Dynamic evaluation method for rehabilitation training effect driven by operation behavior characteristics

Abstract

The invention belongs to the technical field of artificial intelligence, and discloses an operation behavior characteristic-driven rehabilitation training effect dynamic evaluation method, which comprises the following steps of: extracting compensatory behavior characteristics and recessive behavior mode characteristics by acquiring multi-dimensional operation behavior data and physiological sensor data of a patient on a simulated driving device; and constructing a behavior pattern migration feature set. And calculating a behavior pattern migration index based on the compensatory behavior characteristics in the characteristic set, and accurately identifying the progressive transformation process of the patient from the compensatory behavior to the normal behavior. Further constructing a rehabilitation capability dynamic evaluation model, generating a real-time rehabilitation capability score, dynamically adjusting the complexity of a simulated driving scene, collecting adaptive behavior data, constructing a behavior mode migration trend curve and a rehabilitation effect prediction model, and realizing closed-loop optimization of a personalized training scheme. According to the method, subjectivity and static limitation of traditional rehabilitation evaluation are broken through, and the accuracy and effectiveness of neural rehabilitation are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and more specifically, to a method for dynamically evaluating rehabilitation training effects driven by work behavior characteristics. Background Art

[0002] With the advancement of neurorehabilitation medicine, simulated driving environments are increasingly being used as a rehabilitation training method in clinical practice. For patients with systemic diseases, regaining driving ability is not only crucial for their independence but also crucial for social participation and psychological well-being. Driving, as a complex task involving the integration of multiple neurological functions, including visual perception, attention allocation, cognitive judgment, and motor coordination, is an ideal vehicle for assessing the extent of neurological recovery.

[0003] Currently, a variety of driving simulators are used in clinical rehabilitation training, ranging from simple desktop controllers to highly immersive virtual cockpits. These devices provide patients with realistic driving scenarios, allowing them to practice driving skills in a safe environment. Traditional assessment methods rely primarily on on-site observation by therapists combined with simple driving performance indicators (such as task completion time and number of errors). Some advanced centers also incorporate standardized neuropsychological assessment scales for additional assessment.

[0004] However, existing rehabilitation training assessment systems have difficulty effectively identifying the patient's gradual transition from compensatory behavior to normal behavior. In clinical practice, patients with neurological injuries often develop a series of compensatory behavioral strategies to complete daily tasks, such as hemiplegic patients using the non-injured limbs to overcompensate, brain-injured patients using non-standard movement sequences to complete operational tasks, or patients with cervical spondylosis using trunk rotation instead of neck rotation while driving. Although these compensatory behaviors can help patients regain functional independence in the short term, their long-term existence will hinder neural remodeling and the reconstruction of normal movement patterns. Traditional assessment methods mainly rely on intermittent observations and subjective rating scales by doctors, which cannot capture the subtle changes and gradual disappearance of patients' compensatory behaviors during the rehabilitation process. More importantly, existing technologies lack the ability to analyze the correlation between physiological sensor data and operational behavior, making it difficult to reveal the changes in cognitive load and neural control strategy hidden under the surface behavior. For example, during simulated driving rehabilitation training, a patient may appear to complete a task, but the system cannot detect trends in implicit indicators such as heart rate variability, galvanic skin response, and electromyographic activity during the task. This makes it impossible to accurately determine whether the patient continues to rely on compensatory strategies or has begun to restore normal neural control mechanisms. The lack of such a sophisticated behavioral pattern migration assessment mechanism makes it difficult for clinicians to determine the optimal intervention time and adjust rehabilitation strategies, often leading to the solidification of compensatory behaviors or inefficient training, seriously affecting rehabilitation outcomes and patient quality of life.

[0005] In view of this, the present invention proposes a dynamic evaluation method for rehabilitation training effects driven by work behavior characteristics to solve the above problems. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics, comprising: Acquire multi-dimensional operational behavior data of patients on the driving simulator and synchronously collected physiological sensor data; Extracting compensatory behavior features based on the temporal distribution and force changes of the operation actions in the multi-dimensional operation behavior data; extracting latent behavior pattern features based on the fluctuation characteristics of the physiological indicators in the physiological sensor data; and constructing a behavior pattern migration feature set based on the compensatory behavior features and the latent behavior pattern features; Calculating a behavior pattern migration index based on the attenuation trend of compensatory behavior characteristics and the strengthening trend of latent behavior pattern characteristics in the behavior pattern migration feature set; and identifying the patient's gradual transition process from compensatory behavior to normal behavior based on the behavior pattern migration index; Constructing a dynamic assessment model for patient rehabilitation ability based on the behavioral pattern migration index and a preset healthy population behavioral pattern benchmark; generating a real-time rehabilitation ability score based on the dynamic assessment model for patient rehabilitation ability; designing an adaptive adjustment strategy for a virtual driving scenario based on the real-time rehabilitation ability score, dynamically adjusting complexity parameters of the simulated driving scenario, and simultaneously collecting adaptive behavior data of the patient in the adjusted simulated driving scenario; constructing a behavior pattern migration trend curve based on the adaptive behavior data and the behavior pattern migration index; and generating a rehabilitation effect prediction model based on the behavior pattern migration trend curve; According to the rehabilitation effect prediction model and the behavior pattern migration trend curve, a personalized rehabilitation training report is generated, and the training program parameters for the next stage are optimized to achieve closed-loop dynamic optimization of rehabilitation training.

