An Alzheimer's disease incidence risk prediction system

Abstract

This invention belongs to the field of biomedical technology, specifically an Alzheimer's disease risk prediction system. It includes a cognitive parameter generation module, which determines the period of stable cognitive function based on historical follow-up records and generates a cognitive parameter set; a challenge task execution module, which obtains conversion reaction time, conversion accuracy, and memory retention accuracy based on the cognitive parameter set; a practice correction labeling module, which obtains a practice correction index and generates a cognitive state label; a decline cross-validation module, which generates a cross-validation index based on the matching degree between brain metabolic data and language function decline data; a difference score calculation module, which obtains the difference score between self-assessment and other-assessment answers; a subjective verification and standardization module, which verifies the difference score based on the cross-validation index and generates standardized subjective data; and a risk parameter tuning and prediction module, which inputs the risk prediction model to obtain the risk level and adjusts the cognitive parameter set for the next cycle accordingly, thereby improving the accuracy of early prediction of brain diseases.

Description

Technical Field

[0001] This invention relates to the field of biomedical technology, specifically to an Alzheimer's disease risk prediction system. Background Technology

[0002] An Alzheimer's disease risk prediction system is an intelligent tool that uses machine learning or deep learning models to calculate an individual's future probability of developing the disease before symptoms appear, based on clinical data, biomarkers, genetic information, and cognitive assessment. It can integrate multi-dimensional data such as blood indicators, brain imaging, and APOE genes to output risk stratification reports for early warning and intervention decisions, helping to advance the diagnostic window by several years.

[0003] In existing technologies, a data input module first integrates multimodal data, including clinical data, neuropsychological assessments, blood biomarkers, and genetic information. The system then uses machine learning or deep learning models, such as logistic regression, decision trees, generative artificial intelligence, or graph neural networks, to process and analyze this data to build predictive models. Ultimately, the system outputs risk assessment results, disease progression trajectories, or personalized intervention recommendations, thereby supporting early screening and clinical decision-making for Alzheimer's disease.

[0004] The above-mentioned approach still has some problems in practical application. In the long-term risk assessment of Alzheimer's disease, it is often necessary to conduct multiple repeated cognitive tests on the same subject. However, the existing assessment methods are difficult to distinguish between the performance improvement brought about by familiarity with the test content or mastery of answering skills and the subject's real changes in cognitive function. In other words, if a person's score increases after multiple tests, it is only because he "learned how to do the test" rather than because his brain function has truly improved or remained stable. Underlying signs of decline will be masked by this false gain. This can lead to systematic bias in the risk assessment results under long-term dynamic monitoring, making it difficult to accurately reflect the subject's true risk status of developing the disease.

[0005] Therefore, the present invention provides an Alzheimer's disease incidence risk prediction system. Summary of the Invention

[0006] This application provides an Alzheimer's disease incidence risk prediction system that can achieve interlocked verification in both subjective and objective dimensions, reduce the risk of misjudgment from a single data source, improve the authenticity and accuracy of the data input into the risk prediction model, and enhance the system's ability to stably output the risk level of brain diseases in complex clinical situations.

[0007] To achieve the above objectives, this application adopts the following technical solution: This application provides an Alzheimer's disease incidence risk prediction system, the module of which includes: The cognitive parameter generation module is used to determine the corresponding period of cognitive function stability of the subject based on the subject's historical follow-up records and generate a set of cognitive parameters. The challenge task execution module is used to implement a cognitive challenge task on the subject based on the cognitive parameter set, and to obtain the subject's conversion reaction time, conversion accuracy and memory retention accuracy in the cognitive challenge task; The practice correction labeling module is used to input the conversion reaction time, conversion accuracy and memory retention accuracy into a preset comprehensive evaluation model to obtain a practice correction index, and generate a cognitive state label based on the comparison result of the practice correction index and a first preset threshold. The cognitive decline cross-validation module is used to collect cognitive decline data of subjects, including brain metabolism data and language function decline data. Based on the matching degree between the brain metabolism data and language function decline data, a cognitive decline cross-validation index is generated. The difference score calculation module is used to obtain the self-rated answers of the subjects and the other-rated answers of informed individuals, and calculates the difference score based on the degree of deviation between the self-rated answers and the other-rated answers; The subjective verification and standardization module is used to verify the credibility of the difference scores based on the cognitive decline cross-validation index, and generate standardized subjective data based on the verification results. The risk parameter tuning and prediction module inputs the practice correction index, the cognitive state label, the cognitive decline cross-validation index, and the standardized subjective data into a preset Alzheimer's disease risk prediction model to obtain the subject's brain disease risk level, i.e., the subject's Alzheimer's disease risk level, and adjusts the cognitive parameter set in the next follow-up period based on the Alzheimer's disease risk level.

[0008] In some possible implementations, the challenge task execution module is specifically used for: The duration range of the task is determined based on the information on the stability period of cognitive function in the cognitive parameter set. Based on the historical task completion data in the cognitive parameter set, the subject's task completion fluctuations in different cognitive tasks are statistically analyzed, and the task switching frequency is determined based on the task completion fluctuations. Based on the historical memory retention data in the cognitive parameter set, the changes in memory retention of the subjects in different memory stages are statistically analyzed, and the memory retention detection interval is determined based on the changes in memory retention. Based on the duration range of the task, the frequency of task switching, and the memory retention detection interval, different cognitive tasks are arranged and combined to form a cognitive challenge task sequence; For each cognitive task in the cognitive challenge task sequence, the task switching relationship between the current cognitive task and the next cognitive task is determined, and the task response behavior data of the subject during the task switching process is extracted based on the task switching relationship to obtain the corresponding task response time data and task answer result data. The task response time data is used to statistically analyze the changes in the subjects' response time during different task switching processes, and the switching reaction time is obtained based on the results of the response time changes during each task switching process. The task response data is used to statistically analyze the subjects' correctness of task responses after task switching, and the conversion accuracy rate is obtained based on the statistical results. After the cognitive challenge task sequence is completed, the subject is tested for memory retention and the corresponding memory response results are obtained. The accuracy of the subjects' memory responses after a preset memory retention time is statistically analyzed based on the memory response results, and the memory retention accuracy rate is obtained based on the statistical results.

[0009] In some possible implementations, the practice correction marker module is specifically used for: The conversion reaction time, conversion accuracy, and memory retention accuracy are obtained, and correlation matching processing is performed on the conversion reaction time, conversion accuracy, and memory retention accuracy according to the preset data correspondence to form cognitive challenge assessment data; Based on the cognitive challenge assessment data, the changes in the subjects' task performance at different cognitive challenge stages were statistically analyzed, and task adaptation change characteristics were extracted based on the changes in task performance. Based on the task adaptation and change characteristics, the study analyzes the changes in the impact of practice on the subjects during continuous cognitive challenges, and generates a practice impact assessment result based on the changes in the impact of practice. The conversion reaction time, conversion accuracy, and memory retention accuracy are input into a preset comprehensive evaluation model, and the evaluation process of the comprehensive evaluation model is corrected in combination with the evaluation results of the practice impact, so as to generate a corresponding practice correction index. Obtain the first preset threshold, and determine the change in the subject's cognitive state based on the comparison result between the practice correction index and the first preset threshold; Generate cognitive state tags based on the changes in cognitive state.

[0010] In some possible implementations, the practice correction marker module is specifically used for: Calculate the difference between the practice correction index and the first preset threshold; When the difference is within the first preset difference range, the change in the subject's cognitive state is determined to be a stable cognitive state. When the difference is greater than the upper limit of the first preset difference range, the change in the subject's cognitive state is determined to be an improvement in cognitive state; When the difference is less than the lower limit of the first preset difference interval, the change in the subject's cognitive state is determined to be a decline in cognitive state.

[0011] In some possible implementations, the fading cross-validation module is specifically used for: Brain metabolism data of the subjects were collected during the follow-up period, and brain metabolism changes in different brain regions of the subjects were statistically analyzed based on the brain metabolism data to form brain metabolism change data. Data on the decline of language function in the subjects during the follow-up period were collected, and the changes in language function in different language tasks were statistically analyzed based on the data on the decline of language function to form data on changes in language function. Based on the brain metabolism change data, the direction of brain metabolism change in different brain regions of the subjects was extracted, and based on the language function change data, the direction of language function change in different language tasks of the subjects was extracted. The direction of brain metabolic changes is correlated and matched with the direction of language function changes according to the corresponding follow-up period to form a corresponding relationship of decline changes; Based on the aforementioned correlation between the decline changes, the consistency between the direction of brain metabolism changes and the direction of language function changes is statistically analyzed, and brain-language matching results are generated based on the statistical results. The matching degree between the brain metabolism data and the language function decline data is calculated based on the brain language matching results, and a cognitive decline cross-validation index is generated based on the matching degree.

[0012] In some possible implementations, the difference score calculation module is specifically used for: Obtain the self-assessment answers of the subjects during the pre-set questionnaire assessment process, as well as the other-assessment answers of the informed persons corresponding to the subjects during the same pre-set questionnaire assessment process; Based on the correspondence between the self-assessment answers and the peer-assessment answers under the same assessment dimension, the self-assessment answers and the peer-assessment answers are matched item by item to form an assessment answer matching sequence; For each matching item in the evaluation answer matching sequence, calculate the degree of deviation between the self-evaluation answer and the other-evaluation answer, and generate a single-item deviation result; The overall deviation level is calculated based on the aggregated statistical results of each individual deviation, and the difference score is obtained.

