Cognitive level multivariate adaptive evaluation method and system based on heart rate variability

Abstract

The invention discloses a cognitive level multivariate self-adaptive evaluation method and system based on heart rate variability, and relates to the field of cognitive evaluation and prediction, individual indexes are calculated based on personal information, equivalent environmental indexes are calculated based on environmental parameters, and HRV indexes are calculated based on electrocardiosignals; based on individual indexes and related thresholds, performing adaptive optimization from the HRV indexes, and screening to obtain three types of HRV sensitive indexes; respectively calculating correction parameters of the three types of normalized HRV sensitive indexes, and dynamically correcting the correction parameters based on the working time of the target user to obtain corrected correction parameters; calculating a comprehensive index based on the equivalent environment index; on the basis of the three types of HRV sensitive indexes and the corrected correction parameters, feature vectors are calculated; and selecting an optimal classification model based on historical data of the target user, and inputting the feature vector into the optimal classification model to obtain a current cognitive evaluation result. Personalized cognitive level evaluation is realized through real-time analysis and self-adaptive optimization of heart rate variability index characteristics.

Description

Technical Field

[0001] This invention relates to the field of cognitive assessment and prediction, and in particular to a multivariate adaptive assessment method and system for cognitive levels based on heart rate variability. Background Technology

[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.

[0003] Cognitive performance is a core indicator for measuring an individual's ability to perform cognitive tasks, and it plays a crucial role in fields with high mental demands, such as driving, medicine, aviation, and education. With the accelerating pace of society and the increasing intensity and difficulty of daily work, especially in typical office environments, people often face continuous cognitive loads, leading to a decline in cognitive performance and consequently affecting work efficiency and even mental and physical health. Therefore, finding methods suitable for evaluating the cognitive level of people engaged in daily mental work is of great significance for ensuring work efficiency, designing scientific and rational work plans, and implementing occupational health interventions.

[0004] Currently, cognitive performance can be evaluated using behavioral methods to directly observe a person's reaction time and accuracy in performing a task, or using physiological techniques such as electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS). However, behavioral indicators are based on a person's feedback on the results of a specific task, making it difficult to evaluate a person's cognitive level in real time. EEG technology has extremely high temporal resolution and highly sensitive signals, but it requires electrodes to be attached to the human head and has strict requirements on the person's activity level. FNIRS technology has lower requirements on the human activity level, but it still requires sensors to be attached to the head, and the sensors are expensive and difficult to make portable, and there is a certain delay in the signal. Heart rate variability (HRV) has also been applied to the assessment of cognitive load and cognitive state and is relatively mature. It usually only requires a chest strap or optical sensors (smartwatch), has minimal interference with the user, is non-invasive, and has high wearing comfort, making it suitable for long-term continuous monitoring. Moreover, compared with devices such as EEG and fNIRS, HRV is inexpensive, easy to deploy, and can be quickly put on by a single person, making it easy to promote. Meanwhile, in scenarios where slight physical movement is permissible, electrocardiogram (ECG) signals are more stable than electroencephalogram (EEG) signals and are less affected by motion artifacts.

[0005] However, existing HRV-based cognitive assessment methods / systems all rely on a single HRV indicator, failing to consider individual differences in HRV indicators that are sensitive to cognition, leading to significant cognitive assessment errors. Secondly, existing cognitive assessment methods / systems do not consider the impact of changes in environmental conditions; furthermore, existing methods assume that human cognitive capacity is constant and independent of working hours, failing to account for cognitive loss over time. Additionally, existing methods use a single machine learning approach to judge the output, lacking the ability to select the optimal model. Therefore, existing methods and systems cannot adapt to the confounding effects of environment, time, and individual differences, and are limited by a single model, often resulting in significant deviations in the output. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, this invention provides a multivariate adaptive evaluation method and system for cognitive level based on heart rate variability, which realizes personalized cognitive level evaluation through real-time analysis and adaptive optimization of the characteristics of heart rate variability indicators.