[0007] Preferably, the compensatory behavior characteristics include: Segmenting the temporal distribution of the operation actions in the multi-dimensional operation behavior data to obtain the frequency, intensity, and duration of the operation actions in each time period; Calculate the smoothness index of the operation action based on the frequency, strength and duration of the operation action in each time period; the smoothness index is used to characterize the coordination of the operation action; Extracting abnormal fluctuation characteristics of the operation action according to the deviation between the smoothness index and a preset threshold value of the smoothness of the operation action of healthy people; Compensatory behavior characteristics are determined based on the distribution density and duration of the abnormal fluctuation characteristics; the compensatory behavior characteristics include the unnecessary number of repetitions of the operating action and the force overload ratio.

[0008] Preferably, the extracting of implicit behavior pattern features includes: Performing time-frequency analysis on the fluctuation characteristics of the physiological indicators in the physiological sensor data to obtain the frequency domain energy distribution and time domain fluctuation amplitude of the physiological indicators in each time period; Calculating a dynamic balance index of a physiological indicator based on the concentration of the frequency domain energy distribution and the stability of the time domain fluctuation amplitude; the dynamic balance index is used to characterize the patient's potential cognitive load during the operation behavior; According to the difference between the dynamic balance index and a preset dynamic balance benchmark of healthy people, latent behavioral pattern characteristics are extracted; the latent behavioral pattern characteristics include cognitive load overload characteristics and physiological stress response characteristics.

[0009] Preferably, the calculation of the behavior pattern migration index includes: quantifying the attenuation trend of the compensatory behavioral characteristic to obtain the attenuation rate of the compensatory behavioral characteristic; the attenuation rate is obtained by calculating the slope of change of the compensatory behavioral characteristic in continuous time periods; quantifying the enhancement trend of the latent behavior pattern feature to obtain the enhancement rate of the latent behavior pattern feature; the enhancement rate is obtained by calculating the change slope of the latent behavior pattern feature in a continuous time period; The behavior pattern migration index is calculated based on the decay rate of the compensatory behavior characteristics and the enhancement rate of the latent behavior pattern characteristics; the behavior pattern migration index is a weighted ratio of the decay rate to the enhancement rate, where the weight is determined by the influencing factors of the compensatory behavior characteristics and the latent behavior pattern characteristics on the rehabilitation ability.

[0010] Preferably, the construction of a dynamic assessment model for patient rehabilitation ability includes: Constructing a behavior pattern migration state space according to the behavior pattern migration index; the behavior pattern migration state space includes a compensatory behavior dominant state, a transition state, and a normal behavior dominant state; Calculating a dynamic score of rehabilitation ability based on the distance between the patient's current state in the behavioral pattern migration state space and the healthy population behavioral pattern benchmark; the distance is calculated using Euclidean distance; According to the time series changes of the dynamic rehabilitation ability score, a patient rehabilitation ability dynamic evaluation model is constructed; the patient rehabilitation ability dynamic evaluation model uses a long short-term memory network to model the time series changes.

[0011] Preferably, the adaptive adjustment strategy for designing a virtual driving scene includes: determining, based on the real-time rehabilitation ability score, a direction for adjusting complexity parameters of the simulated driving scenario; the complexity parameters including road curvature, obstacle density, and traffic flow; determining an adjustment range of a complexity parameter of a simulated driving scenario according to the behavior pattern migration index; wherein the adjustment range is positively correlated with an absolute value of the behavior pattern migration index; An adaptive adjustment strategy for a virtual driving scene is generated according to the adjustment direction and the adjustment amplitude; the adaptive adjustment strategy includes prioritizing reduction of road curvature and obstacle density when the behavior pattern migration index is lower than a preset migration threshold.

[0012] Preferably, constructing the behavior pattern migration trend curve includes: Performing cluster analysis on the adaptive behavior data to obtain the patient's behavior pattern migration subsets under different complexity parameters; the behavior pattern migration subsets include an attenuation subset of compensatory behavior characteristics and an enhancement subset of implicit behavior pattern characteristics; Fitting a behavior pattern migration trend curve according to the time series changes of the behavior pattern migration subset; the behavior pattern migration trend curve is fitted using a polynomial regression model; The inflection point of the behavior pattern migration is determined according to the curvature change of the behavior pattern migration trend curve; the inflection point is used to represent the significant transition stage of the patient from compensatory behavior to normal behavior.

[0013] Preferably, generating a rehabilitation effect prediction model includes: Extracting long-term trend features and short-term fluctuation features of the trend curve according to the behavior pattern migration trend curve; the long-term trend features are obtained by low-pass filtering, and the short-term fluctuation features are obtained by high-pass filtering; Constructing a rehabilitation effect prediction model based on the long-term trend characteristics and the short-term fluctuation characteristics; the rehabilitation effect prediction model uses an autoregressive integrated moving average model to jointly model the long-term trend characteristics and the short-term fluctuation characteristics; Based on the output of the rehabilitation effect prediction model, the patient's behavior pattern migration index and rehabilitation ability score in the future training cycle are predicted.

[0014] Preferably, the optimizing the training program parameters for the next stage includes: Determining the optimization direction of the training program parameters based on the deficiencies of the behavioral pattern transfer index in the personalized rehabilitation training report; the deficiencies include insufficient attenuation of compensatory behavioral characteristics and insufficient enhancement of implicit behavioral pattern characteristics; Determining the optimization range of the training program parameters based on the prediction results of the rehabilitation effect prediction model; the optimization range is positively correlated with the improvement space of the predicted behavior pattern migration index; According to the optimization direction and the optimization amplitude, the training program parameters for the next stage are adjusted; the training program parameters include complexity parameters of the simulated driving scene, training duration, and training frequency.

