheimer's disease (AD). An Aβ-PET scan is recommended. (An Aβ-PET scan is a non-invasive, visual method that uses positron emission tomography (PET) to detect the deposition of β-amyloid protein (Aβ) in the brain using a radiolabeled specific imaging agent. It is considered the "gold standard" for early diagnosis of Alzheimer's disease (AD).)

[0165] Case B (suspected VaD):

[0166] AD module: 45 points for similar interference items + 20 points for old items = 65 points (low risk).

[0167] VaD module: 15 points for connection test + 10 points for risk factors = 25 points (high risk).

[0168] Assessment: High risk of VaD, cerebrovascular evaluation recommended.

[0169] Case C (Hybrid):

[0170] AD module: 30 points for similar interference items + 20 points for old items = 50 points (medium risk).

[0171] VaD module: Connection test score 0 + risk factor score 10 = 10 (high risk).

[0172] Assessment: High risk of VaD with moderate risk of AD; comprehensive evaluation recommended.

[0173] Case D (Hybrid):

[0174] AD module: 45 points for similar interference items + 10 points for old items = 55 points (medium risk).

[0175] VaD module: 15 points for connectivity test + 20 points for risk factors = 35 points (medium risk).

[0176] MRI images showed hippocampal atrophy, but no obvious cerebrovascular lesions.

[0177] Assessment: VaD of medium risk with AD of high risk, a comprehensive evaluation is recommended.

[0178] Case E (Boundary Conditions):

[0179] AD module: Episode memory 50 points (medium risk).

[0180] VaD module: 25 points for connectivity test (high risk) + 20 points for risk factors (1 stroke history).

[0181] Imaging findings: MRI showed a lacunar lesion in the right basal ganglia, with normal hippocampal volume.

[0182] Assessment: High risk of VaD, cerebrovascular evaluation recommended.

[0183] In some embodiments of this application, the training module includes:

[0184] A first training module is configured to output at least one of the following first-type training programs for the candidate AD patients:

[0185] Virtual scene memory training program;

[0186] Similar face memory training program;

[0187] Training program for identifying differences between similar animals;

[0188] Family photo album reconstruction training program.

[0189] Here, the first type of cognitive impairment patient can be Alzheimer's disease (AD). Compared to VaD, AD causes relatively more severe impairment in memory, language, and other areas, so the training program focuses more on memory and related areas.

[0190] Specifically, the parameters of the first type of training procedure, such as difficulty or training intensity, can be determined by referring to the measurement results of the AD8 scale in the first screening module 11. This is determined by the program and does not require manual determination. Of course, it can also be generated based on the learning results of the subsequent reinforcement learning model.

[0191] Understandably, the virtual scene can be Virtual Reality (VR) or Augmented Reality (AR). Understandably, the virtual scene memorization program primarily activates the hippocampus in the human brain: for example, the virtual scene could be a supermarket, where the target audience needs to memorize the locations of 10 items (e.g., apples in the produce section, milk in the refrigerated section), repeat the process after 5 minutes, and complete the checkout. The difficulty can be adjusted by increasing or decreasing the number of items (±3 items) or shortening the memorization time (±1 minute) based on performance.

[0192] Understandably, a similar face memory program is a form of pattern separation training. Specifically, it displays highly similar face images, the target subject memorizes them, and then determines whether a newly appearing face has appeared before. Difficulty adjustments are made by increasing or decreasing the number of images displayed or the number of images shown.

[0193] Similarly, the similar animal difference recognition program is also a type of pattern separation training. Specifically, the similar animal difference recognition program can display highly similar animal pictures (such as different breeds of dogs), and the target object needs to click on the differences in details (such as ear shape, fur color) and recall the classification. When adjusting the difficulty, the number of similar items can be increased (e.g., from 3 to 5) or the display time can be shortened (from 5 seconds to 3 seconds).

[0194] The family album reconstruction program is used to activate autobiographical memories. Specifically, the program can upload old family photos of the target individual, generate a timeline using AI, arrange them chronologically by event, and describe details (e.g., "This is my son's wedding; you're wearing a red cheongsam"). Autobiographical memory training can slow down the degradation of the Default Mode Network (DMN).

[0195] In some embodiments of this application, the training module 14 further includes:

[0196] The second training module is used to output at least one of the following second-type training programs for the candidate VaD patients:

[0197] Gait correction training program;

[0198] Virtual multitasking execution training program;

[0199] Rapid information processing training program;

[0200] Training program for motor coordination and decision-making speed.

[0201] Here, the second type of cognitive impairment patient can be VaD. Compared to AD, VaD has relatively more severe impairment in executive functions, so the training program is more focused on executive functions.