[0013] In some possible implementations, the subjective verification standardization module is specifically used for: Obtain the difference score and the cognitive decline cross-validation index, and determine the validation weight coefficient based on the confidence distribution characteristics corresponding to the cognitive decline cross-validation index. The difference scores are weighted and verified based on the verification weight coefficients to generate corresponding credibility verification results. The credibility status of subjective data is determined based on the comparison between the credibility verification result and the preset credibility threshold. Based on the credibility status of the subjective data, the difference scores are subjected to credibility standardization processing to generate standardized subjective data.

[0014] In some possible implementations, the risk parameter tuning and prediction module is specifically used for: The practice correction index, the cognitive state label, the cognitive decline cross-validation index, and the standardized subjective data are obtained, and the input data are uniformly time-series aligned to form a risk prediction input data set. Based on the risk prediction input data set, the practice correction index, the cognitive decline cross-validation index, and the standardized subjective data are subjected to feature fusion processing to generate a comprehensive risk feature vector; The practice correction index, the cognitive state marker, the cognitive decline cross-validation index, and the standardized subjective data are input into a preset Alzheimer's disease risk prediction model, and the model outputs the corresponding risk score result based on the comprehensive risk feature vector. The risk level of the subject for Alzheimer's disease is determined based on the risk score results, and the set of cognitive parameters in the next follow-up period is adjusted based on the risk level of Alzheimer's disease.

[0015] In some possible implementations, the cognitive parameter generation module is specifically used for: To obtain the historical cognitive assessment results and historical memory retention results of the subjects in their historical follow-up records; Based on the historical cognitive assessment results, the changes in the subjects' cognitive assessment during the continuous follow-up period were statistically analyzed, and the stable range of cognitive assessment was determined based on the changes in cognitive assessment. Based on the historical memory retention results, the changes in memory retention of the subjects during the continuous follow-up period were statistically analyzed, and the memory retention stability interval was determined based on the changes in memory retention. By matching the cognitive assessment stability interval with the memory retention stability interval, the corresponding cognitive function stability period of the subject is determined; Based on the stable period of cognitive function, the historical task completion status and historical memory retention status in the corresponding follow-up period are extracted, and a set of cognitive parameters is generated.

[0016] The beneficial effects of this invention are as follows: 1. The Alzheimer's disease risk prediction system of the present invention inputs the subject's conversion reaction time, conversion accuracy, and memory retention accuracy in a cognitive challenge task into a preset comprehensive evaluation model through a practice correction labeling module. The system calculates a practice correction index and generates a cognitive state label based on a comparison with a first preset threshold. This effectively removes the practice effect caused by familiarity with content or mastery of skills during repeated testing. Furthermore, by incorporating the practice correction index and the cognitive state label into the subsequent risk prediction model, the system can distinguish between real cognitive function changes and spurious test gains. This effectively suppresses systematic bias caused by multiple tests during long-term longitudinal follow-up, prevents potential decline signals from being masked by learning effects, improves the dynamic accuracy and individualized tracking ability of Alzheimer's disease risk assessment, and enhances the credibility of long-term monitoring results in clinical decision-making. 2. The Alzheimer's disease incidence risk prediction system described in this invention collects brain metabolic data and language function decline data through a cognitive decline cross-validation module and generates a cognitive decline cross-validation index. This quantifies the matching degree between objective physiological metabolism and language function decline. Furthermore, a subjective verification and standardization module verifies the credibility of the difference scores between self-report and other-report answers based on the cross-validation index and outputs standardized subjective data. This effectively identifies and corrects memory bias, denial tendency, or oversensitivity reports in subjective questionnaire information, thereby achieving interlocked verification in both subjective and objective dimensions. This reduces the risk of misjudgment from a single data source, improves the authenticity and accuracy of the data input into the risk prediction model, and enhances the system's ability to stably output the risk level of brain diseases in complex clinical situations. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 This is a flowchart of an Alzheimer's disease incidence risk prediction system according to the present invention; Detailed Implementation

[0019] The terms "first," "second," and "third," etc., used in this application specification, claims, and drawings are used to distinguish different objects, not to limit a specific order.

[0020] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0021] Research has found that long-term risk assessment for Alzheimer's disease often requires repeated cognitive tests on the same subject. However, existing assessment methods struggle to distinguish between performance improvements resulting from familiarity with test content or mastery of test-taking skills and genuine changes in cognitive function. In other words, a person's score may increase after multiple tests simply because they "learned how to do the test," rather than because their brain function has truly improved or stabilized. Potential signs of decline can be masked by this false gain, leading to systematic biases in risk assessments under long-term dynamic monitoring, making it difficult to accurately reflect the subject's true risk of developing the disease. Example

[0022] To solve the above problems, such as Figure 1 As shown, this application provides an Alzheimer's disease incidence risk prediction system, the method comprising: Step 1: Cognitive parameter generation module, used to determine the corresponding period of stable cognitive function of the subject based on the subject's historical follow-up records, and generate a set of cognitive parameters; Further: Obtain the historical cognitive assessment results and historical memory retention results of the subjects in their historical follow-up records; Based on the historical cognitive assessment results, the changes in the subjects' cognitive assessment during the continuous follow-up period were statistically analyzed, and the stable range of cognitive assessment was determined based on the changes in cognitive assessment. Based on the historical memory retention results, the changes in memory retention of the subjects during the continuous follow-up period were statistically analyzed, and the memory retention stability interval was determined based on the changes in memory retention. By matching the cognitive assessment stability interval with the memory retention stability interval, the corresponding cognitive function stability period of the subject is determined; Based on the stable period of cognitive function, the historical task completion status and historical memory retention status in the corresponding follow-up period are extracted, and a set of cognitive parameters is generated.

[0023] In step one: Historical follow-up records: refer to the set of historical cognitive test information formed by the subject over multiple consecutive follow-up periods; Historical cognitive assessment results: refers to the cognitive ability assessment results obtained by the subjects after completing cognitive assessment tasks in each follow-up period; Historical memory retention results: refers to the memory retention ability results formed after the subject completes the memory retention test in each follow-up period; Continuous follow-up period: refers to multiple time periods formed continuously according to a preset follow-up time interval; Changes in cognitive assessment: refers to the trend of changes in the cognitive assessment results of subjects over multiple consecutive follow-up periods; Cognitive assessment stability interval: refers to the time interval during which cognitive assessment results remain stable within a continuous follow-up period; Changes in memory retention: refers to the trend of changes in the memory retention outcomes of subjects over multiple consecutive follow-up periods; Memory retention stability interval: refers to the time interval during which memory retention results remain in a stable state of change within a continuous follow-up period; Cognitive function stabilization period: refers to the continuous time period in which both the cognitive assessment stabilization interval and the memory retention stabilization interval meet the requirements of a stable state. Historical task completion status: refers to the subject's task completion efficiency, task response, and task answering in historical cognitive tasks; Cognitive parameter set: refers to the set of parameter data used for configuring subsequent cognitive challenge tasks.

[0024] For example: The cognitive parameter generation module first establishes a historical follow-up record reading sequence for the subject, and reads and organizes the historical cognitive assessment results and historical memory retention results in each follow-up period according to the chronological order of the continuous follow-up periods. After the data reading is completed, the cognitive parameter generation module will further perform continuous change analysis on the historical cognitive assessment results according to the time sequence corresponding to the continuous follow-up period. It will statistically analyze the changes in the cognitive assessment of the subjects in multiple continuous follow-up periods. During the statistical process, it will extract the magnitude of the changes in the cognitive assessment results, the direction of the changes in the cognitive assessment results, and the continuity of the changes in the cognitive assessment results in each continuous follow-up period. Based on the consistency of the changes in the cognitive assessment results in different continuous follow-up periods, it will identify continuous time segments with small fluctuations and stable changes in direction, thereby forming the corresponding stable cognitive assessment intervals. Subsequently, the cognitive parameter generation module will continue to perform continuous change analysis on the historical memory retention results according to the time sequence corresponding to the continuous follow-up period, and statistically analyze the changes in memory retention of the subjects in multiple continuous follow-up periods. During the statistical process, the cognitive parameter generation module will extract the magnitude of changes in memory retention results, the direction of changes in memory retention results, and the continuity of changes in memory retention results in each continuous follow-up period. Based on the consistency of changes in memory retention results in different continuous follow-up periods, the module will identify continuous time segments with small fluctuations and stable directions of change, thereby forming the corresponding stable memory retention intervals. After the stable intervals of cognitive assessment and memory retention are determined, the cognitive parameter generation module will further match the stable intervals of cognitive assessment and memory retention according to the corresponding continuous follow-up periods, and count the time overlap between the stable intervals of cognitive assessment and memory retention. When the cognitive assessment stability interval and the memory retention stability interval simultaneously meet the stable state within the same continuous follow-up period, the cognitive parameter generation module determines the corresponding continuous time period as the cognitive function stability period. After the cognitive function stability period is determined, the cognitive parameter generation module will further extract the historical task completion status and historical memory retention status in the continuous follow-up period corresponding to the cognitive function stability period. Based on the historical task completion status, the module will statistically analyze the changes in task completion, task response, and task answering status of the subjects in different historical cognitive tasks. At the same time, based on the historical memory retention status, the module will statistically analyze the changes in memory retention status of the subjects in different memory retention tests. Subsequently, the cognitive parameter generation module correlates and organizes the patterns of change in task completion, task response, task answering, and memory retention to form a corresponding cognitive parameter set. After the cognitive parameter set is generated, the cognitive parameter generation module outputs the cognitive parameter set to the challenge task execution module.