[0015] Preferably, the determining of compensatory behavioral characteristics comprises: Performing kernel density estimation on the distribution density of the abnormal fluctuation feature to obtain a probability density function of the abnormal fluctuation feature; Determining the pattern type of the compensatory behavior characteristics according to the number of peaks and the peak spacing of the probability density function; the pattern type includes single-mode compensation and multi-mode compensation; A significance score of the compensatory behavior feature is calculated based on the pattern type and the duration of the abnormal fluctuation feature; the significance score is used to characterize the contribution of the compensatory behavior feature to the rehabilitation ability assessment.

[0016] The technical effects and advantages of the dynamic evaluation method of rehabilitation training effect driven by work behavior characteristics of the present invention are as follows: By accurately identifying the patient's gradual transition from compensatory strategies to normal behavioral patterns, the present invention enables clinicians to seize the golden opportunity for intervention and avoid the solidification of compensatory behaviors and the secondary injuries they bring, such as musculoskeletal imbalances and abnormal postural control. This assessment method breaks through the limitations of traditional subjective assessments, realizes the visualization and quantification of the rehabilitation process, and significantly enhances patients' confidence and compliance with rehabilitation treatment. In addition, the present invention dynamically adjusts the training difficulty to keep the patient in the "zone of proximal development", which not only avoids rehabilitation delays caused by insufficient training intensity, but also prevents frustration and secondary injuries caused by excessive training, thereby optimizing rehabilitation efficiency and patient experience. Most importantly, the present invention helps patients restore functional patterns that are closer to normal rather than relying solely on compensatory strategies by accurately guiding the neural remodeling process, significantly improving patients' long-term quality of life and ability to participate in society. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a schematic flow chart of the steps of the method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics of the present invention. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] This application provides a method for dynamically evaluating rehabilitation training effectiveness driven by behavioral characteristics. The method's execution entities include, but are not limited to, a driving simulator, a rehabilitation training system, physiological data acquisition equipment, and an evaluation server, which can be considered general computing nodes in this application. The rehabilitation training system includes, but is not limited to, at least one of virtual reality training equipment, an interactive rehabilitation device, and an intelligent feedback system.

[0020] The present invention provides a method for dynamic evaluation of rehabilitation training effects driven by operational behavior characteristics. By collecting the patient's operational behavior data and physiological sensor data on a simulated driving device, compensatory behavior characteristics and latent behavior pattern characteristics are extracted, a behavior pattern migration feature set is constructed, a behavior pattern migration index is calculated, and a dynamic evaluation model for rehabilitation ability is constructed. Based on this, the rehabilitation training plan is dynamically adjusted to achieve closed-loop dynamic optimization of rehabilitation training.

[0021] The present invention achieves consistency in feature extraction and analysis through standardized data acquisition interfaces and processing procedures, and adopts a feature-driven assessment method to improve the objectivity and accuracy of the assessment; a built-in intelligent tracking mechanism is used to monitor the patient's gradual transition from compensatory behavior to normal behavior, supporting accurate assessment of the rehabilitation process; a multi-layer assessment architecture (compensatory behavior assessment and implicit behavior pattern assessment) is used to achieve separation and coordination of assessment dimensions, meeting the strict requirements of the rehabilitation training process for objectivity, adaptability and personalization.

[0022] See also Figure 1 In an embodiment of the present invention, the detailed implementation steps of the method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics include: Acquire multidimensional behavioral data from the patient's simulated driving device and simultaneously collected physiological sensor data. Multidimensional behavioral data includes driving parameters such as steering wheel angle, steering force, pedal pressure distribution, and operation timing, reflecting the patient's overt behavioral performance during the simulated driving task. Physiological sensor data, including heart rate variability, galvanic skin response, electromyographic activity, and eye tracking, reflects changes in the patient's intrinsic physiological state during the task. This data is collected through standardized interfaces and synchronized with timestamps to ensure data consistency and integrity.

[0023] Compensatory behavior characteristics are extracted based on the temporal distribution and force variations of manipulative actions in multidimensional manipulative behavior data. Compensatory behaviors refer to nonstandard manipulative patterns adopted by patients to compensate for functional deficits, typically manifesting as uncoordinated manipulative actions, excessive force, or unnecessary repetitive movements. The extraction process identifies patients' compensatory strategies by analyzing fluctuations in manipulative timing, uneven force distribution, and repetitive patterns.

[0024] Based on the fluctuation characteristics of physiological indicators in physiological sensor data, latent behavioral pattern features are extracted. These patterns reflect the patient's cognitive load and physiological stress during task execution, which are not directly reflected in their behavioral performance. They are important supplementary indicators for assessing recovery status. The extraction process analyzes the frequency and time domain characteristics of physiological signals to identify physiological response patterns associated with cognitive effort and emotional state.

[0025] Based on the characteristics of compensatory and latent behavioral patterns, a behavioral pattern transition feature set was constructed. This feature set comprehensively reflects the characteristics of the patient's transition from compensatory behaviors to normal behaviors during the rehabilitation process, providing a data foundation for evaluating rehabilitation progress. Through standardization and feature fusion, this feature set forms a multi-dimensional evaluation indicator system.

