Cognitive function evaluation system for neurology department
By generating assessment tasks, acquiring user interaction datasets, extracting multimodal features, constructing personalized digital reports, predicting the risk of cognitive decline, and automatically generating intervention plans, the system solves the problems of subjective bias and lack of dynamic prediction in traditional systems, and achieves personalized intervention and accurate assessment.
Patent Information
- Application Number
- CN202511812835.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-04
- Publication Date
- 2026-01-27
AI Technical Summary
Traditional cognitive function assessment systems rely on subjective assessment results and lack dynamic predictive capabilities, leading to assessment results that deviate from reality. Furthermore, they lack personalized intervention plans, affecting early warning and treatment outcomes.
By generating assessment tasks, acquiring user interaction datasets, extracting multimodal features, constructing personalized digital reports, using artificial intelligence models to predict the risk of cognitive decline, automatically generating personalized intervention plans, and dynamically adjusting cognitive tasks.
It enables objective quantitative assessment, dynamic prediction of cognitive changes, and provision of personalized intervention plans, thereby improving the accuracy of assessment and the targeting of treatment, and extending the treatment time.
Smart Images
Figure CN121400780A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of cognitive function assessment technology, and more specifically, to a cognitive function assessment system for neurology. Background Technology
[0002] Cognitive function assessment systems are important tools in neurology for screening, diagnosing and monitoring cognitive impairment caused by various neurological diseases. They are mainly used to quantitatively assess patients' cognitive domains such as memory, executive function, attention and language through a series of standardized tasks.
[0003] However, traditional cognitive function assessment systems often suffer from the following shortcomings in practice. First, they rely heavily on patients' subjective responses and doctors' on-site observations and scoring. This approach fails to consider the dynamic details of the patient's cognitive process, making the assessment results susceptible to influence from multiple factors such as doctor's experience, patient emotions, and education level, leading to deviations from reality. Second, traditional systems typically present assessments at a specific point in time, often using static scores. This static approach fails to reflect changes in the patient's cognitive function over time. The changing cognitive functions make it difficult for doctors to provide early warnings and implement interventions for patients. Thirdly, traditional systems mostly use general data as a reference for assessment reports to provide intervention plans for patients. However, such "generalized" and "one-size-fits-all" intervention plans are not customized based on the patient's own cognitive deficits and future risks, which easily leads to poor intervention results. In general, how to effectively solve the problems of subjective and one-sided assessment data, lack of dynamic prediction ability and lack of personalized intervention plans in traditional systems has become the problem that current cognitive function assessment systems used in neurology need to face and solve.
[0004] In view of this, the present invention proposes a cognitive function assessment system for neurology to solve the above problems. Summary of the Invention
[0005] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution, including: The raw interaction data acquisition module is used to generate evaluation tasks and acquire raw interaction datasets during user interaction. The raw interaction datasets include clock drawing test datasets, voice test data, and selection test datasets. Furthermore, the steps for generating the evaluation task and obtaining the raw interaction dataset during the user interaction process include: S1.1: Based on the task database and the evaluation report, select a set of cognitive tasks and output the specified cognitive tasks to the user interface in step S1.2; S1.2: The system activates the user interface corresponding to the cognitive task and records the interaction events when the user performs the cognitive task. The user interface includes a touch screen, a voice input tool, and multiple-choice options. When the patient performs the cognitive task of drawing a clock, the clock drawing test dataset is obtained by recording the second-by-second coordinate sequence of the click point, the click event, the start time stamp of the handwriting, and the end time stamp of the handwriting. When the patient performs a cognitive task that is a speech test, the audio recording tool is activated to record the user's audio and obtain speech test data. When a patient performs a cognitive task involving a selection test, the patient's option click events and reaction times are recorded to obtain a selection test dataset. Package the clock drawing test dataset, the voice test data, and the selection test dataset to obtain the original interactive dataset; S1.3: Integrate the behavior timestamps into the original interaction dataset and output them to the feature data integration and calculation module; The feature data integration and calculation module is used to preprocess the original interactive dataset and extract multimodal features to obtain a feature vector report. Furthermore, the steps of preprocessing the original interactive dataset and extracting multimodal features include: S2.1: Use a low-pass filter to denoise the second-by-second coordinate sequence of clicks in the original interactive dataset, and remove silent segments from the speech test data to obtain a preprocessed dataset; S2.2: Extract multimodal features based on the preprocessed dataset to obtain an initial feature vector report; The total drawing time feature data is obtained by subtracting the handwriting start time from the handwriting end