[0025] Among them: Historical follow-up record reading sequence: refers to the data reading sequence established according to the time order of continuous follow-up period; Continuous variation analysis: refers to the process of continuously statistically analyzing the changes in data over multiple consecutive follow-up periods; Magnitude of change in cognitive assessment results: refers to the magnitude of change in cognitive assessment results over adjacent consecutive follow-up periods; Direction of change in cognitive assessment results: refers to whether the cognitive assessment results show an upward, downward, or stable trend. Continuity of changes in cognitive assessment results: refers to the degree of consistency in the direction of changes in cognitive assessment results across multiple consecutive follow-up periods; Time overlap: refers to the corresponding overlap between the stable interval of cognitive assessment and the stable interval of memory retention in the same continuous follow-up period; Task completion variation pattern: refers to the changes in the stability of task completion by subjects in different historical cognitive tasks; Task response variation pattern: refers to the changes in task response time of subjects in different historical cognitive tasks; Task response variation pattern: refers to the variation pattern of the correctness of subjects' task responses in different historical cognitive tasks; Patterns of memory retention variation: This refers to the patterns of changes in the stability of memory retention in different memory retention tests.

[0026] It should be noted that by performing continuous change analysis on the historical cognitive assessment results and historical memory retention results respectively, and further jointly matching the stable intervals of cognitive assessment and memory retention, the misjudgment problem caused by fluctuations in a single cognitive assessment result can be avoided, thereby improving the stability and reliability of the determination of the stable period of cognitive function. At the same time, by generating a cognitive parameter set based on the historical task completion status and historical memory retention status extracted from the stable period of cognitive function, the task duration, task switching frequency, and memory retention detection interval of subsequent cognitive challenge tasks can be made more consistent with the subject's current cognitive ability change status, thereby improving the individual adaptability and early risk identification accuracy in the Alzheimer's disease risk prediction process.

[0027] Step 2: Challenge Task Execution Module, used to implement cognitive challenge tasks on the subject based on the cognitive parameter set, and to obtain the subject's conversion reaction time, conversion accuracy, and memory retention accuracy in the cognitive challenge tasks; Furthermore: Based on the information on the duration of stable cognitive function in the cognitive parameter set, the range of task duration is determined; Based on the historical task completion data in the cognitive parameter set, the subject's task completion fluctuations in different cognitive tasks are statistically analyzed, and the task switching frequency is determined based on the task completion fluctuations. Based on the historical memory retention data in the cognitive parameter set, the changes in memory retention of the subjects in different memory stages are statistically analyzed, and the memory retention detection interval is determined based on the changes in memory retention. Based on the duration range of the task, the frequency of task switching, and the memory retention detection interval, different cognitive tasks are arranged and combined to form a cognitive challenge task sequence; For each cognitive task in the cognitive challenge task sequence, the task switching relationship between the current cognitive task and the next cognitive task is determined, and the task response behavior data of the subject during the task switching process is extracted based on the task switching relationship to obtain the corresponding task response time data and task answer result data. The task response time data is used to statistically analyze the changes in the subjects' response time during different task switching processes, and the switching reaction time is obtained based on the results of the response time changes during each task switching process. The task response data is used to statistically analyze the subjects' correctness of task responses after task switching, and the conversion accuracy rate is obtained based on the statistical results. After the cognitive challenge task sequence is completed, the subject is tested for memory retention and the corresponding memory response results are obtained. The accuracy of the subjects' memory responses after a preset memory retention time is statistically analyzed based on the memory response results, and the memory retention accuracy rate is obtained based on the statistical results.

[0028] In step two: information on the duration of the stable period of cognitive function: refers to the duration of the stable period of cognitive function in the continuous follow-up period; Task duration range: refers to the allowed duration range for executing a cognitive challenge task; Task completion fluctuation: refers to the changes in the stability of task completion by subjects in different cognitive tasks; Task switching frequency: refers to the frequency of switching between different cognitive tasks; Historical memory retention status: refers to the changes in the subject's memory retention ability during historical memory retention tests; Changes in memory retention: refers to the trend of changes in memory retention outcomes in different memory stages; Memory retention test interval: refers to the time interval between adjacent memory retention tests; Cognitive challenge task sequence: refers to a sequence of multiple cognitive tasks executed in a pre-arranged order; Task switching relationship: refers to the switching correspondence between the current cognitive task and the next cognitive task; Task response behavior: refers to the task response actions of subjects during cognitive task switching; Task response time data: refers to the task response time records generated by the subject during task switching; Task response results data: refers to the recorded data of subjects' task response results in various cognitive tasks; Response time variation: refers to the trend of task response time changes during the switching process between different tasks; Switching reaction time: refers to the average response time of subjects during the switching process between different cognitive tasks; Conversion accuracy rate: refers to the percentage of subjects who answer the task correctly after switching cognitive tasks; Memory response results: refers to the data on the subjects' memory response results in memory retention tests; Preset memory retention time: refers to the preset memory retention waiting time before performing the memory retention test; Memory retention accuracy rate: refers to the percentage of subjects who correctly answer questions about memory after a preset memory retention time.

[0029] For example: The challenge task execution module first obtains the cognitive parameter set and extracts the duration information of the stable period of cognitive function from the cognitive parameter set. Then, the challenge task execution module counts the changes in the duration of the subject in the continuous stable cognitive state based on the duration information of the stable period of cognitive function, and determines the range of task duration corresponding to the subject's current level of cognitive stability based on the duration change results. After determining the duration of the task, the challenge task execution module continues to extract historical task completion data from the cognitive parameter set, and statistically analyzes the fluctuation of the subject's task completion in different cognitive tasks based on the historical task completion data. During the statistical process, the challenge task execution module analyzes the changes in the continuity of task completion, the stability of task completion, and the changes in task response in different cognitive tasks, and determines the corresponding task switching frequency based on the degree of fluctuation difference in different cognitive tasks. Subsequently, the challenge task execution module further extracts historical memory retention from the cognitive parameter set and statistically analyzes the changes in memory retention of the subjects in different memory stages based on the historical memory retention. During the statistical process, the challenge task execution module analyzes the decline and stability of memory retention in different memory stages and determines the corresponding memory retention detection interval based on the memory retention change results in different memory stages. After the duration of the task, the frequency of task switching, and the memory retention detection interval are determined, the challenge task execution module further performs permutation and combination processing on different cognitive tasks according to the duration of the task, the frequency of task switching, and the memory retention detection interval to form a corresponding cognitive challenge task sequence. Subsequently, the challenge task execution module executes the corresponding cognitive tasks on the subject in sequence according to the cognitive challenge task sequence, and determines the task switching relationship between the current cognitive task and the next cognitive task during the execution of the cognitive challenge task sequence. During the task switching process, the challenge task execution module extracts the task response behavior execution data of the subjects during the task switching process according to the task switching relationship, and obtains the corresponding task response time data and task answer result data. Subsequently, the challenge task execution module statistically analyzes the changes in the subjects' response time during different task switching processes based on the task response time data, and calculates the average task response time of the subjects during the cognitive task switching process based on the changes in response time during each task switching process, thereby obtaining the switching reaction time. After the reaction time is obtained, the challenge task execution module will statistically analyze the correctness of the subject's task answers after the task switch based on the task answer results data, and obtain the conversion accuracy rate based on the correct answer ratio in different cognitive tasks. After the entire sequence of cognitive challenge tasks is completed, the challenge task execution module continues to perform memory retention tests on the subjects according to the preset memory retention time, and obtains the corresponding memory response results during the memory retention test. Subsequently, the challenge task execution module statistically analyzes the correctness of the subjects' memory responses after a preset memory retention time based on the memory response results, and obtains the corresponding memory retention accuracy rate based on the proportion of correct responses. After the conversion reaction time, conversion accuracy rate, and memory retention accuracy rate are all obtained, the challenge task execution module outputs the conversion reaction time, conversion accuracy rate, and memory retention accuracy rate to the practice correction marking module.

[0030] Among them: Duration change results: refers to the change in the duration of the stable period of cognitive function in different consecutive follow-up periods; Changes in the continuity of task completion: refers to the changes in the degree of stability in the continuous completion of tasks by subjects in continuous cognitive tasks; Task completion stability variation: refers to the degree of fluctuation in the subject's task completion across different cognitive tasks; Task response variation: refers to the trend of task response time changes in different cognitive tasks; Changes in memory retention: This refers to the declining trend in a subject's memory retention ability at different memory stages; Stable changes in memory retention: refers to the changes in the stability of a subject's memory retention ability across different memory stages; Permutation and combination processing: refers to the process of combining and sorting multiple cognitive tasks according to a preset order; Average task response time: refers to the average response time of the subject during the switching between multiple tasks; The percentage of correct answers refers to the proportion of correct answers out of the total number of answers.