[0026] The Behavior Pattern Migration Index (BMI) is calculated based on the attenuation trend of compensatory behavioral traits and the strengthening trend of latent behavioral patterns within the Behavior Pattern Migration Feature Set. This index quantifies the degree of behavioral pattern shift during a patient's rehabilitation process and is a core indicator for assessing rehabilitation effectiveness. The calculation process considers both the rate and direction of feature change, generating a comprehensive score through weighted fusion.

[0027] The Behavior Pattern Migration Index (BPMI) identifies patients' gradual transition from compensatory behaviors to normal behaviors. This process tracks the temporal changes in the BPMI, identifies key transition stages and recovery milestones, and provides an objective basis for clinical decision-making.

[0028] A dynamic assessment model for patients' rehabilitation capabilities is constructed based on the Behavioral Pattern Migration Index and pre-defined healthy population behavioral pattern benchmarks. This model compares patients' current behavioral patterns with healthy standards, assesses rehabilitation progress and potential, and provides quantitative rehabilitation capacity assessment results. The model uses machine learning algorithms to establish a mapping between behavioral patterns and rehabilitation capacity, enabling dynamic assessment.

[0029] A real-time rehabilitation ability score is generated based on a dynamic assessment model of the patient's rehabilitation ability. This score directly reflects the patient's current rehabilitation status, providing a basis for adjusting training plans. It also serves as a feedback indicator of the patient's rehabilitation progress, enhancing their confidence and motivation for rehabilitation.

[0030] Based on real-time rehabilitation ability scores, an adaptive adjustment strategy for the virtual driving scenario is designed. Complexity parameters of the simulated driving scenario are dynamically adjusted, while simultaneously collecting data on the patient's adaptive behavior in the adjusted simulated driving scenario. This adaptive adjustment ensures that the training difficulty remains within the patient's "zone of proximal development," ensuring it is challenging but within the patient's capabilities, maximizing training effectiveness.

[0031] Based on adaptive behavior data and the behavioral pattern migration index, a behavioral pattern migration trend curve is constructed. This curve describes the long-term trend of behavioral pattern changes during the patient's rehabilitation process, providing a data basis for predicting rehabilitation outcomes. The trend curve is constructed through data fitting and trend analysis methods to reflect the dynamic changes in the rehabilitation process.

[0032] A rehabilitation outcome prediction model is generated based on behavioral pattern migration trend curves. This model predicts future rehabilitation progress based on historical rehabilitation data, providing a scientific basis for long-term rehabilitation planning. The prediction model utilizes time series analysis, taking into account trends and fluctuations, to provide reliable rehabilitation predictions.

[0033] Based on the rehabilitation effect prediction model and behavioral pattern migration trend curve, a personalized rehabilitation training report is generated and the parameters of the next phase of training plan are optimized, achieving a closed-loop dynamic optimization of rehabilitation training. The personalized report provides detailed rehabilitation assessment results and suggestions. Training plan optimization is based on the assessment results and prediction model, adjusting training content, intensity, and frequency, forming a continuous improvement closed-loop rehabilitation loop.

[0034] In the embodiment of the present invention, the detailed implementation steps of extracting compensatory behavior characteristics include: The temporal distribution of operational actions in multidimensional operational behavior data is segmented to obtain the frequency, intensity, and duration of the actions within each time period. This segmentation process utilizes a sliding window technique, with the window length determined based on the task type and operational characteristics, typically 5-30 seconds. For the operational data within each window, the frequency of the action is counted, the mean and standard deviation of intensity are calculated, and the distribution characteristics of duration are measured. These metrics comprehensively reflect the spatiotemporal characteristics of the action.

[0035] The smoothness index of the operation action is calculated based on the frequency, strength and duration of the operation action in each time period.

[0036] Smoothness index of operating action The specific calculation formula is: ; in: , which indicates frequency stability, is the frequency standard deviation, is the frequency mean; , indicating the uniformity of force, is the standard deviation of the force, is the mean value of strength; , indicating duration consistency, is the standard deviation of duration, is the mean duration; 、 、 are the weight coefficients for frequency, intensity, and duration, respectively; The default value is , , .

[0037] To ensure that the calculation result is between 0 and 1, normalization processing is adopted. Then, the closer the smoothness index is to 1, the smoother and more coordinated the operation action is; the closer it is to 0, the more unstable the operation action is.

[0038] The smoothness index is a comprehensive metric that quantifies the coordination of manual movements, reflecting the stability and fluency of the patient's manual behavior. The smoothness index characterizes the coordination of manual movements. The calculation formula considers the stability of the manipulation frequency, the uniformity of the force, and the consistency of the movement duration, generating a normalized score between 0 and 1, with higher scores indicating smoother manipulation. In calculating the smoothness index, outliers in the manipulation data are filtered to avoid evaluation bias caused by temporary interference.

[0039] Abnormal fluctuations in operating behavior are identified based on the deviation of the smoothness index from a preset threshold for smoothness in healthy individuals. The smoothness threshold for healthy individuals is statistically derived from a large amount of normal driving data and serves as an evaluation benchmark. By calculating the difference between the patient's smoothness index and the healthy threshold, abnormal fluctuations in operating behavior are identified, which are often a direct manifestation of compensatory behavior. Abnormal fluctuation characteristics are determined by the volatility of the smoothness index over time and the degree of deviation from the healthy threshold, taking into account both the amplitude and pattern of the fluctuations.