timestamp. By substituting into the calculation formula: , obtained the Instantaneous velocity feature data of each sampling point ,in, For the first The coordinates of each sampling point For the first The timestamps corresponding to each sampling point; Based on instantaneous velocity feature data, average velocity feature data The calculation is performed using the following formula: ; in, This represents the total number of samples in the handwriting trajectory. Speed flow characteristic data is obtained by dividing the standard deviation of instantaneous speed characteristic data by the average speed characteristic data; Based on instantaneous velocity characteristic data, the proportion of pause time characteristic data The calculation is performed using the following formula: ; in, For indicator functions, For speed threshold, Total drawing time; Pack together the total time feature data, instantaneous velocity feature data, average velocity feature data, velocity smoothness feature data, and pause time ratio feature data to obtain an initial feature vector report; S2.3: Perform Z-socre normalization on the initial feature vector report and all other feature data to obtain the feature vector report; S2.4: Output the feature vector report to the personalized digital mapping construction module; The personalized digital mapping construction module is used to construct digital mapping content that reflects the patient's cognitive state based on the feature vector report, and obtain a personalized digital report. Furthermore, the steps for constructing a digital mapping reflecting the patient's cognitive state based on the feature vector report include: S3.1: Obtain the historical personalized digital report of the specified patient based on the database. If the historical personalized digital report exists in the database, proceed to step S3.2. If the historical personalized digital report does not exist in the database, proceed to step S3.3. S3.2: Integrate the feature vector reports into the historical personalized digital reports based on time sorting to obtain a new feature matrix, where each row represents an evaluation and each column represents a feature data; S3.3: Based on the new feature matrix in step S3.2, convert the feature vector report into the format of the new feature matrix; S3.4: The exponentially weighted moving average method is used to assign higher weights to each feature data in the new feature matrix to obtain a new personalized digital report; S3.5: Package the feature vector report and the new personalized digital report from step S3.3 to obtain the personalized digital report; S3.6: Store personalized digital reports in the database. When a personalized digital report is a feature vector report, create a record file for the specified patient. When a personalized digital report is a new personalized digital report, overwrite the previous new personalized digital report with the new personalized digital report. S3.7: Output personalized digital reports to the cognitive trajectory prediction module; The cognitive trajectory prediction module is used to predict the patient's future cognitive decline risk based on personalized digital reports, and to obtain a cognitive risk prediction report. Furthermore, the steps for predicting a patient's future risk of cognitive decline based on personalized digital reports include: S4.1: Load the pre-trained cognitive trajectory prediction model; S4.2: Input personalized digital reports into the cognitive trajectory prediction model; S4.3: The cognitive trajectory prediction model outputs the cognitive prediction score for the Kth time point in the future; At the same time, the model outputs risk probability data; S4.4: Package the cognitive prediction score and risk probability data to obtain a cognitive risk prediction report; S4.5: Output the cognitive risk prediction report to the intervention plan matching and generation module; The intervention program matching and generation module is used to match intervention programs based on the cognitive risk prediction report and obtain a personalized intervention program report. Furthermore, the steps for matching intervention programs based on cognitive risk prediction reports include: S5.1: Based on the risk threshold, when the risk probability data is greater than or equal to the risk threshold, a high-intensity intervention instruction is generated; when the risk probability data is less than the risk threshold, a low-intensity intervention instruction is generated. S5.2: Identify the feature data that declines the fastest in the cognitive prediction score curve and generate an intervention report for the specified feature data; S5.3: Based on high-intensity intervention instructions, low-intensity intervention instructions, and intervention reports with specified characteristic data, and by querying and matching according to the intervention knowledge database, an intervention plan is obtained; S5.4: Package all intervention plans to obtain a personalized intervention plan report; S5.5: Output the personalized intervention plan report to the adaptive cognitive task scheduling module; The adaptive cognitive task scheduling module is used to execute personalized intervention plan reports and dynamically adjust cognitive tasks based on real-time user feedback during execution, thereby obtaining an intervention process log. Furthermore, the steps involved in implementing personalized intervention plans and dynamically adjusting cognitive tasks based on real-time user feedback during implementation include: S6.1: By analyzing the personalized intervention plan report, obtain the type of task and the difficulty level of the task to be performed in this training; S6.2: Based on the task database, retrieve a set of cognitive tasks that match the task type and difficulty level in step S6.1, and push them to the user interface. S6.3: While the patient is performing the cognitive task in step S6.2, record the number of consecutive correct answers the patient receives; When the patient gets the correct answer 3 times or more consecutively, a difficulty level increase instruction is generated. When a patient fails a task, a command to reduce the