[0031] It should be noted that by dynamically determining the task duration range, task switching frequency, and memory retention detection interval based on the cognitive parameter set, the cognitive challenge task can be made more consistent with the subject's current cognitive ability changes, thereby improving the adaptability of the cognitive challenge process. At the same time, by continuously extracting data on task response behavior during task switching and jointly obtaining conversion reaction time, conversion accuracy, and memory retention accuracy, the changes in the subject's cognitive switching ability, task execution ability, and memory retention ability can be reflected simultaneously, thereby improving the comprehensiveness of early cognitive abnormality identification in Alzheimer's disease and the accuracy of risk prediction.

[0032] Step 3: The practice correction labeling module is used to input the conversion reaction time, conversion accuracy, and memory retention accuracy into a preset comprehensive evaluation model to obtain a practice correction index, and generate a cognitive state label based on the comparison result of the practice correction index and a first preset threshold. Further: the conversion reaction time, the conversion accuracy, and the memory retention accuracy are obtained, and the conversion reaction time, the conversion accuracy, and the memory retention accuracy are matched according to a preset data correspondence to form cognitive challenge assessment data; Based on the cognitive challenge assessment data, the changes in the subjects' task performance at different cognitive challenge stages were statistically analyzed, and task adaptation change characteristics were extracted based on the changes in task performance. Based on the task adaptation and change characteristics, the study analyzes the changes in the impact of practice on the subjects during continuous cognitive challenges, and generates a practice impact assessment result based on the changes in the impact of practice. The conversion reaction time, conversion accuracy, and memory retention accuracy are input into a preset comprehensive evaluation model, and the evaluation process of the comprehensive evaluation model is corrected in combination with the evaluation results of the practice impact, so as to generate a corresponding practice correction index. Obtain the first preset threshold, and determine the change in the subject's cognitive state based on the comparison result between the practice correction index and the first preset threshold; Generate cognitive state labels based on the changes in cognitive state; Furthermore: calculate the difference between the practice correction index and the first preset threshold; When the difference is within the first preset difference range, the change in the subject's cognitive state is determined to be a stable cognitive state. When the difference is greater than the upper limit of the first preset difference range, the change in the subject's cognitive state is determined to be an improvement in cognitive state; When the difference is less than the lower limit of the first preset difference interval, the change in the subject's cognitive state is determined to be a decline in cognitive state.

[0033] In step three: the preset data correspondence refers to the preset correlation between conversion reaction time, conversion accuracy, and memory retention accuracy; Association matching processing: refers to the process of associating multiple data according to a preset data correspondence; Cognitive challenge assessment data refers to comprehensive assessment data formed by the correlation between conversion reaction time, conversion accuracy, and memory retention accuracy. Task performance changes: refers to the changes in the subject's task performance at different cognitive challenge stages; Task adaptation change characteristics: refers to the characteristics of changes in a subject's task adaptation ability during a continuous cognitive challenge process; Changes in the impact of practice: refers to the changing trend of the impact of practice on subjects due to continuous cognitive challenges; Practice impact assessment results: refers to the results of practice impact analysis based on changes in practice impact; Comprehensive evaluation model: refers to a model used to comprehensively evaluate conversion reaction time, conversion accuracy, and memory retention accuracy; Correction processing: refers to the process of correcting the output of the comprehensive evaluation model based on the results of the exercise impact assessment; Practice Correction Index: refers to the cognitive assessment results after correction through practice; First preset threshold: refers to the preset reference threshold used to judge changes in cognitive state; Changes in cognitive state: refers to the change in a subject's current cognitive state relative to their historical cognitive state; Cognitive state labeling: refers to state identification information used to represent the results of changes in cognitive state; First preset difference range: refers to the preset difference range used to determine the range of changes in cognitive state; Cognitive stability: refers to a state in which the subject's current cognitive state does not change significantly; Improved cognitive state: refers to a state outcome in which the subject's current cognitive state shows an improving trend compared to their historical cognitive state; Cognitive decline: refers to a state in which the subject's current cognitive state shows a downward trend compared to his / her historical cognitive state.

[0034] For example: The practice correction labeling module acquires the conversion reaction time, conversion accuracy, and memory retention accuracy, and performs association matching processing on the conversion reaction time, conversion accuracy, and memory retention accuracy according to the preset data correspondence. In the association matching process, the conversion reaction time, conversion accuracy, and memory retention accuracy are associated according to the task execution order corresponding to the same cognitive challenge stage to form corresponding cognitive challenge assessment data. After the cognitive challenge assessment data is formed, the changes in the subject's task performance in different cognitive challenge stages are statistically analyzed based on the cognitive challenge assessment data. In the statistical process, the practice correction labeling module analyzes the changes in the subject's task response, task answering, and memory retention in continuous cognitive challenge stages, and extracts the corresponding task adaptation change features based on the continuity of changes in different cognitive challenge stages. Subsequently, the practice correction labeling module continued to analyze the changes in the impact of practice on the subjects during the continuous cognitive challenge based on the characteristics of task adaptation changes. During the analysis, the practice correction labeling module statistically analyzed the changes in the subjects' task familiarity, task response adaptation, and memory retention adaptation in multiple consecutive cognitive challenges, and formed corresponding practice impact assessment results based on the changes in each item. After the practice impact assessment results are generated, the practice correction labeling module further constructs corresponding comprehensive assessment input sequences based on conversion reaction time, conversion accuracy, and memory retention accuracy, and divides the comprehensive assessment input sequences in different cognitive challenge stages according to a preset time sequence. Subsequently, the practice correction labeling module statistically analyzes the changes in conversion reaction time, conversion accuracy, and memory retention accuracy of the subjects in different cognitive challenge stages, and generates corresponding stage change results based on the changes in each stage. After the stage change results are generated, the practice correction labeling module further analyzes the convergence of changes in the subjects during the continuous cognitive challenge process based on the stage change results in the continuous cognitive challenge stages. The system assesses the subject's tendency to conform to changes and their stability, and generates corresponding practice influence correction coefficients based on these coefficients. Subsequently, the practice correction labeling module adjusts the conversion reaction time, conversion accuracy, and memory retention accuracy based on these practice influence correction coefficients to reduce the impact of continuous repetitive cognitive challenges on task familiarity. During the adjustment process, when the subject's tendency to conform to changes and their stability increase continuously during the continuous cognitive challenge phase, the practice correction labeling module increases the correction intensity corresponding to the practice influence correction coefficient; conversely, when the subject's tendency to conform to changes decreases or their stability declines during the continuous cognitive challenge phase, the practice correction labeling module decreases the correction intensity corresponding to the practice influence correction coefficient. After the practice impact assessment results are generated, the practice correction labeling module further inputs the conversion reaction time, conversion accuracy, and memory retention accuracy into the comprehensive assessment model, and performs correction processing on the comprehensive assessment model's assessment process in conjunction with the practice impact assessment results. During the correction process, the practice correction labeling module corrects the task familiarity impact caused by continuous cognitive challenges based on the practice impact assessment results, so as to reduce the interference of practice impact on cognitive assessment results. Subsequently, the comprehensive evaluation model outputs the corresponding practice correction index based on the conversion reaction time, conversion accuracy, and memory retention accuracy after correction processing. After the practice correction index is generated, the practice correction labeling module further obtains a first preset threshold and calculates the difference between the practice correction index and the first preset threshold. Then, the practice correction labeling module determines the corresponding cognitive state change of the subject based on the correspondence between the difference and the first preset difference interval. When the difference is within the first preset difference interval, the practice correction labeling module determines the subject's cognitive state change as stable. When the difference is greater than the upper limit of the first preset difference interval, the practice correction labeling module determines the subject's cognitive state change as improved. When the difference is less than the lower limit of the first preset difference interval, the practice correction labeling module determines the subject's cognitive state change as decreased. After the cognitive state change is determined, the practice correction labeling module generates the corresponding cognitive state label based on the corresponding cognitive state change result and outputs the practice correction index and cognitive state label to the decline cross-validation module and the risk parameter tuning prediction module.

[0035] Specifically: The comprehensive evaluation model adopts a linear weighted model: First, the conversion reaction time is standardized to obtain a standardized value; simultaneously, the conversion accuracy and memory retention accuracy are directly used. The original comprehensive score equals the standardized value of the conversion reaction time multiplied by a first preset weight, plus the conversion accuracy multiplied by a second preset weight, plus the memory retention accuracy multiplied by a third preset weight. Here, the first, second, and third preset weights are three pre-defined weight parameters, and their sum is one. The correction process employs a multiplicative correction method: the practice correction index equals the product of the original composite score and the correction coefficient, where the correction coefficient equals the preset correction parameter multiplied by the practice impact assessment result. The practice impact assessment result reflects the magnitude of the practice effect produced by the repetitive task during continuous cognitive challenges. The specific calculation method for the impact of practice on the assessment results is as follows: The results are obtained by statistically analyzing the trends in reaction time and accuracy of subjects across multiple cognitive challenges. The regression slopes of reaction time and accuracy with the number of challenges are calculated. These two slopes are then weighted and combined, and then transformed to the range of zero to one using a pre-defined mapping function. The larger the absolute value of the slope and the more consistent the direction with the practice effect (decreasing reaction time, increasing accuracy), the closer the assessment result of the impact of practice is to one; conversely, the smaller the slope, the closer it is to zero. The specific forms of the three weighting parameters, the pre-defined correction parameters, and the mapping function are pre-calibrated using the subjects' historical data.