[0202] Similarly, the parameters of the second type of training program, such as difficulty or training intensity, can also be determined by referring to the data in the previous screening module, and will not be described in detail here.

[0203] In some embodiments of this application, the gait correction training procedure is used for:

[0204] Obtain the gait data of the target object;

[0205] Based on the gait data, abnormal features were identified;

[0206] Based on the aforementioned abnormal characteristics, a gamified training task is output.

[0207] Specifically, obtaining gait data of a target object can include:

[0208] The target object's current walking video is obtained through the mobile terminal's camera.

[0209] In some embodiments of this application, obtaining the gait data of the target object includes:

[0210] Use the mobile device's camera to capture walking video of the target object;

[0211] Based on the walking video, the skeletal key points of the target object are identified to obtain the gait data of the target object; the gait data includes at least the toe-to-ground clearance and the heel-to-ground angle;

[0212] The step of identifying abnormal features based on the gait data includes:

[0213] Based on the toe-off-ground clearance and heel-to-ground angle, abnormal features are identified, and a fall risk assessment value for the target object is generated.

[0214] The current walking video of the target object includes: acquiring video of the target object walking in a straight line (natural arm swing) for 3-5 meters (round trip), which can be repeated three times for more accurate data acquisition. The acquisition method includes setting a camera with its height aligned with the target object's waist to collect gait data. The camera can be a mobile phone or tablet camera.

[0215] After acquiring the gait data of the target individual, it is compared with normal gait data to identify abnormal features. For gait data including toe-to-ground clearance and heel-to-ground angle, a fall risk assessment value can also be generated for the target individual. For example, an excessively small toe-to-ground clearance or an abnormally large heel-to-ground angle indicates a high risk of fall. For instance, in VaD patients, due to decreased strength of the gastrocnemius muscle contraction and abnormal nerve control in the posterior calf, this clearance is significantly reduced, making it a key risk indicator for tripping. Furthermore, VaD patients often exhibit a "slapping" landing (reduced heel-to-ground angle) or a flat landing, which also easily leads to gait instability.

[0216] Compared to traditional clinical observation or wearable devices, which cannot conveniently measure these subtle yet crucial biomechanical parameters, this method achieves, for the first time, automated home monitoring of these parameters under label-free and contactless conditions, providing unprecedented insights into fall risk. This also allows for the generation of more targeted training programs.

[0217] Specifically, the gait data may also include at least one of the following:

[0218] Stride length, stride width, cadence, percentage of time spent on both feet, gait cycle, knee flexion angle, and trunk tilt angle.

[0219] The following examples illustrate gait data such as stride length, stride width, and the percentage of time spent supporting each foot. Please refer to Table 3 for details.

[0220] Table 3

[0221]

[0222] Specifically, gait data of the target object can be obtained through a human posture estimation model. This involves analyzing the recorded video using the model to output the 2D coordinates of the hip, knee, and ankle joints; calculating stride length, stride width, cadence, bipedal support time ratio, toe-to-ground clearance, and heel-to-ground angle; identifying the gait cycle (ground contact phase and swing phase) through the hip joint motion trajectory; and extracting abnormal data from these gait data to obtain abnormal features.

[0223] In addition, gait correction training programs are also used for:

[0224] Construct a lower limb kinetic chain model and calculate the knee flexion angle (normal walking: 60°±5°).

[0225] Detect trunk tilt angle (abnormal threshold: >5° and lasting for more than 3 steps).

[0226] AR overlays are displayed on the camera screen to mark abnormal areas (such as legs that do not move sufficiently, highlighted in red).

[0227] Understandably, these steps can also be performed using human pose estimation models.

[0228] Furthermore, voice prompts can be provided for abnormalities during training, such as outputting voice messages like "Left leg swing is insufficient, try raising your knee" or "Stride length is less than 50cm, please increase your stride length."

[0229] Specifically, the aforementioned human pose estimation model can be a lightweight MobileNetV3 model.

[0230] In some embodiments of this application, the step of outputting a gamified training task based on the abnormal features includes:

[0231] A sequence of virtual footprints synchronized with the real-time gait of the target object is overlaid on the screen of the mobile terminal; wherein the stride length, occurrence sequence and horizontal offset of the virtual footprint sequence are set after analysis based on the abnormal characteristics, so as to guide the target object to adjust its gait pattern during walking.

[0232] Therefore, the feedback from training module 14 is not a simple voice command like "faster" or "slower," but rather an overlay of a dynamically moving sequence of virtual footprints onto the real-time camera feed using AR technology. This sequence has the following characteristics: its spatiotemporal parameters are adjustable, and the position (controlling stride length), timing (controlling stride frequency), and horizontal offset of each footprint can all be dynamically adjusted.