difficulty level is generated. S6.4: Dynamically adjust the next cognitive task performed by the patient based on instructions to increase or decrease the difficulty level; S6.5: Record the difficulty level, completion status, and reaction time of all cognitive tasks performed by the patient in step S6.3 to obtain the intervention process log; S6.6: Output the intervention process log to the visualization display closed-loop management module and the personalized digital mapping construction module respectively; The visual display closed-loop management module is used to process current patient-related data; Furthermore, the steps for processing current patient-related data include: S7.1: Based on the database, obtain the current patient's feature vector report, personalized digital report, cognitive risk prediction report, personalized intervention plan report, and intervention process log; S7.2: Generate a cognitive trajectory map based on the cognitive prediction score curve; Generate an intervention content table based on the intervention process log; S7.3: Display the cognitive trajectory map and intervention content table through the visualization panel on the doctor's receiving end; S7.4: Based on the current patient's preset assessment cycle and the difficulty level of the most recent cognitive task in step S6.5, generate a patient assessment report. The patient assessment report includes the next assessment time point for the current patient and the difficulty level of the cognitive task. S7.5: Output the patient assessment report to the raw interactive data acquisition module; Further, S1: Generate the evaluation task and obtain the raw interaction dataset during the user interaction process. The raw interaction dataset includes the clock drawing test dataset, the voice test data, and the selection test dataset. S2: Preprocess the original interactive dataset and extract multimodal features to obtain a feature vector report; S3: Construct a digital mapping content reflecting the patient's cognitive state based on the feature vector report to obtain a personalized digital report; S4: Based on personalized digital reports, predict the patient's future cognitive decline risk and obtain a cognitive risk prediction report; S5: Match intervention plans based on the cognitive risk prediction report to obtain a personalized intervention plan report; S6: Implement the personalized intervention plan report and dynamically adjust the cognitive task based on the user's real-time feedback during the implementation process to obtain the intervention process log; S7: Process the current patient-related data.
[0006] The technical effects and advantages of the cognitive function assessment system for neurology of this invention are as follows: This invention generates assessment tasks and acquires raw interaction datasets from user interactions, including clock-drawing test datasets, voice test datasets, and selection test datasets. Based on these raw datasets, preprocessing and multimodal feature extraction are performed to obtain feature vector reports. Based on these feature vector reports, a digital mapping reflecting the patient's cognitive state is constructed, resulting in a personalized digital report. Based on these personalized digital reports, the patient's future cognitive decline risk is predicted, resulting in a cognitive risk prediction report. Based on this cognitive risk prediction report, an intervention plan is matched, resulting in a personalized intervention plan report. The personalized intervention plan is executed, and the cognitive task is dynamically adjusted based on real-time user feedback during execution, resulting in an intervention process log. This process processes current patient-related data, enabling the system to transform the single-outcome scoring of traditional systems into a task-based assessment system. The multi-dimensional objective interactive data of the task execution process greatly reduces the subjective bias caused by the single "outcome-based" evaluation method in traditional systems. In addition, this invention also links the results of a single evaluation into a time series of "numerical tables" and uses artificial intelligence models to predict the trajectory of patients' cognitive changes. This enables the system to transform traditional post-event judgment into pre-event warning, greatly expanding the valuable time for patients' treatment. Finally, through the prediction report, highly targeted patient intervention plans are automatically generated and executed, and can be dynamically adjusted according to the patient's real-time performance. This ensures that the cognitive task and the patient's ability are kept at the same level to the greatest extent, providing the most appropriate and realistic data support for the patient's subsequent treatment. Overall, this invention has significant advantages such as good objective quantitative effect of evaluation indicators, strong dynamic prediction ability of cognitive risk, and great effect of personalized intervention plans. Attached Figure Description
[0007] Figure 1 This is a schematic diagram of a cognitive function assessment system for neurology according to the present invention; Figure 2 This is a schematic diagram of a cognitive function assessment method for neurology according to the present invention. Detailed Implementation
[0008] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0009] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” used in the embodiments of this invention are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.
[0010] Depending on the context, the words “if” or “suppose” as used here can be interpreted as “when” or “in response to determination” or “in response to detection.” Similarly, depending on the context, the phrases “if determination” or “if detection (of the stated condition or event)” can be interpreted as “when determination” or “in response to determination” or “when detection (of the stated condition or event)” or “in response to detection (of the stated condition or event).”
[0011] Furthermore, the timing of the steps in the following method embodiments is merely an example and not a strict limitation.