[0036] Among them: Task execution order: refers to the order in which tasks are executed in the cognitive challenge task sequence; Task response variation: refers to the trend of task response time changes in subjects during the continuous cognitive challenge phase; Task response changes: refers to the trend of changes in the correctness of the subjects' task responses during the continuous cognitive challenge phase; Changes in memory retention: refers to the trend of changes in the subject's memory retention ability during the continuous cognitive challenge phase; Continuity of change: refers to the degree to which the trend of change remains consistent across different stages of cognitive challenge; Changes in task familiarity: This refers to the trend of changes in task familiarity among subjects during multiple consecutive cognitive challenges; Task response adaptation changes: refers to the trend of changes in the degree of adaptation of subjects to tasks in continuous cognitive challenges; Memory retention adaptation changes: refers to the trend of changes in the degree of memory retention adaptation of the subject in continuous memory retention tests; Task familiarity effect: refers to the task familiarity effect formed due to repeated cognitive challenges.

[0037] It should be noted that by performing correlation matching on conversion reaction time, conversion accuracy, and memory retention accuracy, and further analyzing the changes in the impact of practice on subjects during continuous cognitive challenges, the interference of practice on cognitive assessment results caused by repeated cognitive challenges can be effectively reduced, thereby improving the accuracy of the practice correction index in reflecting changes in real cognitive state. At the same time, by generating cognitive state labels based on the difference between the practice correction index and the first preset threshold, the trend of changes in the subject's current cognitive state can be classified and identified, thus providing a more stable and reliable reference for cognitive state in the subsequent Alzheimer's disease risk prediction process.

[0038] Step 4: Collect cognitive decline data of subjects, including brain metabolism data and language function decline data. Based on the matching degree between the brain metabolism data and language function decline data, generate a cognitive decline cross-validation index. Further: Collect brain metabolism data of the subjects during the follow-up period, and statistically analyze the changes in brain metabolism in different brain regions based on the brain metabolism data to form brain metabolism change data; Data on the decline of language function in the subjects during the follow-up period were collected, and the changes in language function in different language tasks were statistically analyzed based on the data on the decline of language function to form data on changes in language function. Based on the brain metabolism change data, the direction of brain metabolism change in different brain regions of the subjects was extracted, and based on the language function change data, the direction of language function change in different language tasks of the subjects was extracted. The direction of brain metabolic changes is correlated and matched with the direction of language function changes according to the corresponding follow-up period to form a corresponding relationship of decline changes; Based on the aforementioned correlation between the decline changes, the consistency between the direction of brain metabolism changes and the direction of language function changes is statistically analyzed, and brain-language matching results are generated based on the statistical results. The matching degree between the brain metabolism data and the language function decline data is calculated based on the brain language matching results, and a cognitive decline cross-validation index is generated based on the matching degree.

[0039] In step four: Cognitive decline data: refers to the set of data used to characterize changes in the cognitive abilities of the subjects; Brain metabolic data: refers to data on brain metabolic activity in different brain regions of the subject; Language function decline data: refers to data on changes in the language function ability of subjects in different language tasks; Changes in brain metabolism: refers to the trend of changes in brain metabolic activity in different brain regions of the subject; Brain metabolic change data: refers to statistical results of brain metabolism based on changes in brain metabolism; Changes in language function: refers to the trend of changes in the language ability of test takers in different language tasks; Data on changes in language function: refers to statistical results of language function based on changes in language function; Direction of brain metabolic changes: refers to the trend of brain metabolic activity increasing, decreasing, or remaining stable; Direction of language function change: refers to the trend of language function ability changing towards improvement, decline, or stabilization; Association matching: refers to the process of associating different data according to the corresponding follow-up period; Correspondence in decline and change: refers to the correspondence between the direction of changes in brain metabolism and the direction of changes in language function; Consistent change: refers to the degree of consistency between the direction of changes in brain metabolism and the direction of changes in language function; Brain-language matching results: These refer to the matching analysis results between brain metabolism and language function based on consistent changes. Matching degree: refers to the degree of correspondence between brain metabolic data and language function decline data; Cross-validation index for cognitive decline: refers to an assessment index for cognitive decline formed by cross-validating brain metabolism data with language function decline data.

[0040] For example: The decay cross-validation module acquires the subject's brain metabolism data during the current follow-up period, and performs continuous statistical analysis on the changes in brain metabolism activity in different brain regions based on the detection results corresponding to different brain regions; during the statistical analysis, the decay cross-validation module extracts the magnitude of changes in brain metabolism activity, the direction of changes in brain metabolism activity, and the continuity of changes in brain metabolism activity in different brain regions, and forms corresponding brain metabolism change data based on the results of changes in brain metabolism activity in different brain regions; Subsequently, the decline cross-validation module further obtained language function decline data of the subjects in the current follow-up period, and performed continuous statistical analysis on the changes in language function of the subjects in different language tasks based on the language response results corresponding to different language tasks. During the statistical analysis, the decline cross-validation module extracted the magnitude of language function changes, the direction of language function changes, and the continuity of language function changes in different language tasks, and formed corresponding language function change data based on the language function change results in different language tasks. After the brain metabolism change data and language function change data are generated, the decay cross-validation module further extracts the direction of brain metabolism change in different brain regions based on the brain metabolism change data, and extracts the direction of language function change in different language tasks based on the language function change data. Subsequently, the decay cross-validation module performs association matching processing on the direction of brain metabolism change and the direction of language function change according to the time sequence of the corresponding follow-up period. During the association matching process, the decay cross-validation module counts the correspondence between the direction of brain metabolism change and the direction of language function change within the same follow-up period, and forms the corresponding decay change correspondence based on the correspondence. After the correlation of declining changes is established, the decline cross-validation module continues to statistically analyze the consistency between the direction of brain metabolism changes and the direction of language function changes. During the statistical process, when the direction of brain metabolism changes and the direction of language function changes both show a downward trend within the same follow-up period, the decline cross-validation module determines that the degree of consistency in the corresponding follow-up period is enhanced; when the direction of brain metabolism changes and the direction of language function changes show different trends within the same follow-up period, the decline cross-validation module determines that the degree of consistency in the corresponding follow-up period is reduced. Subsequently, the decline cross-validation module generates corresponding brain-language matching results based on the consistency in different follow-up periods. After the brain-language matching results are generated, the decline cross-validation module further calculates the matching degree between brain metabolic data and language function decline data based on the brain-language matching results. During the calculation process, the decline cross-validation module counts the proportion of consistent changes in different follow-up periods and determines the corresponding matching degree based on the proportion of consistent changes. Subsequently, the decline cross-validation module generates the corresponding cognitive decline cross-validation index based on the matching degree and outputs the cognitive decline cross-validation index to the subjective verification and standardization module and the risk parameter tuning and prediction module.

[0041] Among them: the magnitude of change in brain metabolic activity: refers to the magnitude of change in brain metabolic activity during different follow-up periods; Continuity of changes in brain metabolic activity: refers to the degree of consistency in the direction of changes in brain metabolic activity across multiple follow-up periods; Range of language function changes: refers to the magnitude of changes in language function across different language tasks; Continuity of language function change: refers to the degree of consistency in the direction of language function change across multiple language tasks; Increased consistency in changes: This refers to an increased degree of consistency between the direction of changes in brain metabolism and the direction of changes in language function. Decreased consistency in change: This refers to a decrease in the degree of consistency between the direction of changes in brain metabolism and the direction of changes in language function. Percentage of consistent changes: This refers to the percentage of consistent changes across all follow-up periods.

[0042] It should be noted that by jointly collecting brain metabolism data and language function decline data, and further performing cross-validation analysis on the consistency between the direction of changes in brain metabolism and the direction of changes in language function, the problem of misjudgment of cognitive decline caused by single cognitive test results can be avoided, thereby improving the stability and reliability of cognitive decline assessment results. At the same time, by generating cognitive decline cross-validation indicators based on the matching degree between brain metabolism data and language function decline data, the correlation between brain function decline and language function decline in subjects can be more accurately reflected, thereby improving the early cognitive decline identification ability and risk prediction accuracy in the Alzheimer's disease risk prediction process.

[0043] Step 5: Obtain the self-reported answers from the subjects and the peer-reported answers from informed individuals. Calculate the difference score based on the degree of deviation between the self-reported answers and the peer-reported answers. Further: Obtain the self-assessment answers of the subjects during the pre-set questionnaire assessment process, as well as the other-assessment answers of the informed parties corresponding to the subjects during the same pre-set questionnaire assessment process; Based on the correspondence between the self-assessment answers and the peer-assessment answers under the same assessment dimension, the self-assessment answers and the peer-assessment answers are matched item by item to form an assessment answer matching sequence; For each matching item in the evaluation answer matching sequence, calculate the degree of deviation between the self-evaluation answer and the other-evaluation answer, and generate a single-item deviation result; The overall deviation level is calculated based on the aggregated statistical results of each individual deviation, and the difference score is obtained.