[0233] The virtual footprint sequence can adaptively shift: when an asymmetry is detected in the user (such as a smaller left stride), the virtual footprint sequence will automatically shift to the left, using the user's visual-motor coordination instinct to induce them to take a larger left step to land on the footprint, thereby achieving active correction.

[0234] Therefore, this is a "guided" rather than "command-based" rehabilitation intervention. It utilizes the inherent "visual tracking" and "gait adaptation" instincts of the human nervous system, resulting in a more natural and profound training effect, representing a completely new interactive paradigm for digital therapy.

[0235] Furthermore, based on the aforementioned anomaly characteristics, the output gamified training task may also include:

[0236] Virtual obstacles appear randomly on the screen, with a height of 10cm-20cm. The target must lift their leg to step over them and maintain their balance.

[0237] Difficulty Adjustment: Increase the difficulty by completing the task correctly three times in a row, for example, by increasing the stride length by 5cm or raising the obstacle height by 2cm. Decrease the difficulty by failing twice in a row, for example, by decreasing the stride length by 5cm.

[0238] Understandably, a virtual multi-task execution training program is a type of execution function training, which may specifically include:

[0239] Simulate a kitchen scenario where the target user must simultaneously operate a cooking timer (ringing every 30 seconds), answer a virtual phone call to record a shopping list, and prevent the cookware from burning. Adjustments to the difficulty can be made by increasing the number of concurrent tasks, such as adding a "plate after turning off the heat" step or shortening the task interval.

[0240] Specifically, a rapid information processing training program may include:

[0241] The program outputs randomly flashing symbol-number combinations to the screen, along with training rules. For example, it requires the target object to be classified according to the rules within one second. Classification can be achieved by pressing different buttons, such as "left-click for red shapes, right-click for odd numbers." The symbol-number combinations can include colored shapes, such as ▲-3, ●-5, etc.

[0242] Specifically, a training program for motor coordination and decision-making speed may include:

[0243] The program outputs a virtual scene of a virtual character navigating a maze. The target user is required to control the virtual character, avoiding moving obstacles and quickly choosing the correct path to reach the finish line. The target user can control the virtual character by tilting and moving their body. The motion coordination and decision-making speed training program can use a camera, such as a mobile phone camera, to capture the target user's body movements and enable control of the virtual character. Difficulty adjustments can be made by increasing the speed of obstacle movement or increasing the maze's complexity (e.g., multi-level structures).

[0244] The following table lists the defects of the target objects corresponding to various training programs. Please refer to Table 4 for details.

[0245] Table 4

[0246]

[0247] In some embodiments of this application, the training performance of the target object includes: the training record of the target object in the training program, and / or the facial expressions of the target object. This allows for timely adjustment of the difficulty of the current project based on the target object's training record, achieving better training results. The training record may include the patient's test accuracy, reaction time, etc.

[0248] For patients with special circumstances, such as those whose training is still unsatisfactory after adjusting the difficulty level, it is necessary to consider the patient's emotions and modify the training program accordingly. This is because, based on the neuropsychological theory of "learned helplessness," patients with cognitive impairment are highly susceptible to giving up due to failure and thus engaging in passive training. The core innovation of this solution lies in elevating "user experience" and "emotional management" to the same level of importance as "cognitive training," proactively preventing users from falling into frustration by adjusting the difficulty level—a feature not considered in conventional rehabilitation software.

[0249] Specifically, traditional rehabilitation training only displays an "Error" message when a user fails. This system employs a composite response mechanism:

[0250] First failure: The system reduces the difficulty of the project (e.g., reduces the number of distracting items).

[0251] A second failure: Instead of frustrating the user again on the same task, the system intelligently switches to a task type that the user previously performed well on (e.g., switching from a memory task to a visuospatial task), allowing the user to experience a sense of accomplishment and control, thus maintaining training motivation.

[0252] Specifically, the patient's facial expressions can be captured using the mobile device's camera, and their emotions can be determined based on these expressions. If the patient's emotions are characterized by frustration or irritability, the training program may be modified or the test may be terminated.

[0253] Specifically, the training items in the device are not from a pre-set, fixed question bank, but are generated in real-time based on user errors. This is a self-evolving, highly personalized training system. It is no longer a static "trainer," but an "intelligent coach" capable of discovering unique cognitive deficiencies in users and generating targeted "remedies," thanks to powerful backend algorithms and content generation capabilities. For example, the device incorporates many basic elements, such as basic images, which have only a shape feature and no other characteristics. These basic images can be customized with various features, such as color, brightness, background, atmosphere, etc. Based on the patient's condition, corresponding features can be added to generate test items tailored to that patient, making the test items more personalized. Direct effects: Training completion rate increased to 85% (enhanced user engagement), and cognitive function improvement rate increased by 30% (e.g., improved episodic memory scores).