[0007] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: In a first aspect, the present invention provides a multivariate adaptive evaluation method for cognitive level based on heart rate variability, including: Obtain the target user's personal information, and collect environmental parameters and the target user's electrocardiogram (ECG) signal; calculate the individual index based on the personal information, calculate the equivalent environmental index based on the environmental parameters, and calculate the HRV index based on the ECG signal; Based on individual indices and relevant thresholds, adaptive optimization is performed on HRV indicators to screen out three types of HRV sensitive indicators, and then normalized three types of HRV sensitive indicators are obtained. The correction parameters for the three types of HRV sensitivity indicators after normalization are calculated respectively, and the correction parameters are dynamically adjusted based on the target user's working time to obtain the corrected parameters; the comprehensive index is calculated based on the equivalent environmental index. Based on the normalized three types of HRV sensitivity indicators and the corrected parameters, the feature vector is calculated; The optimal classification model is selected based on the target user's historical data. The feature vector is then input into the optimal classification model to obtain the current cognitive evaluation result.

[0008] In a further technical solution, the individual index is expressed as:

[0009] in, For the target user age, For gender index, BMI index Years of education , , These are the weighting coefficients.

[0010] A further technical solution is that the equivalent environmental index is expressed as:

[0011] in, For ambient temperature, For ambient relative humidity, For local airflow velocity, For illuminance, This refers to the concentration of carbon dioxide. , , , , These are the weighting coefficients.

[0012] In a further technical solution, the three types of HRV sensitivity indicators include sympathetic nerve activity characteristic indicators, parasympathetic nerve activity characteristic indicators, and autonomic nerve balance characteristic indicators.

[0013] A further technical solution, based on individual indices and relevant thresholds, adaptively optimizes the HRV index as follows: when an individual index is detected... Less than the preset individual threshold At that time, the three types of system datasets corresponding to the current target user will be... , , Extract and output; when the individual threshold is reached. Greater than or equal to the preset individual threshold But less than or equal to the preset individual threshold At that time, the three types of system datasets corresponding to the current target user will be... , , Extract and output; otherwise, extract and output the three types of system datasets corresponding to the current target user. , , Extract and output.

[0014] A further technical solution is to calculate the feature vector as follows:

[0015]

[0016]

[0017] in, , , The feature vectors of the three types of HRV sensitivity indicators , , These are the correction coefficients for the three types of HRV sensitivity indicators. , , These are the normalized values ​​of the three types of HRV sensitivity indicators. As a composite index, , , It is a constant.

[0018] Further technical solutions include classification models such as support vector machines, random forests, decision trees, and k-nearest neighbors. The performance metrics of multiple classification models are updated in real time based on the current user's historical training data. By comparing the performance metrics of different classification models, the optimal model is selected.

[0019] Secondly, this invention provides a multivariate adaptive assessment system for cognitive levels based on heart rate variability, including: The data acquisition module is configured to: acquire the target user's personal information, and collect environmental parameters and the target user's electrocardiogram (ECG) signal; calculate the individual index based on the personal information, calculate the equivalent environmental index based on the environmental parameters, and calculate the HRV index based on the ECG signal; The indicator screening module is configured to: based on individual indices and relevant thresholds, adaptively optimize HRV indicators to screen three types of HRV sensitive indicators, and then obtain the three types of normalized HRV sensitive indicators. The comprehensive index calculation module is configured to: calculate the correction parameters of the three types of HRV sensitivity indicators after normalization, and dynamically adjust the correction parameters based on the target user's working time to obtain the corrected parameters; and calculate the comprehensive index based on the equivalent environmental index. The feature vector calculation module is configured to calculate feature vectors based on the normalized three types of HRV sensitivity indicators and the corrected correction parameters. The cognitive level identification module is configured to: select the optimal classification model based on the target user's historical data, input the feature vector into the optimal classification model, and obtain the current cognitive evaluation result.

[0020] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the cognitive level multivariate adaptive evaluation method based on heart rate variability as described in the first aspect.