[0040] Compensatory behavior characteristics are determined based on the distribution density and duration of abnormal fluctuation characteristics. The distribution density reflects the prevalence of abnormal fluctuations in operational behavior, while the duration reflects the stability and persistence of compensatory behavior. Compensatory behavior characteristics include the number of unnecessary repetitions of the operational action and the force overload ratio. These two indicators intuitively reflect the main manifestations of compensatory behavior. The number of unnecessary repetitions counts the number of unnecessary actions required to achieve the same operational goal, while the force overload ratio quantifies the proportion of operations that exceed the required force level. Together, these two indicators constitute a quantitative description of compensatory behavior.

[0041] In the embodiment of the present invention, the detailed implementation steps of extracting implicit behavior pattern features include: Time-frequency analysis is performed on the fluctuation characteristics of physiological indicators in physiological sensor data to obtain the frequency domain energy distribution and time domain fluctuation amplitude of the physiological indicators within each time period. Time-frequency analysis uses methods such as short-time Fourier transform or wavelet transform to decompose physiological signals into different frequency components and analyze their temporal variation characteristics. The frequency domain energy distribution reflects the intensity of physiological activity at different frequencies, while the time domain fluctuation amplitude reflects the degree of change of physiological indicators over time. Together, they provide a comprehensive description of physiological status.

[0042] The dynamic balance index of physiological indicators is calculated based on the concentration of frequency domain energy distribution and the stability of time domain fluctuation amplitude.

[0043] Dynamic balance index The specific calculation formula is: ; in, is the frequency domain energy entropy, and the calculation formula is , It is Normalized energy value of each frequency band; is the time domain variation coefficient, and the calculation formula is: , is the standard deviation of the time domain fluctuations, is the mean of the time domain fluctuation; is a weight coefficient used to balance the importance of frequency domain and time domain features. Its value range is [0,1] and is usually 0.5.

[0044] Frequency domain energy entropy The smaller the value, the more concentrated the energy distribution; the time domain coefficient of variation The smaller the value, the more stable the fluctuation range. Both are subtracted from 1, so that the higher the index value, the better the dynamic balance. The value range is [0,1].

[0045] The dynamic balance index is a comprehensive indicator for assessing physiological system stability, reflecting a patient's ability to regulate their physiology while performing tasks. It is used to characterize the potential cognitive load on patients during tasks. The calculation formula comprehensively considers the entropy of frequency-domain energy distribution and the coefficient of variation of time-domain fluctuations to generate a standardized index value. A higher dynamic balance index indicates a more stable physiological system and a more moderate cognitive load; a lower index indicates significant physiological fluctuations and a higher cognitive load.

[0046] Based on the difference between the dynamic balance index and the preset dynamic balance benchmark of healthy people, the characteristics of implicit behavioral patterns are extracted. The dynamic balance benchmark of healthy people is derived based on the physiological data statistics of normal people under the same task conditions and serves as the reference standard for evaluation. By comparing the patient's dynamic balance index with the healthy benchmark, the patient's implicit behavioral pattern characteristics are identified. These characteristics reflect the patient's adaptive state at the cognitive and emotional levels. The implicit behavioral pattern characteristics include cognitive load overload characteristics and physiological stress response characteristics. The former reflects the consumption of cognitive resources when performing tasks, and the latter reflects the degree of stress response when facing task challenges. The two together constitute the core indicators of the patient's implicit behavioral pattern.

[0047] In the embodiment of the present invention, the detailed implementation steps of calculating the behavior pattern migration index include: The attenuation trend of compensatory behavioral characteristics was quantified to obtain the attenuation rate of compensatory behavioral characteristics. The attenuation rate reflects the speed at which compensatory behaviors decrease with the progress of rehabilitation and is a key indicator for evaluating rehabilitation effectiveness. The attenuation rate is obtained by calculating the slope of the change in compensatory behavioral characteristics over consecutive time periods. Linear regression is used to fit the trend of characteristic values ​​over time. The negative value of the slope is the attenuation rate. A larger attenuation rate indicates a faster reduction in compensatory behaviors and more significant rehabilitation progress. A attenuation rate close to zero or a positive value indicates that compensatory behaviors persist or are increasing, indicating poor rehabilitation results.

[0048] The reinforcement trend of latent behavioral pattern characteristics is quantified to obtain the reinforcement rate of latent behavioral pattern characteristics. The reinforcement rate reflects the speed at which normal behavioral patterns are established and is an important indicator for assessing rehabilitation quality. The reinforcement rate is obtained by calculating the slope of change of latent behavioral pattern characteristics over consecutive time periods. Linear regression is also used to fit the temporal trend of the characteristic values. The positive value of the slope is the reinforcement rate. A larger reinforcement rate indicates faster establishment of normal behavioral patterns and higher rehabilitation quality. A reinforcement rate close to zero or a negative value indicates slow establishment or regression of normal behavioral patterns and a need for improvement in rehabilitation quality.

[0049] The behavioral pattern migration index is calculated based on the decay rate of compensatory behavioral characteristics and the enhancement rate of latent behavioral pattern characteristics. The behavioral pattern migration index is a weighted ratio of the decay rate to the enhancement rate, where the weight is determined by the impact factors of compensatory behavioral characteristics and latent behavioral pattern characteristics on rehabilitation ability. The calculation formula is: Behavioral pattern migration index = (w1 × decay rate) / (w2 × enhancement rate); Here, w1 and w2 are weight coefficients, determined based on the importance of the feature to the rehabilitation assessment. In this calculation logic, the ratio of the decay rate to the enhancement rate reflects the balance between the reduction of compensatory behavior and the establishment of normal behavior. Weighted processing considers the relative importance of different features to generate more meaningful evaluation indicators.