[0254] In some embodiments of this application, the adjustment module 15 is further configured to:

[0255] Obtain physiological parameter values ​​measured by the wearable device of the target object;

[0256] Based on the physiological parameter values, adjust the difficulty of the current project, change the training program, or terminate the training.

[0257] Understandably, since the training programs in training module 14 require the patient's full engagement, it's necessary to monitor the patient's physiological parameters, such as heart rate and blood oxygen saturation, while recording the training data. These parameters can be obtained through the patient's wearable device, such as a wristband. If abnormal physiological parameters are detected, such as a heart rate exceeding 120 or blood oxygen saturation below 90%, training should be terminated immediately. This is because, on the one hand, continuing training may be less effective if the patient's physiological parameters are abnormal; on the other hand, continuing training might harm the patient's health.

[0258] Specifically, a patient's safe heart rate can be determined as (220 - age) * 70%. For example, if a patient is 50 years old, the safe heart rate is 119. Therefore, if the heart rate exceeds 119, training should be stopped immediately.

[0259] Furthermore, wearable devices can be used to integrate wearable data (heart rate, blood oxygen, etc.), rehabilitation training logs, and image data to construct a full-cycle health profile for patients.

[0260] In some embodiments of this application, the adjustment module 15 is further configured to:

[0261] Obtain environmental information about the environment in which the target object is located;

[0262] Adjust the difficulty of the current project, change the training program, or terminate the training based on the environmental information.

[0263] Understandably, the adjustment module 15 can add environmental awareness functionality, such as obtaining relevant information about the target's environment (e.g., ambient noise, light intensity, social interaction, etc.) through interaction with the training mobile terminal (e.g., a mobile phone) or wearable device (e.g., a smart bracelet). The training program can then be adjusted based on this environmental information. For example, when excessive ambient noise affects the patient's attention, the difficulty of the training program can be automatically reduced or training can be paused until the environment improves.

[0264] Furthermore, the training program can be comprehensively adjusted based on environmental information, the patient's training performance, physiological parameter values, and other factors. Not all of these parameters are necessarily present; they can be used as a reference for adjustment when available, and adjustments can be made based on other information when not present.

[0265] In some embodiments of this application, the adjustment module 15 is further configured to:

[0266] Acquire training records for multiple target objects and input them into a pre-defined reinforcement learning model for learning;

[0267] When the change in the reward function of the reinforcement learning model is less than the convergence threshold, the iterative training of the reinforcement learning model is stopped; the reinforcement learning model is used to adjust the generation rules of the training items based on the input training records.

[0268] To make the rules for generating training programs more scientific and improve rehabilitation outcomes, a dynamic algorithm and reinforcement learning framework are introduced to establish an adaptive difficulty system.

[0269] The reinforcement learning framework, namely the reinforcement learning model in this embodiment, can adjust the generation rules of the training items according to the real-time performance of the target object (such as accuracy and reaction time).

[0270] Specifically, each learning state space in the reinforcement learning framework encompasses the target object's accuracy, average reaction time, and task type from the last five training iterations; the action space allows adjustment of the number of distractors (±1~3) and the task time limit (±2 seconds); the reward function comprehensively considers the difference between the current accuracy and the baseline accuracy, as well as the reciprocal of the reaction time. Based on a large number of patients' historical training logs, the system iteratively optimizes using the Proximal Policy Optimization (PPO) algorithm, stopping iteration when the reward function change is less than a convergence threshold, thereby achieving intelligent adaptive adjustment of task difficulty.

[0271] Specifically, the reward function can be referred to as the following expression (1):

[0272] reward=0.7×(current_accuracy-baseline_accuracy)+0.3×(1 / reaction_time) (1)

[0273] Where reward is the reward function, current_accuracy is the current accuracy, baseline_accuracy is the baseline accuracy, and reaction_time is the reaction time in seconds. The baseline accuracy can be set according to different target objects and is not limited; the calculation result of the expression ignores the unit and only uses the numerical value as the basis for iteration.

[0274] Training data: Based on the historical training logs of 1000 patients, the PPO algorithm iterated 1000 times, and the convergence threshold Δreward < 0.01.

[0275] Example: If the accuracy rate of the episodic memory task is >90% for 3 consecutive times, then increase the number of distractors (e.g., from 10 items to 15 items).

[0276] Specifically, refer to Figure 3 The learning process of a reinforcement learning model can include:

[0277] It acquires training records, PPO algorithm decisions, model parameter updates, and feedback to the user.

[0278] In this context, the PPO algorithm decision is the input reinforcement learning framework, and the model parameter update includes the update of the item generation rule.