[0021] Fourthly, the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the cognitive level multivariate adaptive evaluation method based on heart rate variability as described in the first aspect.

[0022] The above one or more technical solutions have the following beneficial effects: This invention proposes a personalized cognitive level evaluation method based on multiple indicators of heart rate variability with adaptive function. It also considers the influence of factors such as working hours and physical environment on human cognition, which significantly improves the reliability and accuracy of cognitive level evaluation under long-term tasks.

[0023] This invention abandons the traditional single-threshold evaluation model and fully considers the interference of age, gender, BMI, educational background, and environmental factors such as temperature and illumination on heart rate variability (HRV) by introducing individual and environmental indices. By dynamically adjusting the selection and correction parameters of sensitive indicators, personalized cognitive level assessment is achieved, significantly improving the accuracy and robustness of the evaluation results.

[0024] This invention not only considers static physiological indicators but also innovatively introduces working time as a dynamic correction factor to track the consumption of cognitive resources in real time. Simultaneously, by combining the equivalent environmental index with weighted calculations of the HRV feature vector, it achieves a deep fusion of multiple dimensions—physiological, environmental, and time—making the evaluation results more closely reflect the actual cognitive state in real-world work scenarios.

[0025] This invention introduces a multi-model competition mechanism in the classification stage and dynamically selects the best-performing model based on historical user data. This dynamic selection strategy avoids the problem of insufficient generalization ability of a single model and ensures the stability and reliability of evaluation results across different individuals and time periods.

[0026] This invention ultimately divides cognitive levels into three clear levels (high, medium, and low), and provides differentiated human-computer interaction feedback (such as text prompts and vibration alerts) for the medium and low levels. This not only enables precise monitoring of cognitive states, but also provides timely and effective early warning and intervention methods to prevent operational errors caused by cognitive decline.

[0027] This invention proposes a cognitive evaluation system based on multiple indicators of heart rate variability. The system has high accuracy in identifying and monitoring multiple indicators, and can achieve professional-level functions by wearing a smartwatch and utilizing existing popular hardware. It is highly practical, has high ecological validity, and is easy to implement and promote.

[0028] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0029] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0030] Figure 1 This is a flowchart of the cognitive level multivariate adaptive evaluation method based on heart rate variability according to an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the interaction between a smartwatch and a mobile terminal according to an embodiment of the present invention. Detailed Implementation

[0031] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0032] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0033] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0034] Example 1 like Figure 1 As shown in the figure, this embodiment discloses a multivariate adaptive evaluation method for cognitive level based on heart rate variability, which includes the following steps: S1: Obtain the target user's personal information, and collect environmental parameters and the target user's electrocardiogram (ECG) signal; calculate the individual index based on the personal information, calculate the equivalent environmental index based on the environmental parameters, and calculate the HRV index based on the ECG signal; In this embodiment, the user enters their personal information or extracts or updates it based on historical data, including age. Gender Index BMI index Years of education Calculate the individual index of the target user. To achieve adaptive adaptation to individual differences, it is represented as:

[0035] in, For the target user age, For gender index, BMI index Years of education , , These are weighting coefficients. Age, gender, BMI, and years of education have been shown to be significant factors affecting HRV. Additionally, , , All data are derived from long-term experimental data and are representative and statistically significant. Long-term experimental data includes test data and individual baseline calibration data collected in real-world work scenarios or laboratory simulation scenarios.

[0036] In addition, data is collected synchronously through sensors built into the wearable detection device. First, the air temperature at the current location is continuously collected using an environmental variable sensor. ,humidity Local airflow velocity Illuminance and carbon dioxide concentration Environmental parameters, and calculate the equivalent environmental index. :

[0037] in, For ambient temperature, For ambient relative humidity, For local airflow velocity, For illuminance, This refers to the concentration of carbon dioxide. , , , , These are weighting coefficients. Temperature, humidity, wind speed, illuminance, and carbon dioxide concentration have been proven to be factors that significantly affect perception. Additionally, , , , , All data are derived from long-term experimental data and are representative and statistically significant. Long-term experimental data includes environmental data collected in real-world work scenarios or laboratory simulation scenarios.