[0050] It should be noted that the weights are determined by the factors affecting the rehabilitation ability of compensatory behavioral characteristics and implicit behavioral pattern characteristics. The specific determination method adopts a combination of data-driven and expert evaluation: w1=β1×r1+(1-β1)×e1; w2=β1×r2+(1-β1)×e2; in: r1 and r2 are the absolute values ​​of the Pearson correlation coefficients between compensatory behavior characteristics and implicit behavior pattern characteristics and rehabilitation outcomes, respectively; e1 and e2 are the importance scores of compensatory behavior characteristics and implicit behavior pattern characteristics evaluated by experts, with a value range of [0,10]; r1, r2, e1 and e2 are the above-mentioned influencing factors; β1 is the coefficient for balancing data-driven and expert evaluation, with a value range of [0,1] and usually 0.7.

[0051] In practical applications, the initial weights can be set to w1=0.6 and w2=0.4. As more rehabilitation data are accumulated, the weight coefficients are recalculated regularly to achieve adaptive optimization of the weights.

[0052] In an embodiment of the present invention, the detailed implementation steps of constructing a dynamic assessment model for patient rehabilitation ability include: Based on the behavioral pattern migration index, a behavioral pattern migration state space is constructed. This state space is a multidimensional representation of behavioral pattern changes during a patient's rehabilitation process, providing a theoretical framework for assessing rehabilitation ability. The behavioral pattern migration state space includes a compensatory behavior-dominant state, a transitional state, and a normal behavior-dominant state, reflecting the major stages of the rehabilitation process. The compensatory behavior-dominant state indicates that the patient primarily relies on compensatory strategies to complete tasks, the transitional state indicates a coexistence of compensatory and normal behaviors, and the normal behavior-dominant state indicates that the patient primarily uses normal behavior patterns to perform tasks.

[0053] The dynamic rehabilitation ability score is calculated based on the distance between the patient's current state in the behavioral pattern migration state space and the baseline behavioral pattern of healthy people. This distance is calculated using Euclidean distance, reflecting the degree of difference between the patient's current behavioral pattern and the healthy standard. A smaller distance indicates a closer approach to the healthy standard and a higher degree of recovery; a larger distance indicates a significant difference and significant room for recovery. The dynamic rehabilitation ability score is nonlinearly mapped based on the distance value, generating a standardized score from 0 to 100 that intuitively reflects the recovery status.

[0054] A dynamic assessment model for patients' rehabilitation ability is constructed based on the time series changes in the dynamic rehabilitation ability scores. This model uses a long-short-term memory (LSTM) network to model time series changes, capturing both long-term trends and short-term fluctuations in the rehabilitation process. LSTM networks are a special type of recurrent neural network with long-term memory, making them suitable for processing time series data used in rehabilitation assessments. The model is trained using historical rehabilitation data, optimizing network parameters through supervised learning methods to establish a mapping between the score series and rehabilitation ability. The model output includes a current rehabilitation ability assessment and short-term predictions, providing real-time guidance for rehabilitation training.

[0055] In an embodiment of the present invention, the detailed implementation steps of designing an adaptive adjustment strategy for a virtual driving scene include: Based on the real-time rehabilitation ability score, the system determines the direction of adjustment for the complexity parameters of the simulated driving scenario. Complexity parameters are key factors in controlling the difficulty of the driving task, and appropriate difficulty adjustment is an important means of optimizing training effectiveness. Complexity parameters include road curvature, obstacle density, and traffic flow, which directly affect the cognitive and operational difficulty of the driving task. Based on the rehabilitation ability score, the system determines whether to increase or decrease the task difficulty. When the score is low, the complexity is reduced to avoid frustration; when the score is high, the complexity is appropriately increased to create a moderate challenge.

[0056] The Behavior Pattern Migration Index (BMI) determines the magnitude of adjustments to the complexity parameters of the simulated driving scenario. This magnitude is positively correlated with the absolute value of the BMI, reflecting the intensity and speed of the adjustments. A high BMI indicates a rapid change in the patient's behavior pattern, and the system should make larger adjustments to match the rate of change. A low BMI indicates a slow change in behavior, and the system should adopt smaller, more gradual adjustments to avoid excessive intervention.

[0057] Based on the adjustment direction and amplitude, an adaptive adjustment strategy for the virtual driving scenario is generated. This adaptive adjustment strategy includes prioritizing the reduction of road curvature and obstacle density when the behavioral pattern migration index falls below a preset migration threshold. This strategy takes into account the impact of different complexity parameters on the driving task. Road curvature and obstacle density require high operational precision. Prioritizing the reduction of these parameters to reduce the operational burden when the patient's behavioral pattern is not yet stable. Traffic flow primarily affects cognitive load and can be gradually adjusted after basic operational capabilities are established. Adaptive adjustment uses a gradual approach to avoid maladaptation caused by drastic changes. The adjustment effects are recorded at the same time to form a feedback optimization mechanism.

[0058] In the embodiment of the present invention, the detailed implementation steps of constructing the behavior pattern migration trend curve include: Cluster analysis of adaptive behavior data was performed to identify subsets of behavioral pattern shifts under different complexity parameters. Cluster analysis, using algorithms such as K-means or hierarchical clustering, grouped behavioral data based on similarity and identified behavioral pattern characteristics under different task conditions. Behavioral pattern shift subsets included a subset of attenuated compensatory behavioral characteristics and a subset of enhanced latent behavioral pattern characteristics. These two subsets reflect the process of reducing compensatory behavior and establishing normal behavior, respectively, providing detailed data for trend analysis.