[0279] Specifically, the device can generate stage goals:

[0280] Generate personalized goals each week (e.g., "This week, focus on improving connection test speed to under 12 seconds").

[0281] Specifically, adjustment module 15 can also set up data-driven training optimization.

[0282] A. Multidimensional data dashboard, see Table 5.

[0283] Table 5

[0284]

[0285] B. Family doctor collaboration

[0286] Family-side APP:

[0287] Review the training report and receive suggestions (such as "Grandma Wang's recent memory has improved, so the complexity of kitchen tasks can be increased").

[0288] Doctor-side platform:

[0289] Export structured data, such as comma-separated values ​​(CSV) and PDF, and support comparative analysis with clinical scales (such as MoCA).

[0290] Furthermore, the training module 14 can introduce human-computer dialogue in the form of voice and set up a virtual AI coach. The following are simple examples using common phrases from the AI ​​coach:

[0291] For example, in a virtual scene memory program, an AI coach:

[0292] "Aunt Zhang, today we'll play a supermarket shopping game. Remember where the apples and milk are on the shelf!"

[0293] "Great! You've memorized 6 items, 2 more than last week! How about trying 8 tomorrow?"

[0294] In addition to task introductions and feedback, AI coaches can also provide emotional feedback:

[0295] For example, trigger encouraging voice messages (such as "Significant progress!") or adjust the difficulty prompts (such as "Shall we slow down and try again?") based on the patient's performance.

[0296] Furthermore, it can also acquire patients' facial expressions and, combined with facial expression recognition, automatically switch tasks when detecting patients' frustration.

[0297] To better understand the screening and rehabilitation device 10 provided in this application embodiment, the execution process of the screening and rehabilitation device 10 provided in this application embodiment will be described below. Figure 4 This is a schematic flowchart illustrating the execution process of the cognitive impairment patient screening and rehabilitation device provided in this application embodiment, with reference to... Figure 4 As shown, the execution process may include:

[0298] Step 201: Screening is conducted based on the target subject's AD8 scale score; if the score is greater than or equal to the preset assessment threshold, screening is terminated and the target subject is identified as a candidate AD patient; otherwise, screening continues.

[0299] Step 202: Based on the target subject's test scores in hippocampal memory ability tests, candidate AD patients are screened out. It is also used to screen out candidate VaD patients based on the target subject's test scores in non-memory cognitive tests.

[0300] Step 203: For target subjects who are both candidate AD patients and candidate VaD patients, obtain their brain medical images for further screening.

[0301] Step 204: For candidate AD patients and candidate VaD patients, output the corresponding training programs that can be executed on mobile terminals.

[0302] Step 205: Obtain the training performance of the target object and / or the physiological parameters of the target object, and adjust the training program according to the training performance and / or physiological parameters to adjust the difficulty of the current project, change the training project, or terminate the training.

[0303] Figure 5 This is a detailed flowchart illustrating the execution process of the cognitive impairment patient screening and rehabilitation device provided in the embodiments of this application. (Refer to...) Figure 5 As shown, the execution process may include:

[0304] Step 301: Obtain the AD8 scale.

[0305] Step 302: Is the score ≥ 2 points? If yes, proceed to step 307; otherwise, proceed to step 303.

[0306] Step 303: Obtain the image recognition test results. That is, the test results of the similar image recognition program.

[0307] Step 304: Obtain the test results of the connection program. That is, the test results of the program that identifies patterns in different items and connects them.

[0308] Step 305: Obtain the medical history questionnaire. This is the cerebrovascular risk questionnaire procedure. Sometimes, this step can precede step 304. It should be noted that step 306 is optional, meaning it is only performed on target subjects who are simultaneously candidate AD patients and candidate VaD patients. Therefore, if step 306 is skipped, this step can proceed directly to step 307. The wiring procedure can be replaced by a clock drawing procedure.

[0309] Step 306: Obtain medical images of the brain.

[0310] Step 307: Determine the patient type. Based on steps 302, 303, 304, and 305, the type of patient with cognitive impairment can be preliminarily determined. Step 306 then further clarifies the type to obtain more accurate results.

[0311] Step 308: Output the first type of training program. That is, for those determined to be AD, output the first type of training program.

[0312] Step 309: Output the second type of training program. That is, for those determined to be VaD, output the second type of training program.

[0313] The modules included in this embodiment can be implemented using a processor in a computer; alternatively, they can be implemented using logic circuits in a computer. The processor can be a general-purpose processor, a digital signal processor (DSP), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a central processing unit (CPU), a microprocessor (MPU), or any other conventional processor.

[0314] Example 2

[0315] This application provides a computing device 50, with reference to... Figure 6 The computing device 50 includes: a storage unit 51, a communication bus 52, and a processing unit 53, wherein:

[0316] The storage component 51 is used to store the operating program of the cognitive impairment patient screening and rehabilitation device;

[0317] The communication bus 52 is used to realize the connection and communication between the storage component 51 and the processing component 53.