[0044] In step five: self-assessment answers: refer to the subjects' autonomous responses to each assessment question during the pre-designed questionnaire assessment process; Informed person: refers to a person who has long been aware of changes in the subject's daily cognitive state; Others' evaluation of answers: refers to the evaluation results formed by informed individuals in response to the cognitive state of the subjects during the pre-designed questionnaire assessment process; Pre-set questionnaire assessment process: refers to the cognitive state assessment process carried out based on a pre-set cognitive assessment questionnaire; Assessment dimensions: These refer to the different cognitive assessment directions corresponding to the pre-designed questionnaire assessment process; Item-by-item matching: refers to matching self-assessment answers with other-assessment answers one-to-one based on the same assessment dimensions; The evaluation answer matching sequence refers to the sequence of answers formed according to the order of evaluation dimensions. Degree of deviation: refers to the degree of difference between self-assessed and other-assessed answers; Single-item deviation results: refers to the deviation analysis results of answers in a single assessment dimension; Overall deviation level: refers to the combined degree of deviation of all individual deviation results; Difference score: refers to the assessment result of cognitive state differences based on the overall level of deviation.

[0045] For example: The difference score calculation module obtains the self-evaluation answers of the subjects in the preset questionnaire assessment process, and simultaneously obtains the other-evaluation answers of the corresponding informed persons in the same preset questionnaire assessment process; during the acquisition process, the difference score calculation module performs corresponding reading and sorting processing on the self-evaluation answers and other-evaluation answers according to the assessment dimension order in the preset questionnaire assessment process to form the corresponding assessment answer basic data; Subsequently, the difference score calculation module performs item-by-item matching on the self-assessment answers and the other-assessment answers based on the correspondence between the self-assessment answers and the other-assessment answers under the same assessment dimension. During the item-by-item matching process, the difference score calculation module associates the self-assessment answers and the other-assessment answers according to the order of the assessment questions corresponding to the same assessment dimension, and forms the corresponding assessment answer matching sequence based on the correspondence in each assessment dimension. After the evaluation answer matching sequence is formed, the difference score calculation module further calculates the degree of deviation between the self-evaluation answer and the other-evaluation answer for each matching item in the evaluation answer matching sequence. During the calculation process, the difference score calculation module respectively counts the direction of answer difference, the magnitude of answer difference, and the continuity of answer difference between the subject's self-evaluation result and the informed other-evaluation result, and generates corresponding single-item deviation results based on the differences in different evaluation dimensions. Subsequently, the difference score calculation module performs summary statistical analysis based on the deviation results of each item. During the statistical analysis, the difference score calculation module statistically analyzes the distribution of individual deviation results, the concentration of individual deviation results, and the continuous changes of individual deviation results in different evaluation dimensions, and calculates the overall deviation level based on the statistical results. After the overall deviation level is calculated, the difference score calculation module further generates the corresponding difference score based on the overall deviation level and outputs the difference score to the subjective verification and standardization module.

[0046] Among them: Basic assessment data: refers to the set of basic assessment data formed by self-assessment answers and other-assessment answers; Direction of answer difference: refers to the direction of change in the difference between self-evaluated and other-evaluated answers; Scope of difference in answers: refers to the magnitude of the difference between self-evaluated and other-evaluated answers; Continuity of answer differences: refers to the degree of consistency in the variation of differences between self-assessed and other-assessed answers across multiple assessment dimensions; Distribution of individual deviation results: refers to the distribution of individual deviation results across different evaluation dimensions; Clustering of individual deviations: refers to the degree of clustering among multiple individual deviations; Continuous change of individual deviation results: refers to the continuous change trend of individual deviation results during the continuous evaluation process.

[0047] It should be noted that by jointly obtaining the self-assessment answers of the subjects and the peer-assessment answers of informed individuals, and further performing item-by-item matching analysis on the degree of deviation of answers in different assessment dimensions, the differences between the subjects' subjective cognitive state and the externally observed cognitive state can be identified more accurately, thereby reducing the error impact of a single subjective cognitive assessment. At the same time, by generating a difference score based on the summary statistical results of each individual deviation result, the abnormal changes in the subjects' cognitive state perception can be reflected more stably, thereby improving the ability to identify subjective cognitive abnormalities and the accuracy of risk prediction in the process of predicting the risk of Alzheimer's disease.

[0048] Step Six: Based on the cognitive decline cross-validation index, perform credibility verification on the difference score, and generate standardized subjective data based on the verification results; Further: Obtain the difference score and the cognitive decline cross-validation index, and determine the validation weight coefficient based on the confidence distribution characteristics corresponding to the cognitive decline cross-validation index; The difference scores are weighted and verified based on the verification weight coefficients to generate corresponding credibility verification results. The credibility status of subjective data is determined based on the comparison between the credibility verification result and the preset credibility threshold. Based on the credibility status of the subjective data, the difference scores are subjected to credibility standardization processing to generate standardized subjective data.

[0049] In step six: Credibility distribution characteristics: refers to the distribution of the credibility of the cross-validation index in different cognitive decline states; Verification weight coefficient: refers to the credibility correction weight parameter determined based on the credibility distribution characteristics; Weighted confidence verification processing: refers to the process of correcting the confidence of difference scores based on the verification weight coefficients; Credibility verification result: refers to the verification analysis result formed after the difference scores have undergone credibility verification; Preset credibility threshold: refers to the preset reference threshold used to judge the credibility status of subjective data; Subjective data credibility status: refers to the credibility level of the subjective evaluation results corresponding to the difference scores; Credibility standardization: refers to the process of adjusting the credibility of differing scores according to the credibility status of subjective data. Standardized subjective data: refers to subjective cognitive assessment data formed after credibility standardization processing.

[0050] For example: The subjective verification standardization module obtains the difference score and the cross-validation index of cognitive decline, and performs statistical analysis on the changes in credibility in different cognitive decline states based on the credibility distribution characteristics corresponding to the cross-validation index of cognitive decline; during the statistical analysis, the subjective verification standardization module extracts the credibility concentration, credibility fluctuation and credibility continuous change in different cognitive decline states, and determines the corresponding verification weight coefficient based on the credibility distribution results in different cognitive decline states. After the verification weight coefficients are determined, the subjective verification standardization module further performs weighted credibility verification processing on the difference scores based on the verification weight coefficients. During the weighted credibility verification process, the subjective verification standardization module counts the correspondence between the difference scores and the cognitive decline cross-validation index, and adjusts the credibility correction strength corresponding to the difference scores based on the correspondence. When the credibility distribution characteristics corresponding to the cognitive decline cross-validation index remain stable, the subjective verification standardization module increases the credibility correction strength corresponding to the verification weight coefficients. When the credibility distribution characteristics corresponding to the cognitive decline cross-validation index fluctuate, the subjective verification standardization module decreases the credibility correction strength corresponding to the verification weight coefficients. Subsequently, the subjective verification standardization module generates the corresponding credibility verification results based on the results after completing the weighted credibility verification processing. After the credibility verification result is generated, the subjective verification standardization module further obtains the preset credibility threshold and determines the corresponding subjective data credibility status based on the comparison relationship between the credibility verification result and the preset credibility threshold. During the judgment process, when the credibility verification result is higher than the preset credibility threshold, the subjective verification standardization module determines that the subjective data credibility status corresponding to the difference score is a high credibility status; when the credibility verification result is lower than the preset credibility threshold, the subjective verification standardization module determines that the subjective data credibility status corresponding to the difference score is a low credibility status. Subsequently, the subjective verification and standardization module performs credibility standardization processing on the difference scores based on the credibility status of the subjective data. During the credibility standardization process, the subjective verification and standardization module performs corresponding credibility correction processing on the difference scores according to the credibility adjustment rules corresponding to different credibility statuses of subjective data, so as to form standardized subjective data under a unified credibility scale. After the standardized subjective data is generated, the subjective verification and standardization module outputs the standardized subjective data to the risk parameter tuning and prediction module.

[0051] Specifically: The weighted confidence verification process uses a multiplicative weighting method: Let the difference score be one value, and the verification weight coefficient be another value. Then, the weighted difference score equals the difference score multiplied by the verification weight coefficient. The verification weight coefficient ranges from zero to one. The verification weight coefficient is determined based on the matching degree corresponding to the cognitive decline cross-validation index: when the matching degree is greater than 0.7, the verification weight coefficient is 0.9; when the matching degree is less than 0.3, the verification weight coefficient is 0.5; and in all other cases, the verification weight coefficient is 0.7. The credibility standardization process uses linear normalization: Let the minimum weighted difference score in the historical dataset be the minimum reference value, and the maximum weighted difference score be the maximum reference value. Then, the standardized subjective data equals the weighted difference score minus the minimum reference value, divided by the difference between the maximum and minimum reference values. The standardized subjective data ranges from zero to one.