[0059] Based on the time series changes in the behavioral pattern migration subset, a behavioral pattern migration trend curve is fitted. This behavioral pattern migration trend curve is fitted using a polynomial regression model to capture the nonlinear changes in the rehabilitation process. The order of the polynomial regression model is dynamically determined based on data complexity, typically choosing a 3-5th order polynomial to balance fitting accuracy and model complexity. The fitting process uses the least squares method to optimize the polynomial coefficients, while also introducing a regularization term to prevent overfitting and ensure the smoothness and predictive power of the trend curve.

[0060] The inflection point of behavioral pattern migration is determined based on the curvature change of the behavioral pattern migration trend curve. Inflection points are used to indicate a significant transition from compensatory behavior to normal behavior and are key milestones in the rehabilitation process. The mathematical definition of an inflection point is the point where the second-order derivative of the trend curve changes sign, indicating the location where the curve changes from concave to convex or vice versa. In rehabilitation assessment, this corresponds to the turning point in the rate of change of the behavioral pattern. Inflection point analysis uses numerical differentiation and extreme value detection methods to identify key transition locations on the trend curve, providing an objective basis for rehabilitation stage classification and progress assessment.

[0061] In the embodiment of the present invention, the detailed implementation steps of generating a rehabilitation effect prediction model include: Based on the behavioral pattern migration trend curve, we extract the long-term trend characteristics and short-term fluctuation characteristics of the trend curve. Long-term trend characteristics are obtained through low-pass filtering, reflecting the basic direction and speed of the rehabilitation process; short-term fluctuation characteristics are obtained through high-pass filtering, reflecting the fluctuations and adaptive characteristics of the rehabilitation process. The filtering process uses methods such as Butterworth filtering or wavelet decomposition to decompose the trend curve into different frequency components, effectively separating long-term trends from short-term fluctuations. Long-term trend characteristics are used to predict the overall direction of rehabilitation, while short-term fluctuation characteristics are used to assess rehabilitation stability and the ability to cope with changes.

[0062] A rehabilitation outcome prediction model was constructed based on long-term trend characteristics and short-term fluctuation characteristics. This model uses an autoregressive integrated moving average model to jointly model long-term trend characteristics and short-term fluctuation characteristics to predict future rehabilitation progress. The autoregressive integrated moving average model (ARIMA) is a classic time series analysis method suitable for processing data with trends and seasonality. It can effectively capture the regularity and volatility of progress in rehabilitation prediction. The model parameters (p, d, q) are determined through analysis of autocorrelation functions and partial autocorrelation functions. Maximum likelihood estimation is used to optimize the parameter values ​​and establish a mapping between data and predicted values.

[0063] Based on the output of the rehabilitation outcome prediction model, the patient's behavioral pattern transfer index and rehabilitation ability score are predicted within the next training cycle. This prediction utilizes a rolling forecast strategy, first predicting near-term values ​​and then using the predicted results as new inputs to forecast further into the future, gradually extending the prediction timeframe. The prediction results include point predictions and prediction intervals. Point predictions provide the most likely recovery trajectory, while prediction intervals account for prediction uncertainty and provide a range of possible variations. These predictions are used to develop long-term rehabilitation plans and set appropriate rehabilitation goals, guiding clinical decision-making and managing patient expectations.

[0064] In the embodiment of the present invention, the detailed implementation steps of optimizing the training program parameters for the next stage include: Based on the deficiencies in the Behavioral Pattern Transfer Index in the personalized rehabilitation training report, we determine the optimization direction for the training program parameters. These deficiencies include insufficient attenuation of compensatory behavioral characteristics and insufficient enhancement of latent behavioral pattern characteristics, reflecting deficiencies in eliminating compensatory behaviors and establishing normal behaviors, respectively. By analyzing the specific manifestations of deficiencies, such as which compensatory movements persist and which normal behavioral patterns have not been established, we systematically determine the optimization direction for the training program and strengthen the training of specific abilities in a targeted manner.

[0065] The degree of optimization of training program parameters is determined based on the predicted results of the rehabilitation outcome prediction model. The degree of optimization is positively correlated with the predicted potential for improvement in the behavioral pattern transfer index, reflecting the strength and speed of adjustments. The predicted results provide an estimate of future rehabilitation progress under the current training program. By comparing the predicted value with the rehabilitation goal, the required optimization intensity is determined. When the predicted value is far below the target, significant program adjustments are required; when the predicted value is close to the target, smaller adjustments can be made to consolidate existing gains.

[0066] Adjust the training program parameters for the next stage based on the optimization direction and magnitude. The training program parameters include the complexity parameters of the simulated driving scenario, the training duration, and the training frequency. These parameters together determine the intensity and content of the training. Complexity parameter adjustments are targeted at specific ability requirements, such as increasing road curvature to train fine steering control and increasing traffic flow to train multi-tasking capabilities. Training duration and frequency adjustments take into account the needs of fatigue management and learning consolidation, and are personalized according to the patient's tolerance and learning characteristics. The adjustment process adopts a gradual approach to avoid maladaptation caused by drastic changes. At the same time, a tracking and evaluation mechanism for the adjustment effect is established to form a closed-loop management of continuous optimization.