[0318] The processing unit 53 is used to execute the operating program of the cognitive impairment patient screening and rehabilitation device to enable the operation of each module in the cognitive impairment patient screening and rehabilitation device.

[0319] Specifically, the operation of each module in the cognitive impairment screening and rehabilitation device can include, for example: Figure 4 The steps are shown.

[0320] Understandably, computing device 50 can be a specialized smart device, such as a medical device, or a general-purpose smart terminal, such as a mobile phone, tablet, or laptop, or even a cloud server. It can also be any other device capable of performing the aforementioned functions. In practical use, computing device 50 can be a cloud server. On the one hand, it can possess stronger computing power; on the other hand, existing public mobile communication networks, such as 5G networks, can provide sufficient data transmission capacity and shorter response times, ensuring the deployment of the cloud server. When computing device 50 is a cloud server, the user also needs to use a mobile phone or other smart terminal to collect relevant information about the target object through the camera of the mobile phone or other smart terminal, and then send the information to the cloud server through the communication device of the mobile phone or other smart terminal.

[0321] The type or structure of the storage component 51 can be found in the storage medium section below, and will not be repeated here.

[0322] The processing unit 53 can be a general-purpose processor, a digital signal processor (DSP), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor can be a central processing unit (CPU), a microprocessor (MPU), or any other conventional processor.

[0323] In some embodiments, the computing device 50 may further include an input device 54, an output device 55, and an external communication interface 56, which are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0324] In some embodiments, the input device 54 may include, for example, a keyboard, a mouse, a microphone, etc. If the computing device 50 is a smart terminal such as a mobile phone, tablet, or laptop, the input device 54 may also include a camera, etc.

[0325] The output device 55 can output various information to the outside, including displays, speakers, printers, projectors, communication networks and their connected remote output devices, etc.

[0326] The external communication interface 56 can be wired, such as a standard serial port (RS232), a General-Purpose Interface Bus (GPIB) interface, an Ethernet interface, or a Universal Serial Bus (USB) interface, or it can be wireless, such as wireless network communication technology (WiFi) or Bluetooth. If the computing device 50 is a smart terminal such as a mobile phone, tablet, or laptop, external communication can be achieved through public mobile communication networks, i.e., 4G or 5G networks.

[0327] The descriptions of the above device embodiments are similar to those of the above apparatus embodiments, and have similar beneficial effects. For technical details not disclosed in the embodiments of this application, please refer to the descriptions of the apparatus embodiments in this application for understanding.

[0328] Example 3

[0329] This application provides a computer-readable storage medium storing an executable program, which, when executed by a processor, enables the operation of various modules in a cognitive impairment patient screening and rehabilitation device.

[0330] Specifically, the operation of each module in the cognitive impairment screening and rehabilitation device can include, for example: Figure 4 The steps are shown.

[0331] Exemplary examples show that a computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A computer-readable storage medium is a tangible device capable of holding and storing instructions for use by an instruction execution device. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), flash memory, compact disc read-only memory (CD-ROM), digital versatile discs (DVDs), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combinations thereof.

[0332] The RAM includes: Static Random Access Memory (SRAM), Synchronous Static Random Access Memory (SSRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), SyncLink Dynamic Random Access Memory (SLDRAM), and Direct Rambus Random Access Memory (DRRAM).

[0333] The ROM includes: Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), and Electrically Erasable Programmable Read-Only Memory (EEPROM).

[0334] The description of the computer-readable storage medium embodiments above is similar to the description of the device embodiments above, and has similar beneficial effects. For technical details not disclosed in the embodiments of this application, please refer to the description of the device embodiments in this application for understanding.

[0335] Example 4

[0336] This application provides a screening and rehabilitation system for patients with cognitive impairment, referencing... Figure 7 The rehabilitation system includes:

[0337] The computing device 50 described in Example 2;

[0338] The first client 62 is used to collect various screening data of the target object and send them to the computing device;

[0339] The second client 63 is used to export data from the computing device for offline treatment of the target subject.

[0340] Specifically, the computing device 50 may include a screening module 611 and a rehabilitation module 612. The screening module 611 may include the first screening module 11, the second screening module 12, and the third screening module 13 from Embodiment 1. The rehabilitation module 612 may include the training module 14 and the adjustment module 15 from Embodiment 1. Various data collected by the first client 62 can be sent to the screening module 611. The training program for the rehabilitation module 612 can be sent to the first client 62.

[0341] Specifically, the first client 62 can be the client of the target object. For example, it can be a mobile phone, tablet, or laptop. In addition to being used by the patient, family members can also use the first client 62.