[0038] Simultaneously, the ECG sensor continuously collects ECG signals and records the working time T, and the data acquisition module extracts the RR interval sequence and heart rate data in real time.

[0039] Furthermore, the extracted RR interval sequence is preprocessed. First, filtering is performed to remove baseline drift, power line interference, and electromyographic noise. Then, outliers are identified, and interpolation is applied to the removed outliers to obtain a stable RR interval sequence for subsequent calculation of the target user's cognition. A sliding window method is used to segment the processed RR interval sequence into a series of continuous short time intervals, with the window length and sliding step size set accordingly. In the feature extraction and classification stage, HRV time-domain, frequency-domain, and nonlinear indices are calculated based on the processed RR intervals, including multiple types of heart rate variability indices such as SDNN, RMSSD, pNN50, LF, HF, LF / HF ratio, SampEn, SD1, and SD2. These indices are then grouped into datasets according to their corresponding systems, and all indices are used for HRV feature classification.

[0040] By incorporating individual differences such as age, gender, BMI, and years of education, the baseline drift problem of HRV indices varies from person to person, providing a data basis for subsequent adaptive optimization. By quantifying external environmental interferences such as air temperature, humidity, wind speed, illuminance, and CO2 concentration, the confounding effect of environmental factors on cognitive load assessment is eliminated, making HRV changes more purely reflect changes in internal cognitive state. Multi-dimensional HRV features (such as LF, HF, RMSSD, and SampEn) including time domain, frequency domain, and nonlinearity are extracted, providing a rich dataset for comprehensively characterizing autonomic neural activity.

[0041] S2: Based on individual indices and relevant thresholds, adaptive optimization is performed on HRV indicators to select three types of HRV sensitive indicators, and then the normalized three types of HRV sensitive indicators are obtained. In this embodiment, the main purpose of this step is to screen HRV indicators that are sensitive to the current user's cognitive level and to normalize the HRV-sensitive indicator data.

[0042] The extracted HRV features were divided into three categories: sympathetic neural activity feature indicators, etc. Characteristic indicators of parasympathetic nerve activity Characteristic indicators of autonomic nervous system balance This classification method is used to extract sensitive indicators for subsequent calculations of cognitive levels. It is based on sensitive HRV indicators corresponding to three types of neural activity: LF (sympathetic nervous system), RMSSD (parasympathetic nervous system), and LF / HF (autonomic nervous system balance). To avoid the influence of units on subsequent calculations, each indicator is normalized.

[0043] Based on the calculated individual index of the target user Threshold judgment is applied to it: when an individual index is detected. Less than the preset individual threshold At that time, the three types of system datasets corresponding to the current target user will be... , , Extract and output; when the individual threshold is reached. Greater than or equal to the preset individual threshold But less than or equal to the preset individual threshold At that time, the three types of system datasets corresponding to the current target user will be... , , Extract and output; otherwise, extract and output the three types of system datasets corresponding to the current target user. , , Extract and output. Through different individuals The value achieves adaptive optimization of the HRV metric.

[0044] The HRV index in each system dataset should be dynamically determined based on user status, and the filtering rules are as follows: Through experiments involving long-duration cognitive tasks, HRV data was collected during the experiments. Correlation analysis was then performed to identify HRV indicators significantly related to cognition, categorizing them into three groups. Simultaneously, p-values ​​were calculated to summarize which HRV indicators were included in the three system datasets under different p-values. Zero or one of the most significant indicators from each system dataset was selected for subsequent processing. Further filtering was conducted using large databases related to the above experiments and real-time HRV data from users.

[0045] It is important to note that, to ensure the constructed model is free from collinearity, only one or zero sensitive indicators are selected from each of the three sets of indicators representing autonomic neural activities, with preset individual thresholds. and For configurable thresholds, under each threshold... , , All of these were determined through analysis of long-term experimental data and relevant large databases.