[0012] In practice, the server-side equipment deployed for a neurological cognitive function assessment system may consist of one or more devices. This neurological cognitive function assessment system can be implemented as a business instance, a virtual machine, or a hardware device. For example, the neurological cognitive function assessment system can be implemented as a business instance deployed on one or more devices in a cloud node. Simply put, this neurological cognitive function assessment system can be understood as software deployed on a cloud node to provide the neurological cognitive function assessment system to various user terminals. Alternatively, the neurological cognitive function assessment system can also be implemented as a virtual machine deployed on one or more devices in a cloud node. This virtual machine contains application software for managing various user terminals. Alternatively, the neurological cognitive function assessment system can also be implemented as a server composed of numerous identical or different types of hardware devices, with one or more hardware devices configured to provide the neurological cognitive function assessment system to various user terminals.
[0013] In terms of implementation, the cognitive function assessment system for neurology and the user terminal are mutually compatible. That is, if the cognitive function assessment system for neurology is implemented as an application installed on a cloud service platform, then the user terminal is a client that establishes a communication connection with the application; or if the cognitive function assessment system for neurology is implemented as a website, then the user terminal is implemented as a webpage; or if the cognitive function assessment system for neurology is implemented as a cloud service platform, then the user terminal is implemented as a mini-program in an instant messaging application.
[0014] like Figure 1 The figure shown is a system architecture diagram of a cognitive function assessment system for neurology provided in an embodiment of the present invention.
[0015] The cognitive function assessment system for neurology described in this invention can be hosted on a cloud server. In terms of implementation, it can function as one or more service devices, or as an application installed in the cloud (e.g., a mobile service operator's server, server cluster, etc.), or it can be developed as a website. Depending on the functions implemented, the cognitive function assessment system for neurology may include a raw interactive data acquisition module, a feature data integration and calculation module, a personalized digital mapping construction module, a cognitive trajectory prediction module, an intervention plan matching and generation module, an adaptive cognitive task scheduling module, and a visualization display closed-loop management module. The module described in this invention can also be called a unit, referring to a series of computer program segments that can be executed by an electronic device's processor and perform a fixed function, stored in the electronic device's memory.
[0016] In this embodiment of the invention, in the cognitive function assessment system for neurology, each of the above-mentioned modules can be implemented independently and can call other modules. Here, "calling" can be understood as a module connecting to multiple modules of another type and providing corresponding services to those connected modules. For example, the sharing assessment module can call the same information collection module to obtain information collected by that module. Based on the above characteristics, in the cognitive function assessment system for neurology provided in this embodiment of the invention, without modifying the program code, the applicable scope of the cognitive function assessment system architecture can be adjusted by adding modules and directly calling them, achieving cluster-based horizontal expansion to quickly and flexibly expand the cognitive function assessment system for neurology. In practical applications, the above modules can be set in the same device or different devices, or they can be set in a virtual device, such as a service instance in a cloud server.
[0017] Please refer to Example 1 Figure 1 As shown in this embodiment, a cognitive function assessment system for neurology is described, the system comprising: The original interaction data acquisition module is used to generate evaluation tasks and acquire the original interaction dataset during the user interaction process. The original interaction dataset includes clock drawing test dataset, voice test data and selection test dataset. Furthermore, the steps for generating the evaluation task and obtaining the raw interaction dataset during the user interaction process include: S1.1: Based on the task database and the evaluation report, select a set of cognitive tasks and output the specified cognitive tasks to the user interface in step S1.2; It should be explained that the assessment report refers to the report used to select cognitive tasks based on the patient's assessment results. Here, the assessment report is assumed to be the initial assessment. The cognitive task refers to the document used to assess the patient's cognitive function, such as: "Please remember the following five words, which are banana, mobile phone, house, blue, and happy", and has a voice broadcast function. S1.2: The system activates the user interface corresponding to the cognitive task and records the interaction events when the user performs the cognitive task. The user interface includes a touch screen, a voice input tool, and multiple-choice options. When the patient performs the cognitive task of drawing a clock, the clock drawing test dataset is obtained by recording the second-by-second coordinate sequence of the click point, the click event, the start time stamp of the handwriting, and the end time stamp of the handwriting. When the patient performs a cognitive task that is a speech test, the audio recording tool is activated to record the user's audio and obtain speech test data. When a patient performs a cognitive task involving a selection test, the patient's option click events and reaction times are recorded to obtain a selection test dataset. Package the clock drawing test dataset, the voice test data, and the selection test dataset to