[0052] Among them: Credibility concentration: refers to the degree of concentration of credibility results in different cognitive decline states; Credibility fluctuation: refers to the degree of fluctuation in credibility results across different states of cognitive decline; Continuous changes in credibility: refers to the degree of consistency in the trend of credibility change across multiple states of cognitive decline; Correspondence consistency: refers to the degree of correspondence consistency between the difference score and the cross-validation index of cognitive decline; Confidence correction strength: refers to the degree of adjustment made by the verification weight coefficient when performing confidence correction on the difference score; Highly credible state: refers to a state result corresponding to subjective data with a high degree of credibility; Low credibility state: refers to a state result corresponding to subjective data with a low degree of credibility; Credibility adjustment rules: refer to the preset credibility correction rules for different subjective data credibility states; Unified credibility scale: refers to a standardized reference scale used to unify data standards under different credibility states.

[0053] It should be noted that by performing weighted credibility verification on the difference scores based on the cognitive decline cross-validation index, the impact of subjective cognitive bias and observer bias on the cognitive assessment results can be effectively reduced, thereby improving the reliability of the subjective cognitive assessment results. At the same time, by performing credibility standardization on the difference scores based on the credibility status of the subjective data, the subjective cognitive data formed under different credibility levels can maintain a uniform assessment scale, thereby improving the stability of multi-source cognitive data fusion analysis and the accuracy of risk prediction in the subsequent Alzheimer's disease risk prediction process.

[0054] Step 7: Input the practice correction index, the cognitive state marker, the cognitive decline cross-validation index, and the standardized subjective data into the preset Alzheimer's disease risk prediction model to obtain the subject's brain disease risk level, i.e., the subject's Alzheimer's disease risk level, and adjust the cognitive parameter set in the next follow-up period based on the Alzheimer's disease risk level. Further: the practice correction index, the cognitive state label, the cognitive decline cross-validation index, and the standardized subjective data are obtained, and the input data are uniformly time-series aligned to form a risk prediction input data set; Based on the risk prediction input data set, the practice correction index, the cognitive decline cross-validation index, and the standardized subjective data are subjected to feature fusion processing to generate a comprehensive risk feature vector; The practice correction index, the cognitive state marker, the cognitive decline cross-validation index, and the standardized subjective data are input into a preset Alzheimer's disease risk prediction model, and the model outputs the corresponding risk score result based on the comprehensive risk feature vector. The risk level of the subject for Alzheimer's disease is determined based on the risk score results, and the set of cognitive parameters in the next follow-up period is adjusted based on the risk level of Alzheimer's disease.

[0055] In step seven: Unified time alignment processing: refers to the process of performing unified time sequence processing on multiple input data according to the corresponding follow-up period; Risk prediction input data set: refers to the risk prediction data set after unified time series alignment processing; Feature fusion processing: refers to the process of performing correlation and fusion on multiple risk prediction data; Comprehensive risk feature vector: refers to comprehensive risk feature data formed by fusing multiple risk prediction data; Alzheimer's disease risk prediction model: refers to a risk prediction model used to output the risk of developing Alzheimer's disease; Risk score results: refers to the risk score data output by the Alzheimer's disease risk prediction model; Alzheimer's disease risk level: refers to the risk level result formed based on the risk score; Cognitive parameter set adjustment: refers to the process of adjusting the cognitive parameter set for the next follow-up period based on the risk level of Alzheimer's disease. Next follow-up period: refers to the subsequent consecutive follow-up periods after the current follow-up period.

[0056] For example: The risk parameter tuning and prediction module acquires the practice correction index and cognitive state markers, and simultaneously acquires the cognitive decline cross-validation index and standardized subjective data; subsequently, the risk parameter tuning and prediction module performs unified time-series alignment processing on the practice correction index, cognitive state markers, cognitive decline cross-validation index, and standardized subjective data according to the follow-up period time sequence corresponding to each input data; during the unified time-series alignment processing, the risk parameter tuning and prediction module statistically analyzes the time distribution, time correspondence, and time continuous change of different input data, and forms the corresponding risk prediction input data set based on the time correspondence in the same follow-up period; After the risk prediction input data set is formed, the risk parameter tuning prediction module further performs feature fusion processing on the practice correction index, cognitive decline cross-validation index, and standardized subjective data based on the risk prediction input data set. During the feature fusion processing, the risk parameter tuning prediction module statistically analyzes the cognitive state change trend corresponding to the practice correction index, the cognitive decline consistency change trend corresponding to the cognitive decline cross-validation index, and the subjective cognitive change trend corresponding to the standardized subjective data, and forms a comprehensive risk feature vector based on the correlation between different change trends. Subsequently, the risk tuning and prediction module further inputs the practice correction index, cognitive state markers, cognitive decline cross-validation index, and standardized subjective data into the Alzheimer's disease risk prediction model. During the model processing, the Alzheimer's disease risk prediction model uses the comprehensive risk feature vector to statistically analyze the changes in cognitive abnormalities, the continuous changes in cognitive decline, and the changes in subjective cognitive deviations of the subjects in different follow-up periods, and outputs the corresponding risk score results based on the comprehensive correlation between different changes. After the risk score is generated, the risk parameter tuning and prediction module further determines the corresponding Alzheimer's disease risk level based on the risk score. During the determination process, when the risk score is in the high-risk range, the risk parameter tuning and prediction module determines the subject's Alzheimer's disease risk level as high-risk; when the risk score is in the medium-risk range, the risk parameter tuning and prediction module determines the subject's Alzheimer's disease risk level as medium-risk; and when the risk score is in the low-risk range, the risk parameter tuning and prediction module determines the subject's Alzheimer's disease risk level as low-risk. Subsequently, the risk parameter tuning and prediction module performs cognitive parameter set adjustment processing on the cognitive parameter set for the next follow-up period based on the Alzheimer's disease incidence risk level. During this adjustment, when the Alzheimer's disease incidence risk level is high, the module increases the frequency of cognitive challenge tasks and memory retention tests in the next follow-up period; when the risk level is medium, the module maintains the task configuration intensity corresponding to the current cognitive parameter set; and when the risk level is low, the module decreases the frequency of cognitive challenge tasks and memory retention tests in the next follow-up period. After the cognitive parameter set adjustment is completed, the module outputs the adjusted cognitive parameter set to the cognitive parameter generation module.

[0057] Specifically: The Alzheimer's disease risk prediction model uses a pre-trained extreme gradient boosting classifier, also known as the XGBoost classifier; Feature fusion processing employs a direct concatenation method: the practice correction index, cognitive decline cross-validation index, standardized subjective data, and numerical codes of cognitive state labels are concatenated to form a four-dimensional comprehensive risk feature vector. The numerical coding rule for cognitive state labels is: stable cognitive state is encoded as zero, improved cognitive state as one, and declining cognitive state as two. This feature vector serves as the model input. The model outputs a risk score, ranging from zero to one; a higher risk score indicates a greater risk of developing Alzheimer's disease. The risk scoring ranges are as follows: high risk is 0.7 to 1, medium risk is 0.4 to 0.7 (excluding 0.7), and low risk is 0 to 0.4 (excluding 0.4). The cognitive parameter set adjustment rules are as follows: when the disease risk level is high, the task switching frequency in the next follow-up period is increased by 20%, and the memory retention test interval is shortened by 15%; when the disease risk level is medium, the task switching frequency and memory retention test interval in the current cognitive parameter set are maintained; when the disease risk level is low, the task switching frequency in the next follow-up period is decreased by 20%, and the memory retention test interval is extended by 15%. The adjusted cognitive parameter set is output to the cognitive parameter generation module for configuring cognitive challenge tasks in the next follow-up period.

[0058] Among them: Time distribution: refers to the time distribution of different input data in each follow-up period; Time correspondence: refers to the correspondence between different input data in the same follow-up period; Continuous change over time: refers to the continuous change trend of different input data over multiple consecutive follow-up periods; Cognitive state change trend: refers to the direction of cognitive state change corresponding to the practice correction index; Consistent trend of cognitive decline: refers to the consistent direction of cognitive decline corresponding to the cross-validation index of cognitive decline. Trends in subjective cognition: refers to the direction of change in subjective cognition corresponding to standardized subjective data; Comprehensive correlation degree: refers to the comprehensive correspondence between different changes; High-risk scoring range: refers to the range of risk scores corresponding to a high-risk level; Medium-risk scoring range: refers to the range of risk scores corresponding to the medium-risk level; Low-risk scoring range: refers to the range of risk scores corresponding to the low-risk level; Task configuration intensity: refers to the task execution intensity corresponding to the frequency of cognitive challenge tasks and the frequency of memory retention tests.

[0059] It should be noted that by performing unified temporal alignment processing on the practice correction index, cognitive state markers, cognitive decline cross-validation indicators, and standardized subjective data, and further performing feature fusion processing, unified correlation analysis among multi-source cognitive assessment data can be achieved, thereby improving the stability and accuracy of Alzheimer's disease risk prediction results. At the same time, by dynamically adjusting the cognitive parameter set in the next follow-up cycle based on the Alzheimer's disease risk level, subsequent cognitive challenge tasks can be better adapted to the changes in the subject's current cognitive state, thereby improving individual adaptability and continuous risk tracking capabilities during long-term follow-up of Alzheimer's disease.