[0046] After screening the sensitive indicators, to avoid significant errors in calculations due to different units, the screened indicators are now normalized. The calculation formula is as follows:

[0047]

[0048]

[0049] in, , , This is the current HRV indicator data. , , , , , These are the maximum and minimum values ​​of the corresponding indicators determined through big data or historical user data. , , The adjustment coefficient is calibrated based on big data or experimental data; , , This represents the normalized value of the three types of system indicators (HRV sensitivity indicators).

[0050] By utilizing the aforementioned technical features, different sets of sensitive indicators are dynamically invoked based on the different threshold ranges within which the calculated individual index falls. This mechanism ensures that the selected HRV indicators are the most sensitive and representative for the specific individual at that time, avoiding evaluation distortion caused by using universal indicators.

[0051] S3: Calculate the correction parameters for the three types of HRV sensitivity indicators after normalization, and dynamically adjust the correction parameters based on the target user's working time to obtain the corrected parameters. In this embodiment, based on the three types of system indicators selected... , , Calculate its correction parameters The correction parameters for the three types of systems are dynamically adjusted based on the target user's working time T. The calculation model is as follows: for :

[0052] for :

[0053] for :

[0054] in, For working hours, , , These are the correction coefficients for the three types of system indicators. , , These are the weighting coefficients. , , It is a constant.

[0055] , , , , , All data are derived from long-term experimental data and are representative and statistically significant. Long-term experimental data includes data collected in real-world work scenarios or laboratory simulation scenarios.

[0056] By leveraging the aforementioned technical features and establishing a linear model of working time and correction parameters, a dynamic compensation mechanism is introduced in the time dimension. As working hours increase, even if the task difficulty remains unchanged, an individual's cognitive resources will naturally deplete, and autonomic neural response patterns will drift. This correction parameter adjusts the weights of sensitive indicators in real time, ensuring the continued accuracy of the evaluation model in long-term work scenarios.

[0057] S4: Calculate the comprehensive index based on the equivalent environmental index; In this embodiment, the composite index is used to subsequently calculate the feature vector of the HRV index. Based on the equivalent environmental index... The composite index is calculated dynamically and expressed as follows:

[0058] in, As a composite index, These are the weighting coefficients. It is a constant. , All data are derived from long-term experimental data and are representative and statistically significant. Long-term experimental data includes data collected in real-world work scenarios or laboratory simulation scenarios.

[0059] Using the aforementioned technical features, the equivalent environmental index is transformed into a comprehensive index through linear transformation. This index, acting as a dynamic adjustment coefficient, directly affects subsequent feature vector calculations. When the environment becomes harsh (e.g., high temperature, high CO2 concentration), changes in the comprehensive index amplify the impact of HRV (Human Respiratory Value) changes on cognitive evaluation, reflecting the synergistic effect of environmental stress and physiological stress.

[0060] S5: Calculate the feature vector based on the normalized three types of HRV sensitivity indicators and the corrected correction parameters; In this embodiment, based on the normalized three types of HRV sensitivity indicators and the corrected parameters, the feature vector of the HRV indicator is calculated, and expressed as:

[0061]

[0062]

[0063] in, , , These are the feature vectors of three types of system indicators. , , These are the normalized values ​​of the three types of system indicators. As a composite index, , , It is a constant. , , All data originates from long-term experimental data. Long-term experimental data includes data collected in real-world work scenarios or laboratory simulation scenarios. , , The comprehensive index calculated earlier reflects the impact of the environment on the heart rate variability index, and reflects the adaptive behavior of the heart rate variability index to the environment.

[0064] Based on the comprehensive index and personalized sensitive HRV index coefficient Calculate the feature vector of the user's current HRV index, and input the feature vector of the user's current HRV index into the optimal classification model to evaluate the cognitive level.