obtain the original interactive dataset; S1.3: Integrate the behavior timestamps into the original interaction dataset and output them to the feature data integration and calculation module; It should be explained that the behavior timestamp refers to the time point at which data is collected during the execution of the patient's interactive task in step S1.2; The feature data integration and calculation module is used to preprocess the original interactive dataset and extract multimodal features to obtain a feature vector report. Furthermore, the steps of preprocessing the original interactive dataset and extracting multimodal features include: S2.1: Use a low-pass filter to denoise the second-by-second coordinate sequence of clicks in the original interactive dataset, and remove silent segments from the speech test data to obtain a preprocessed dataset; S2.2: Extract multimodal features based on the preprocessed dataset to obtain an initial feature vector report; It should be explained that this step uses the clock drawing test dataset as an example; The total drawing time feature data is obtained by subtracting the handwriting start time from the handwriting end timestamp. By substituting into the calculation formula: , obtained the Instantaneous velocity feature data of each sampling point ,in, For the first The coordinates of each sampling point For the first The timestamps corresponding to each sampling point; Based on instantaneous velocity feature data, average velocity feature data The calculation is performed using the following formula: ; in, This represents the total number of samples in the handwriting trajectory. Speed flow characteristic data is obtained by dividing the standard deviation of instantaneous speed characteristic data by the average speed characteristic data; Based on instantaneous velocity characteristic data, the proportion of pause time characteristic data The calculation is performed using the following formula: ; in, For indicator functions, For speed threshold, Total drawing time; It needs to be explained that the indicator function determines the first... The instantaneous velocity feature data of each sampling point is checked to see if it is less than the velocity threshold. If the result is yes, the function value is 1; otherwise, the function value is 0. The velocity threshold is set manually and input into the system. Pack together the total time feature data, instantaneous velocity feature data, average velocity feature data, velocity smoothness feature data, and pause time ratio feature data to obtain an initial feature vector report; S2.3: Perform Z-socre normalization on the initial feature vector report and all other feature data to obtain the feature vector report; It should be explained that all other features refer to feature data extracted based on the speech test data in the preprocessed dataset and the selected test dataset; S2.4: Output the feature vector report to the personalized digital mapping construction module; The personalized digital mapping construction module is used to construct digital mapping content reflecting the patient's cognitive state based on the feature vector report, and obtain a personalized digital report. Furthermore, the steps for constructing a digital mapping reflecting the patient's cognitive state based on the feature vector report include: S3.1: Obtain the historical personalized digital report of the specified patient based on the database. If the historical personalized digital report exists in the database, proceed to step S3.2. If the historical personalized digital report does not exist in the database, proceed to step S3.3. S3.2: Integrate the feature vector reports into the historical personalized digital reports based on time sorting to obtain a new feature matrix, where each row represents an evaluation and each column represents a feature data; S3.3: Based on the new feature matrix in step S3.2, convert the feature vector report into the format of the new feature matrix; S3.4: The exponentially weighted moving average method is used to assign higher weights to each feature data in the new feature matrix to obtain a new personalized digital report; S3.5: Package the feature vector report and the new personalized digital report from step S3.3 to obtain the personalized digital report; S3.6: Store personalized digital reports in the database. When a personalized digital report is a feature vector report, create a record file for the specified patient. When a personalized digital report is a new personalized digital report, overwrite the previous new personalized digital report with the new personalized digital report. S3.7: Output the personalized digital report to the cognitive trajectory prediction module; The cognitive trajectory prediction module is used to predict the patient's future cognitive decline risk based on personalized digital reports, and to obtain a cognitive risk prediction report. Furthermore, the steps for predicting a patient's future risk of cognitive decline based on personalized digital reports include: S4.1: Load the pre-trained cognitive trajectory prediction model; It should be explained that the cognitive trajectory prediction model is, for example, a time series model based on LSTM or Transformer. This model is trained on historical patient data in a database and can learn the mapping relationship from a sequence of numbers to future cognitive states. S4.2: Input personalized digital reports into the cognitive trajectory prediction model; S4.3: The cognitive trajectory prediction model outputs the cognitive prediction score for the Kth time point in the future; At the same time, the model outputs risk probability data; It should be explained that the risk probability data refers to the probability that a patient will develop dementia within the next M years; S4.4: Package the cognitive prediction score and risk probability data to obtain a cognitive risk prediction report; S4.5: Output the cognitive risk prediction report to the intervention plan matching and generation module; The intervention plan matching and generation module is used to match intervention plans based on the cognitive risk prediction report and obtain a personalized intervention plan report. Furthermore, the steps for matching intervention programs based on cognitive risk prediction reports include: S5.1: Based on the risk threshold, when the risk probability