[0342] Specifically, the second client 63 can be a client used by the target doctor, such as an office computer or mobile phone. The doctor can use the second client 63 to provide feedback on the results of the examinations performed on the patient or professional judgments, in order to further improve the data collection.

[0343] Furthermore, the second client 63 can also be used to obtain the execution status of the screening module 611 and / or the rehabilitation module 612 in order to provide professional suggestions and improve the training program of the rehabilitation module 612.

[0344] Furthermore, the second client 63 can also be used to export data from the screening module 611 and / or the rehabilitation module 612 for offline treatment of the target subjects. For example, it can export structured data, such as CSV and PDF, supporting comparative analysis with clinical scales (such as MoCA).

[0345] Furthermore, for some high-risk patients identified during screening and training, information can be simultaneously sent to a second client 63 so that professional doctors can intervene in a timely manner.

[0346] The descriptions of the system embodiments above are similar to those of the device embodiments above, and have similar beneficial effects. For technical details not disclosed in the embodiments of this application, please refer to the descriptions of the device embodiments in this application for understanding.

[0347] It should be noted that the various embodiments provided in this application belong to the same concept; the technical features in the technical solutions described in each embodiment can be arbitrarily combined to form new embodiments without conflict.

[0348] The embodiments of this application may be systems, methods, and / or computer program products. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this application. The computer program product may be written in any combination of one or more programming languages ​​to perform operations of the embodiments of this application. Programming languages ​​include object-oriented programming languages ​​such as Java, C++, etc., and also conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code may be executed entirely on a user's computer, partially on a user's device, as a standalone software package, partially on a user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuits, such as programmable logic circuits, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), are personalized by utilizing state information of computer-readable program instructions. These electronic circuits can execute computer-readable program instructions to implement various aspects of this application.

[0349] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0350] Various aspects of this application are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0351] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0352] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0353] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple modules or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or modules can be electrical, mechanical, or other forms.

[0354] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0355] In addition, each functional module in the various embodiments of this application can be integrated into one processing module, or each functional module can be a separate module, or two or more functional modules can be integrated into one module; the integrated module can be implemented in hardware or in the form of hardware plus software functional modules.

[0356] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments.

[0357] Alternatively, if the integrated modules described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this application. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0358] In the above description, the terms "first, second, ..." are used only to distinguish similar objects and do not represent a specific order of objects. Understandably, "first, second, third" can be interchanged in a specific order or sequence where permitted.

[0359] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0360] In the embodiments described in this application, unless otherwise stated and limited, the term "connection" should be interpreted broadly. For example, it can be an electrical connection or a connection between two internal components. It can be a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above term according to the specific circumstances.

[0361] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. When used herein, the singular forms “a,” “an,” and “the” are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “comprising” and / or “including,” when used in this specification, identify the presence of said features, integers, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups. When used herein, the term “and / or” includes any and all combinations of the associated listed items. It should be understood that “an embodiment” or “some embodiments” as used throughout the specification means that a particular feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, “in an embodiment” or “in some embodiments” appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these particular features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the embodiment numbers are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0362] It should be understood that the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0363] It should be understood that the above embodiments are exemplary and are not intended to encompass all possible implementations of the technical solutions contained in this application. Various modifications and changes can be made to the above embodiments without departing from the scope of this application. Similarly, the various technical features of the above embodiments can be arbitrarily combined to form other embodiments of this application that may not be explicitly described. Therefore, the above embodiments merely illustrate several implementations of this application and do not limit the scope of protection of this patent application.

Claims

1. A screening and rehabilitation device for patients with cognitive impairment, applied to Alzheimer's disease (AD) and vascular dementia (VaD), characterized in that, The device includes: The first screening module is used to screen based on the target subject's AD8 scale score; if the score is greater than or equal to a preset threshold, the screening is terminated and the target subject is determined to be a candidate AD patient; otherwise, the screening continues; the test time of the AD8 scale is set to less than 20 seconds. The second screening module is used to screen out candidate AD patients based on the target subject's test scores in hippocampal memory ability tests, and also to screen out candidate VaD patients based on the target subject's test scores in non-memory cognitive tests; the test time of the second screening module is set to be less than 50 seconds; stratified screening is achieved through the first and second screening modules; The second screening module is also used for: Output a first test program for testing the hippocampal memory-related abilities of the target subject and a second test program for testing the non-memory-related cognitive abilities of the target subject; The test scores of the first test program and the second test program are obtained; the first test program includes a similar image recognition program; the second test program includes a line-connecting program and a cerebrovascular risk questionnaire program. For target subjects who are both candidate AD patients and candidate VaD patients, when the test score of the cerebrovascular risk questionnaire program is less than the preset score threshold, the screening weight of the connection program is increased. The third screening module is used to obtain brain medical images of target subjects who are both candidate AD patients and candidate VaD patients for further screening. The training module is used to output training programs executable on a mobile terminal for candidate AD patients and candidate VaD patients, respectively; wherein, for the candidate VaD patients, the output training program includes a gait correction training program; the gait correction training program is used for: Use the mobile device's camera to capture walking video of the target object; Based on the walking video, the skeletal key points of the target object are identified to obtain the gait data of the target object; the gait data includes at least the toe-to-ground clearance and the heel-to-ground angle; Based on the toe-off-ground clearance and heel-to-ground angle, abnormal features are identified, and a fall risk assessment value for the target object is generated. The adjustment module is used to obtain the training performance and physiological parameters of the target object, and adjust the training program according to the training performance and physiological parameters to adjust the difficulty of the current project, change the training project, or terminate the training.