[0065] Through the above technical features, not only is basic HRV information included, but individual differences, time cumulative effects and environmental stress are also cleverly embedded into the feature space, enhancing the representational ability and discriminative power of the feature vector.

[0066] It should be noted that the parameters derived from the above long-term experimental data were determined through correlation analysis and linear regression using relevant experimental data (process-collected HRV) from cognitive tasks performed over a long period of time.

[0067] S6: Select the optimal classification model based on the target user's historical data, input the feature vector into the optimal classification model, and obtain the current cognitive evaluation result.

[0068] In this embodiment, after calculating the feature vectors, the feature vectors are simultaneously fed into four classification models: Support Vector Machine (SVM), Random Forest, Decision Tree, and k-Nearest Neighbors. When using SVM as the classification model, the RBF kernel function is used as the SVM kernel function. During model training, a grid search combined with ten-fold cross-validation is used to optimize the SVM parameters for the optimal penalty factor and radial basis function parameters. When using Random Forest as the classification model, multiple decision trees are used as base learners. During model training, a grid search combined with ten-fold cross-validation is used to optimize the random forest parameters for the optimal number of trees and the maximum depth. When using Decision Tree as the classification model, information gain is used as the splitting criterion. During model training, a grid search combined with ten-fold cross-validation is used to optimize the decision tree parameters for the optimal maximum depth and the minimum number of sample splits. When using k-Nearest Neighbors as the classification model, Euclidean distance is used as the distance metric. During model training, a grid search combined with ten-fold cross-validation is used to optimize the k-nearest neighbor parameters for the optimal k-value and distance metric.

[0069] The training data for the classification model consists of electrocardiogram (ECG) data of users who wear the corresponding product continuously and work normally for a week, as well as ECG data of subjects collected in real-world work scenarios or laboratory simulation scenarios.

[0070] The performance metrics of various classification models, including accuracy, are updated in real time based on the current user's historical training data. Accuracy Recall rate F1 score and AUC value By comparing the performance metrics of different classification models, the optimal model is selected to output the current cognitive evaluation result.

[0071] It should be noted that, considering individual differences, historical data needs to be updated in real time. Therefore, the classification models may change over time, leading to improvements in the optimal model. Thus, the selection of the optimal model is necessary.

[0072] Based on the classification results, the current user's cognitive ability is divided into three levels: high cognitive level, medium cognitive level (requires attention), and low cognitive level (requires intervention).

[0073] Based on the cognitive assessment results output by the classification model, the system updates the target user's current cognitive level assessment in real time on the mobile terminal page and outputs corresponding suggestions based on the current state. When the cognitive level is determined to be moderate, the system begins to display the corresponding text: (1) For target users with a moderate level of cognition, the system displays that the target user's current cognitive state is at a moderate level, but does not vibrate or ring to remind them; (2) For target users with low cognitive level, the system displays that the target user's current cognitive level is low and emits a vibration or ringtone to remind the target user to take a break.

[0074] Through the aforementioned technical features, the system automatically selects the best-performing model for decision-making during each evaluation, significantly improving the generalization ability and environmental adaptability of the evaluation results. The abstract model output is quantified into three intuitive cognitive levels: "high," "medium," and "low," with a progressive warning strategy—from text prompts to vibration alerts—set for the "medium" and "low" states. This user-friendly interactive design avoids frequent interruptions while providing effective reminders at crucial moments, achieving a leap from passive monitoring to proactive intervention.

[0075] Specific examples are as follows: like Figure 2 As shown, an office worker wore and activated a smartwatch equipped with the system of this invention when starting work. The worker was 25 years old and male (in this embodiment). The employee, with a BMI of 20 and 15 years of education, had the following environmental parameters measured after 20 minutes of work: ambient temperature 24℃, relative humidity 50%, local airflow velocity 0.5 m / s, illuminance 400 lx, and carbon dioxide concentration 800 ppm. The HRV features extracted within a sliding time window (1 minute) were: SDNN=40 ms, RMSSD=40 ms, pNN50=12%, LF=30 ms², HF=30 ms², LF / HF=1, SampEn=1.5, SD1=40 ms, SD2=50 ms, and average heart rate HR=80 bpm.