data is greater than or equal to the risk threshold, a high-intensity intervention instruction is generated; when the risk probability data is less than the risk threshold, a low-intensity intervention instruction is generated. S5.2: Identify the feature data that declines the fastest in the cognitive prediction score curve and generate an intervention report for the specified feature data; It should be explained that the cognitive prediction score curve refers to the curve obtained by distributing the cognitive prediction scores based on the prediction time points from near to far. S5.3: Based on high-intensity intervention instructions, low-intensity intervention instructions, and intervention reports with specified characteristic data, and by querying and matching according to the intervention knowledge database, an intervention plan is obtained; S5.4: Package all intervention plans to obtain a personalized intervention plan report; It should be explained that the content of the personalized intervention program report includes, but is not limited to, a list of recommended tasks, weekly training frequency, duration of a single training session, and expected target values. S5.5: Output the personalized intervention plan report to the adaptive cognitive task scheduling module; The adaptive cognitive task scheduling module is used to execute personalized intervention plan reports and dynamically adjust cognitive tasks based on real-time user feedback during execution to obtain an intervention process log. Furthermore, the steps involved in implementing personalized intervention plans and dynamically adjusting cognitive tasks based on real-time user feedback during implementation include: S6.1: By analyzing the personalized intervention plan report, obtain the type of task and the difficulty level of the task to be performed in this training; S6.2: Based on the task database, retrieve a set of cognitive tasks that match the task type and difficulty level in step S6.1, and push them to the user interface. S6.3: While the patient is performing the cognitive task in step S6.2, record the number of consecutive correct answers the patient receives; When the patient gets the correct answer 3 times or more consecutively, a difficulty level increase instruction is generated. When a patient fails a task, a command to reduce the difficulty level is generated. S6.4: Dynamically adjust the next cognitive task performed by the patient based on instructions to increase or decrease the difficulty level; S6.5: Record the difficulty level, completion status, and reaction time of all cognitive tasks performed by the patient in step S6.3 to obtain the intervention process log; S6.6: Output the intervention process log to the visualization display closed-loop management module and the personalized digital mapping construction module respectively; The visualization display closed-loop management module is used to process the current patient-related data; Further steps for processing current patient-related data include: S7.1: Based on the database, obtain the current patient's feature vector report, personalized digital report, cognitive risk prediction report, personalized intervention plan report, and intervention process log; S7.2: Generate a cognitive trajectory map based on the cognitive prediction score curve; Generate an intervention content table based on the intervention process log; S7.3: Display the cognitive trajectory map and intervention content table through the visualization panel on the doctor's receiving end; S7.4: Based on the current patient's preset assessment cycle and the difficulty level of the most recent cognitive task in step S6.5, generate a patient assessment report. The patient assessment report includes the next assessment time point for the current patient and the difficulty level of the cognitive task. S7.5: Output the patient assessment report to the raw interactive data acquisition module; In this embodiment, the beneficial effects are achieved by generating an assessment task and acquiring the original interaction dataset during the user interaction process. This original interaction dataset includes clock-drawing test datasets, voice test data, and selection test datasets. Preprocessing is performed on the original interaction dataset, and multimodal feature extraction is conducted to obtain a feature vector report. Based on the feature vector report, a digital mapping content reflecting the patient's cognitive state is constructed, resulting in a personalized digital report. Based on the personalized digital report, the patient's future cognitive decline risk is predicted, resulting in a cognitive risk prediction report. Based on the cognitive risk prediction report, an intervention plan is matched to obtain a personalized intervention plan report. The personalized intervention plan report is executed, and the cognitive task is dynamically adjusted based on the user's real-time feedback during execution, resulting in an intervention process log. This process processes the current patient-related data, enabling the system to transform the single-outcome scoring of traditional systems into a more dynamic and comprehensive system. Based on multi-dimensional objective interactive data of the task execution process, this invention significantly reduces the subjective bias caused by the single "outcome-based" evaluation method in traditional systems. Furthermore, by stringing together single evaluation results into a time series of "numerical tables" and using artificial intelligence models to predict the trajectory of patients' cognitive changes, the system can transform traditional post-event judgment into pre-event warning, greatly extending the valuable time available for patient treatment. Finally, through predictive reports, highly targeted patient intervention plans are automatically generated and executed, and can be dynamically adjusted based on the patient's real-time performance, ensuring that the cognitive task and the patient's abilities remain at the same level to the greatest extent possible. This provides the most relevant and realistic data support for the patient's subsequent treatment. Overall, this invention has significant advantages such as good objective quantification of evaluation indicators, strong dynamic prediction ability of cognitive risk, and significant effect of personalized intervention plans.