[0060] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. An Alzheimer's disease incidence risk prediction system, characterized in that, Includes the following modules: The cognitive parameter generation module is used to determine the corresponding period of cognitive function stability of the subject based on the subject's historical follow-up records and generate a set of cognitive parameters. The challenge task execution module is used to implement a cognitive challenge task on the subject based on the cognitive parameter set, and to obtain the subject's conversion reaction time, conversion accuracy and memory retention accuracy in the cognitive challenge task; The practice correction labeling module is used to input the conversion reaction time, conversion accuracy and memory retention accuracy into a preset comprehensive evaluation model to obtain a practice correction index, and generate a cognitive state label based on the comparison result of the practice correction index and a first preset threshold. The cognitive decline cross-validation module is used to collect cognitive decline data of subjects, including brain metabolism data and language function decline data. Based on the matching degree between the brain metabolism data and language function decline data, a cognitive decline cross-validation index is generated. The difference score calculation module is used to obtain the self-rated answers of the subjects and the other-rated answers of informed individuals, and calculates the difference score based on the degree of deviation between the self-rated answers and the other-rated answers; The subjective verification and standardization module is used to verify the credibility of the difference scores based on the cognitive decline cross-validation index, and generate standardized subjective data based on the verification results. The risk parameter tuning and prediction module inputs the practice correction index, the cognitive state label, the cognitive decline cross-validation index, and the standardized subjective data into a preset Alzheimer's disease risk prediction model to obtain the subject's brain disease risk level, i.e., the subject's Alzheimer's disease risk level, and adjusts the cognitive parameter set in the next follow-up period based on the Alzheimer's disease risk level.

2. The Alzheimer's disease incidence risk prediction system according to claim 1, characterized in that, The challenge task execution module is specifically used for: The duration range of the task is determined based on the information on the stability period of cognitive function in the cognitive parameter set. Based on the historical task completion data in the cognitive parameter set, the subject's task completion fluctuations in different cognitive tasks are statistically analyzed, and the task switching frequency is determined based on the task completion fluctuations. Based on the historical memory retention data in the cognitive parameter set, the changes in memory retention of the subjects in different memory stages are statistically analyzed, and the memory retention detection interval is determined based on the changes in memory retention. Based on the duration range of the task, the frequency of task switching, and the memory retention detection interval, different cognitive tasks are arranged and combined to form a cognitive challenge task sequence; For each cognitive task in the cognitive challenge task sequence, the task switching relationship between the current cognitive task and the next cognitive task is determined, and the task response behavior data of the subject during the task switching process is extracted based on the task switching relationship to obtain the corresponding task response time data and task answer result data. The task response time data is used to statistically analyze the changes in the subjects' response time during different task switching processes, and the switching reaction time is obtained based on the results of the response time changes during each task switching process. The task response data is used to statistically analyze the subjects' correctness of task responses after task switching, and the conversion accuracy rate is obtained based on the statistical results. After the cognitive challenge task sequence is completed, the subject is tested for memory retention and the corresponding memory response results are obtained. The accuracy of the subjects' memory responses after a preset memory retention time is statistically analyzed based on the memory response results, and the memory retention accuracy rate is obtained based on the statistical results.

3. The Alzheimer's disease incidence risk prediction system according to claim 2, characterized in that, The practice correction marking module is specifically used for: The conversion reaction time, conversion accuracy, and memory retention accuracy are obtained, and correlation matching processing is performed on the conversion reaction time, conversion accuracy, and memory retention accuracy according to the preset data correspondence to form cognitive challenge assessment data; Based on the cognitive challenge assessment data, the changes in the subjects' task performance at different cognitive challenge stages were statistically analyzed, and task adaptation change characteristics were extracted based on the changes in task performance. Based on the task adaptation and change characteristics, the study analyzes the changes in the impact of practice on the subjects during continuous cognitive challenges, and generates a practice impact assessment result based on the changes in the impact of practice. The conversion reaction time, conversion accuracy, and memory retention accuracy are input into a preset comprehensive evaluation model, and the evaluation process of the comprehensive evaluation model is corrected in combination with the evaluation results of the practice impact, so as to generate a corresponding practice correction index. Obtain the first preset threshold, and determine the change in the subject's cognitive state based on the comparison result between the practice correction index and the first preset threshold; Generate cognitive state tags based on the changes in cognitive state.

4. The Alzheimer's disease incidence risk prediction system according to claim 3, characterized in that, The practice correction marking module is specifically used for: Calculate the difference between the practice correction index and the first preset threshold; When the difference is within the first preset difference range, the change in the subject's cognitive state is determined to be a stable cognitive state. When the difference is greater than the upper limit of the first preset difference range, the change in the subject's cognitive state is determined to be an improvement in cognitive state; When the difference is less than the lower limit of the first preset difference interval, the change in the subject's cognitive state is determined to be a decline in cognitive state.

5. The Alzheimer's disease incidence risk prediction system according to claim 4, characterized in that, The decay cross-validation module is specifically used for: Brain metabolism data of the subjects were collected during the follow-up period, and brain metabolism changes in different brain regions of the subjects were statistically analyzed based on the brain metabolism data to form brain metabolism change data. Data on the decline of language function in the subjects during the follow-up period were collected, and the changes in language function in different language tasks were statistically analyzed based on the data on the decline of language function to form data on changes in language function. Based on the brain metabolism change data, the direction of brain metabolism change in different brain regions of the subjects was extracted, and based on the language function change data, the direction of language function change in different language tasks of the subjects was extracted. The direction of brain metabolic changes is correlated and matched with the direction of language function changes according to the corresponding follow-up period to form a corresponding relationship of decline changes; Based on the aforementioned correlation between the decline changes, the consistency between the direction of brain metabolism changes and the direction of language function changes is statistically analyzed, and brain-language matching results are generated based on the statistical results. The matching degree between the brain metabolism data and the language function decline data is calculated based on the brain language matching results, and a cognitive decline cross-validation index is generated based on the matching degree.

6. The Alzheimer's disease incidence risk prediction system according to claim 5, characterized in that, The difference score calculation module is specifically used for: Obtain the self-assessment answers of the subjects during the pre-set questionnaire assessment process, as well as the other-assessment answers of the informed persons corresponding to the subjects during the same pre-set questionnaire assessment process; Based on the correspondence between the self-assessment answers and the peer-assessment answers under the same assessment dimension, the self-assessment answers and the peer-assessment answers are matched item by item to form an assessment answer matching sequence; For each matching item in the evaluation answer matching sequence, calculate the degree of deviation between the self-evaluation answer and the other-evaluation answer, and generate a single-item deviation result; The overall deviation level is calculated based on the aggregated statistical results of each individual deviation, and the difference score is obtained.

7. The Alzheimer's disease incidence risk prediction system according to claim 6, characterized in that, The subjective verification standardization module is specifically used for: Obtain the difference score and the cognitive decline cross-validation index, and determine the validation weight coefficient based on the confidence distribution characteristics corresponding to the cognitive decline cross-validation index. The difference scores are weighted and verified based on the verification weight coefficients to generate corresponding credibility verification results. The credibility status of subjective data is determined based on the comparison between the credibility verification result and the preset credibility threshold. Based on the credibility status of the subjective data, the difference scores are subjected to credibility standardization processing to generate standardized subjective data.

8. The Alzheimer's disease incidence risk prediction system according to claim 7, characterized in that, The risk parameter tuning and prediction module is specifically used for: The practice correction index, the cognitive state label, the cognitive decline cross-validation index, and the standardized subjective data are obtained, and the input data are uniformly time-series aligned to form a risk prediction input data set. Based on the risk prediction input data set, the practice correction index, the cognitive decline cross-validation index, and the standardized subjective data are subjected to feature fusion processing to generate a comprehensive risk feature vector; The practice correction index, the cognitive state marker, the cognitive decline cross-validation index, and the standardized subjective data are input into a preset Alzheimer's disease risk prediction model, and the model outputs the corresponding risk score result based on the comprehensive risk feature vector. The risk level of the subject for Alzheimer's disease is determined based on the risk score results, and the set of cognitive parameters in the next follow-up period is adjusted based on the risk level of Alzheimer's disease.

9. The Alzheimer's disease incidence risk prediction system according to claim 8, characterized in that, The cognitive parameter generation module is specifically used for: To obtain the historical cognitive assessment results and historical memory retention results of the subjects in their historical follow-up records; Based on the historical cognitive assessment results, the changes in the subjects' cognitive assessment during the continuous follow-up period were statistically analyzed, and the stable range of cognitive assessment was determined based on the changes in cognitive assessment. Based on the historical memory retention results, the changes in memory retention of the subjects during the continuous follow-up period were statistically analyzed, and the memory retention stability interval was determined based on the changes in memory retention. By matching the cognitive assessment stability interval with the memory retention stability interval, the corresponding cognitive function stability period of the subject is determined; Based on the stable period of cognitive function, the historical task completion status and historical memory retention status in the corresponding follow-up period are extracted, and a set of cognitive parameters is generated.

Images