[0076] The specific calculation process is as follows: Step 1: (1) The individual index of this employee Calculation:

[0077] Setting this example , , Substituting it into the formula, we get:

[0078] (2) The employee's equivalent environmental index Calculation:

[0079] Setting this example , , , , Substituting it into the formula, we get:

[0080] Step 2: This example assumes an individual threshold. and The indices are 30 and 60 respectively. Since the individual employee index in this example is less than... Therefore, the three types of system datasets corresponding to this employee were... , , Extraction. Assume that in this embodiment, the extracted components are LF, RMSSD, and SampEn.

[0081] Determined based on big data and historical data , , , , , ;and , , Substituting into the formula, we get:

[0082]

[0083]

[0084] Step 3: This example assumes , , All are 0. , , Both are 1, substituting them into the formula yields:

[0085]

[0086]

[0087] Therefore, the employee's corresponding , , All are 1.

[0088] Step 4: This example assumes =0, Substituting 1 into the formula, we get:

[0089] Therefore, the employee's composite index at this time is 1.

[0090] Step 5: Assuming this example =1、 =1、 =1, and substituting the calculated value from the previous steps into the formula, we get:

[0091]

[0092]

[0093] Step 6: The calculated feature vectors are simultaneously fed into four classification models: Support Vector Machine, Random Forest, Decision Tree, and k-Nearest Neighbor, and the accuracy of the different classifier models is compared. Accuracy Recall rate F1 score and AUC value Assuming the support vector machine has optimal performance and accuracy... =0.8, accuracy =0.8, recall rate =0.8, F1 score =0.8 and AUC value =0.8. The current evaluation result output by the support vector machine is "high cognitive level," indicating that the employee's current cognitive level is relatively high and they can continue to handle the current task.

[0094] Based on the results of step 6, the employee can now observe text with a high level of cognitive ability on the mobile terminal without prompting or intervention.

[0095] Example 2 This embodiment discloses a multivariate adaptive assessment system for cognitive level based on heart rate variability, including: The data acquisition module is configured to: acquire the target user's personal information, and collect environmental parameters and the target user's electrocardiogram (ECG) signal; calculate the individual index based on the personal information, calculate the equivalent environmental index based on the environmental parameters, and calculate the HRV index based on the ECG signal; The indicator screening module is configured to: based on individual indices and relevant thresholds, adaptively optimize HRV indicators to screen three types of HRV sensitive indicators, and then obtain the three types of normalized HRV sensitive indicators. The comprehensive index calculation module is configured to: calculate the correction parameters of the three types of HRV sensitivity indicators after normalization, and dynamically adjust the correction parameters based on the target user's working time to obtain the corrected parameters; and calculate the comprehensive index based on the equivalent environmental index. The feature vector calculation module is configured to calculate feature vectors based on the normalized three types of HRV sensitivity indicators and the corrected correction parameters. The cognitive level identification module is configured to: select the optimal classification model based on the target user's historical data, input the feature vector into the optimal classification model, and obtain the current cognitive evaluation result.

[0096] Example 3 The purpose of this embodiment is to provide a computing device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method of Embodiment 1.

[0097] Example 4 The purpose of this embodiment is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the method of Embodiment 1.

[0098] The steps and methods involved in the apparatuses of Embodiments 3 and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0099] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0100] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0101] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A multivariate adaptive evaluation method for cognitive level based on heart rate variability, characterized in that, include: Obtain the target user's personal information, and collect environmental parameters and the target user's electrocardiogram signals; Individual indices are calculated based on personal information, equivalent environmental indices are calculated based on environmental parameters, and HRV indices are calculated based on electrocardiogram signals. Based on individual indices and relevant thresholds, adaptive optimization is performed on HRV indicators to screen out three types of HRV sensitive indicators, and then normalized three types of HRV sensitive indicators are obtained. The correction parameters for the three types of HRV sensitivity indicators after normalization are calculated respectively, and the correction parameters are dynamically adjusted based on the target user's working time to obtain the corrected parameters; the comprehensive index is calculated based on the equivalent environmental index. Based on the normalized three types of HRV sensitivity indicators and the corrected correction parameters, the feature vector is calculated; The optimal classification model is selected based on the target user's historical data. The feature vector is then input into the optimal classification model to obtain the current cognitive evaluation result.