[0018] Please refer to Example 2 Figure 2 As shown, the parts not described in detail in this embodiment are described in Embodiment 1. A cognitive function assessment method for neurology is provided. The method includes: S1: generating an assessment task and obtaining the original interaction dataset during the user interaction process. The original interaction dataset includes a clock drawing test dataset, a voice test dataset, and a selection test dataset. S2: Preprocess the original interactive dataset and extract multimodal features to obtain a feature vector report; S3: Construct a digital mapping content reflecting the patient's cognitive state based on the feature vector report to obtain a personalized digital report; S4: Based on personalized digital reports, predict the patient's future cognitive decline risk and obtain a cognitive risk prediction report; S5: Match intervention plans based on the cognitive risk prediction report to obtain a personalized intervention plan report; S6: Implement the personalized intervention plan report and dynamically adjust the cognitive task based on the user's real-time feedback during the implementation process to obtain the intervention process log; S7: Process the current patient-related data.
[0019] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the present invention.
Claims
1. A cognitive function assessment system for neurology, characterized in that, The system includes: a personalized digital mapping construction module, a cognitive trajectory prediction module, an intervention plan matching and generation module, and an adaptive cognitive task scheduling module, wherein: The personalized digital mapping construction module is used to construct digital mapping content reflecting the patient's cognitive state based on the feature vector report, and obtain a personalized digital report. The cognitive trajectory prediction module is used to predict the patient's future cognitive decline risk based on personalized digital reports, and to obtain a cognitive risk prediction report. The intervention plan matching and generation module is used to match intervention plans based on the cognitive risk prediction report and obtain a personalized intervention plan report. The adaptive cognitive task scheduling module is used to execute personalized intervention plan reports and dynamically adjust cognitive tasks based on real-time user feedback during execution to obtain intervention process logs.
2. The cognitive function assessment system for neurology according to claim 1, characterized in that, The system also includes: a raw interactive data acquisition module, a feature data integration and calculation module, and a visualization display closed-loop management module, wherein: The original interaction data acquisition module is used to generate evaluation tasks and acquire the original interaction dataset during the user interaction process. The original interaction dataset includes clock drawing test dataset, voice test data and selection test dataset. The feature data integration and calculation module is used to preprocess the original interactive dataset and extract multimodal features to obtain a feature vector report. The visualization display closed-loop management module is used to process the current patient-related data.
3. The cognitive function assessment system for neurology according to claim 2, characterized in that, The steps to generate the evaluation task and obtain the raw interaction dataset during the user interaction process include: S1.1: Based on the task database and the evaluation report, select a set of cognitive tasks and output the specified cognitive tasks to the user interface in step S1.2; S1.2: The system activates the user interface corresponding to the cognitive task and records the interaction events when the user performs the cognitive task. The user interface includes a touch screen, a voice input tool, and multiple-choice options. When the patient performs the cognitive task of drawing a clock, the clock drawing test dataset is obtained by recording the second-by-second coordinate sequence of the click point, the click event, the start time stamp of the handwriting, and the end time stamp of the handwriting. When the patient performs a cognitive task that is a speech test, the audio recording tool is activated to record the user's audio and obtain speech test data. When a patient performs a cognitive task involving a selection test, the patient's option click events and reaction times are recorded to obtain a selection test dataset. Package the clock drawing test dataset, the voice test data, and the selection test dataset to obtain the original interactive dataset; S1.3: Integrate the behavior timestamps into the original interaction dataset and output them to the feature data integration and calculation module.
4. A cognitive function assessment system for neurology according to claim 2, characterized in that, The steps for preprocessing the original interactive dataset and extracting multimodal features include: S2.1: Use a low-pass filter to denoise the second-by-second coordinate sequence of clicks in the original interactive dataset, and remove silent segments from the speech test data to obtain a preprocessed dataset; S2.2: Extract multimodal features based on the preprocessed dataset to obtain an initial feature vector report; S2.3: Perform Z-socre normalization on the initial feature vector report and all other feature data to obtain the feature vector report; S2.4: Output the feature vector report to the personalized digital mapping building module.
5. A cognitive function assessment system for neurology according to claim 1, characterized in that, The steps for constructing a digital mapping reflecting a patient's cognitive state based on the feature vector report include: S3.1: Obtain the historical personalized digital report of the specified patient based on the database. If the historical personalized digital report exists in the database, proceed to step S3.
2. If the historical personalized digital report does not exist in the database, proceed to step S3.
3. S3.2: Integrate the feature vector reports into the historical personalized digital reports based on time sorting to obtain a new feature matrix, where each row represents an evaluation and each column represents a feature data; S3.3: Based on the new feature matrix in step S3.2, convert the feature vector report into the format of the new feature matrix; S3.4: The exponentially weighted moving average method is used to assign higher weights to each feature data in the new feature matrix to obtain a new personalized digital report; S3.5: Package the feature vector report and the new personalized digital report from step S3.3 to obtain the personalized digital report; S3.6: Store personalized digital reports in the database. When a personalized digital report is a feature vector report, create a record file for the specified patient. When a personalized digital report is a new personalized digital report, overwrite the previous new personalized digital report with the new personalized digital report. S3.7: Output personalized digital reports to the cognitive trajectory prediction module.