2. The cognitive impairment screening and rehabilitation device according to claim 1, characterized in that, The first screening module is also used for: Output an electronic version of the AD8 scale that can be executed on a mobile device; Obtain the input responses to the AD8 scale to obtain the rating score.

3. The cognitive impairment screening and rehabilitation device according to claim 1, characterized in that, The brain medical images include brain magnetic resonance images and / or brain CT images; The third screening module is also used for: The hippocampal volume or white matter lesion area can be determined by analyzing the acquired brain MRI and / or brain CT images. The target individuals are further screened based on the hippocampal volume or the area of ​​white matter lesions.

4. The cognitive impairment screening and rehabilitation device according to claim 1, characterized in that, The training module also includes: A first training module is configured to output at least one of the following first-type training programs for the candidate AD patients: Virtual scene memory training program; Similar face memory training program; Training program for identifying differences between similar animals; Family photo album reconstruction training program; The second training module, for the candidate VaD patients, also outputs at least one of the following second type of training programs: Virtual multitasking execution training program; Rapid information processing training program; The training program for motor coordination and decision-making speed; the gait correction training program belongs to one of the second type of training programs.

5. The cognitive impairment screening and rehabilitation device according to claim 1, characterized in that, The gait correction training program is also used for: Based on the aforementioned abnormal characteristics, a gamified training task is output.

6. The cognitive impairment screening and rehabilitation device according to claim 5, characterized in that, The step of outputting a gamified training task based on the abnormal features includes: A sequence of virtual footprints synchronized with the real-time gait of the target object is overlaid on the screen of the mobile terminal; wherein the stride length, occurrence sequence and horizontal offset of the virtual footprint sequence are set after analysis based on the abnormal characteristics, so as to guide the target object to adjust its gait pattern during walking.

7. The cognitive impairment screening and rehabilitation device according to claim 1, characterized in that, The training performance of the target object includes: the training record of the target object in the training program, and / or the facial expressions of the target object.

8. The cognitive impairment screening and rehabilitation device according to claim 1, characterized in that, The adjustment module is also used for: Obtain physiological parameter values ​​measured by the wearable device of the target object; Based on the physiological parameter values, adjust the difficulty of the current project, change the training program, or terminate the training.

9. The cognitive impairment screening and rehabilitation device according to claim 1, characterized in that, The adjustment module is also used for: Obtain environmental information about the environment in which the target object is located; Adjust the difficulty of the current project, change the training program, or terminate the training based on the environmental information.

10. The cognitive impairment screening and rehabilitation device according to claim 5, characterized in that, The adjustment module is also used for: Acquire training records for multiple target objects and input them into a pre-defined reinforcement learning model for learning; When the change in the reward function of the reinforcement learning model is less than the convergence threshold, the iterative training of the reinforcement learning model is stopped. The reinforcement learning model is used to adjust the generation rules of the training items based on the input training records.

11. A computing device, characterized in that, The computing device includes: a storage component, a communication bus, and a processing component, wherein: The storage component is used to store the operating program of the screening and rehabilitation device for patients with cognitive impairment; The communication bus is used to enable communication between the storage component and the processing component; The processing unit is used to execute the operating program of the cognitive impairment patient screening and rehabilitation device to enable the operation of each module in the cognitive impairment patient screening and rehabilitation device according to any one of claims 1-10.

12. A computer-readable storage medium, characterized in that, An executable program is stored on the computer-readable storage medium. When the executable program is executed by the processor, it enables the operation of each module in the cognitive impairment patient screening and rehabilitation device according to any one of claims 1-10.

13. A screening and rehabilitation system for patients with cognitive impairment, characterized in that, include: The computing device of claim 11; The first client is used to collect various screening data of the target object and send them to the computing device; The second client is used to export data from the computing device for offline treatment of the target subject.

Images