2. The cognitive level multivariate adaptive evaluation method based on heart rate variability as described in claim 1, characterized in that, The individual index is expressed as: in, For the target user age, For gender index, BMI index Years of education , , These are the weighting coefficients.

3. The cognitive level multivariate adaptive evaluation method based on heart rate variability as described in claim 1, characterized in that, The equivalent environmental index is expressed as: in, For ambient temperature, For ambient relative humidity, For local airflow velocity, For illuminance, This refers to the concentration of carbon dioxide. , , , , These are the weighting coefficients.

4. The cognitive level multivariate adaptive evaluation method based on heart rate variability as described in claim 1, characterized in that, The three types of HRV sensitivity indicators include sympathetic nerve activity characteristic indicators, parasympathetic nerve activity characteristic indicators, and autonomic nerve balance characteristic indicators.

5. The cognitive level multivariate adaptive evaluation method based on heart rate variability as described in claim 1, characterized in that, Adaptive optimization of HRV metrics based on individual indices and relevant thresholds specifically involves: when an individual index is detected... Less than the preset individual threshold At that time, the three types of system datasets corresponding to the current target user will be... , , Extract and output; when the individual threshold is reached. Greater than or equal to the preset individual threshold But less than or equal to the preset individual threshold At that time, the three types of system datasets corresponding to the current target user will be... , , Extract and output; Otherwise, use the three types of system datasets corresponding to the current target user. , , Extract and output.

6. The cognitive level multivariate adaptive evaluation method based on heart rate variability as described in claim 1, characterized in that, The eigenvector is calculated as follows: in, , , The feature vectors of the three types of HRV sensitivity indicators , , These are the correction coefficients for the three types of HRV sensitivity indicators. , , These are the normalized values ​​of the three types of HRV sensitivity indicators. As a composite index, , , It is a constant.

7. The cognitive level multivariate adaptive evaluation method based on heart rate variability as described in claim 1, characterized in that, The classification models include support vector machines, random forests, decision trees, and k-nearest neighbors. The performance metrics of various classification models are updated in real time based on the current user's historical training data. By comparing the performance metrics of different classification models, the optimal model is selected.

8. A multivariate adaptive assessment system for cognitive levels based on heart rate variability, characterized in that, include: The data acquisition module is configured to: acquire the target user's personal information, and collect environmental parameters and the target user's electrocardiogram signal; Individual indices are calculated based on personal information, equivalent environmental indices are calculated based on environmental parameters, and HRV indices are calculated based on electrocardiogram signals. The indicator screening module is configured to: based on individual indices and relevant thresholds, adaptively optimize HRV indicators to screen three types of HRV sensitive indicators, and then obtain the three types of normalized HRV sensitive indicators. The comprehensive index calculation module is configured to: calculate the correction parameters of the three types of HRV sensitivity indicators after normalization, and dynamically adjust the correction parameters based on the target user's working time to obtain the corrected parameters; and calculate the comprehensive index based on the equivalent environmental index. The feature vector calculation module is configured to calculate feature vectors based on the normalized three types of HRV sensitivity indicators and the corrected correction parameters. The cognitive level identification module is configured to: select the optimal classification model based on the target user's historical data, input the feature vector into the optimal classification model, and obtain the current cognitive evaluation result.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the multivariate adaptive evaluation method for cognitive levels based on heart rate variability as described in any one of claims 1-7.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the cognitive level multivariate adaptive evaluation method based on heart rate variability as described in any one of claims 1-7.

Images