6. A cognitive function assessment system for neurology according to claim 1, characterized in that, The steps involved in predicting a patient's future risk of cognitive decline based on personalized digital reports include: S4.1: Load the pre-trained cognitive trajectory prediction model; S4.2: Input personalized digital reports into the cognitive trajectory prediction model; S4.3: The cognitive trajectory prediction model outputs the cognitive prediction score for the Kth time point in the future; At the same time, the model outputs risk probability data; S4.4: Package the cognitive prediction score and risk probability data to obtain a cognitive risk prediction report; S4.5: Output the cognitive risk prediction report to the intervention plan matching and generation module.
7. A cognitive function assessment system for neurology according to claim 1, characterized in that, The steps for matching intervention programs based on cognitive risk prediction reports include: S5.1: Based on the risk threshold, when the risk probability data is greater than or equal to the risk threshold, a high-intensity intervention instruction is generated; when the risk probability data is less than the risk threshold, a low-intensity intervention instruction is generated. S5.2: Identify the feature data that declines the fastest in the cognitive prediction score curve and generate an intervention report for the specified feature data; S5.3: Based on high-intensity intervention instructions, low-intensity intervention instructions, and intervention reports with specified characteristic data, and by querying and matching according to the intervention knowledge database, an intervention plan is obtained; S5.4: Package all intervention plans to obtain a personalized intervention plan report; S5.5: Output the personalized intervention plan report to the adaptive cognitive task scheduling module.
8. A cognitive function assessment system for neurology according to claim 1, characterized in that, The steps for implementing personalized intervention plans and dynamically adjusting cognitive tasks based on real-time user feedback during implementation include: S6.1: By analyzing the personalized intervention plan report, obtain the type of task and the difficulty level of the task to be performed in this training; S6.2: Based on the task database, retrieve a set of cognitive tasks that match the task type and difficulty level in step S6.1, and push them to the user interface. S6.3: While the patient is performing the cognitive task in step S6.2, record the number of consecutive correct answers the patient receives; When the patient gets the correct answer 3 times or more consecutively, a difficulty level increase instruction is generated. When a patient fails a task, a command to reduce the difficulty level is generated. S6.4: Dynamically adjust the next cognitive task performed by the patient based on instructions to increase or decrease the difficulty level; S6.5: Record the difficulty level, completion status, and reaction time of all cognitive tasks performed by the patient in step S6.3 to obtain the intervention process log; S6.6: Output the intervention process log to the visualization display closed-loop management module and the personalized digital mapping construction module respectively.
9. A cognitive function assessment system for neurology according to claim 2, characterized in that, The steps for processing current patient-related data include: S7.1: Based on the database, obtain the current patient's feature vector report, personalized digital report, cognitive risk prediction report, personalized intervention plan report, and intervention process log; S7.2: Generate a cognitive trajectory map based on the cognitive prediction score curve; Generate an intervention content table based on the intervention process log; S7.3: Display the cognitive trajectory map and intervention content table through the visualization panel on the doctor's receiving end; S7.4: Based on the current patient's preset assessment cycle and the difficulty level of the most recent cognitive task in step S6.5, generate a patient assessment report. The patient assessment report includes the next assessment time point for the current patient and the difficulty level of the cognitive task. S7.5: Output the patient assessment report to the raw interactive data acquisition module.
10. A method for assessing cognitive function in neurology, implemented according to any one of claims 1-9, characterized in that, The work includes the following steps: S1: Generate the evaluation task and obtain the raw interaction dataset during the user interaction process. The raw interaction dataset includes the clock drawing test dataset, the voice test data, and the selection test dataset. S2: Preprocess the original interactive dataset and extract multimodal features to obtain a feature vector report; S3: Construct a digital mapping content reflecting the patient's cognitive state based on the feature vector report to obtain a personalized digital report; S4: Based on personalized digital reports, predict the patient's future cognitive decline risk and obtain a cognitive risk prediction report; S5: Match intervention plans based on the cognitive risk prediction report to obtain a personalized intervention plan report; S6: Implement the personalized intervention plan report and dynamically adjust the cognitive task based on the user's real-time feedback during the implementation process to obtain the intervention process log; S7: Process the current patient-related data.