[0067] In the embodiment of the present invention, the detailed implementation steps of determining the compensatory behavior characteristics include: Kernel density estimation is performed on the distribution density of abnormal fluctuation characteristics to obtain the probability density function of the abnormal fluctuation characteristics. Kernel density estimation is a nonparametric estimation method that can effectively describe the distribution characteristics of data without being restricted by specific distribution assumptions. The estimation process uses a Gaussian kernel function to smooth the abnormal fluctuation data, generating a continuous probability density function that intuitively reflects the distribution pattern of abnormal fluctuations. The peak of the probability density function corresponds to the high-frequency abnormal fluctuation pattern and is key information for identifying the main compensatory behavior.

[0068] The pattern type of compensatory behavior characteristics is determined based on the number of peaks and peak spacing of the probability density function. Pattern types include single-mode compensation and multi-mode compensation, which reflect the complexity and stability of the patient's compensatory strategy. Single-mode compensation is characterized by a single main peak in the probability density function, indicating that the patient adopts a relatively consistent compensatory strategy; multi-mode compensation is characterized by multiple distinct peaks, indicating that the patient switches between different compensatory strategies based on task demands, which generally represents a more complex functional disorder. The peak spacing reflects the differentiation of different compensatory patterns. The larger the spacing, the more obvious the pattern difference, which facilitates accurate identification and targeted training.

[0069] The significance score of the compensatory behavioral characteristics was calculated based on the pattern type and the duration of the abnormal fluctuation characteristics.

[0070] The specific calculation formula for the significance score SU of compensatory behavior characteristics is: ; in: is the pattern complexity coefficient, which is calculated using different methods depending on the pattern type: Single mode compensation: ; Multimodal compensation: ; in, is the peak height, is the maximum possible peak height, N is the number of peaks, is the minimum distance between peaks, is the maximum possible distance; is the duration weight, and the calculation formula is: ; Where T is the duration of abnormal fluctuation and Ttotal is the total observation time.

[0071] Finally, the significance scores were standardized to a range of 0–10, with higher significance scores indicating that compensatory behavioral characteristics had a greater impact on the assessment of rehabilitation ability and required a higher priority in the training program.

[0072] The significance score characterizes the contribution of compensatory behavior traits to the assessment of rehabilitation capacity and is an important factor in determining trait weights. The calculation takes into account both the stability and persistence of the pattern. Single, long-lasting compensatory behaviors often indicate entrenched substitution strategies and warrant attention. Multimodal, short-lived compensatory behaviors may be temporary adaptations and warrant a lower priority. The significance score, using a standardized scale of 0-10, intuitively reflects the importance of the trait and guides subsequent assessment and training program design.

[0073] Through the description of the above detailed implementation methods, the present invention constructs a behavioral pattern migration feature set and a rehabilitation ability evaluation model by comprehensively analyzing the patient's explicit operational behavior and implicit physiological reactions, thereby achieving objective evaluation of rehabilitation training effects and dynamic optimization of training programs.

[0074] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art will be able to modify the technical solutions described in the foregoing embodiments or to substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

[0075] It should be noted that the preset parameters and threshold selection are set by those skilled in the art according to actual conditions.

[0076] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

Claims

1. A dynamic evaluation method for rehabilitation training effects driven by work behavior characteristics, characterized by: include: Acquire multi-dimensional operational behavior data of patients on the driving simulator and synchronously collected physiological sensor data; Providing compensatory behavior characteristics based on the temporal distribution and force changes of the operation actions in the multi-dimensional operation behavior data; Extracting latent behavior pattern features based on the fluctuation characteristics of physiological indicators in the physiological sensor data; constructing a behavior pattern migration feature set based on the compensatory behavior features and the latent behavior pattern features; Calculating a behavior pattern migration index according to the attenuation trend of the compensatory behavior characteristics and the enhancement trend of the latent behavior pattern characteristics in the behavior pattern migration characteristics set; Constructing a dynamic assessment model for patient rehabilitation ability based on the behavioral pattern migration index and a preset healthy population behavioral pattern benchmark; generating a real-time rehabilitation ability score based on the dynamic assessment model for patient rehabilitation ability; designing an adaptive adjustment strategy for a virtual driving scenario based on the real-time rehabilitation ability score, dynamically adjusting complexity parameters of the simulated driving scenario, and simultaneously collecting adaptive behavior data of the patient in the adjusted simulated driving scenario; constructing a behavior pattern migration trend curve based on the adaptive behavior data and the behavior pattern migration index; and generating a rehabilitation effect prediction model based on the behavior pattern migration trend curve; According to the rehabilitation effect prediction model and the behavior pattern migration trend curve, a personalized rehabilitation training report is generated, and the training program parameters for the next stage are optimized.

2. The method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics according to claim 1, characterized in that: The characteristics of compensatory behavior include: Segmenting the temporal distribution of the operation actions in the multi-dimensional operation behavior data to obtain the frequency, intensity, and duration of the operation actions in each time period; Calculate the smoothness index of the operation action based on the frequency, strength and duration of the operation action in each time period; Extracting abnormal fluctuation characteristics of the operation action according to the deviation between the smoothness index and a preset threshold value of the smoothness of the operation action of healthy people; Compensatory behavior characteristics are determined based on the distribution density and duration of the abnormal fluctuation characteristics; the compensatory behavior characteristics include the unnecessary number of repetitions of the operating action and the force overload ratio.

3. The method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics according to claim 1, characterized in that: The extracting of implicit behavior pattern features includes: Performing time-frequency analysis on the fluctuation characteristics of the physiological indicators in the physiological sensor data to obtain the frequency domain energy distribution and time domain fluctuation amplitude of the physiological indicators in each time period; Calculating a dynamic balance index of a physiological indicator according to the concentration of the frequency domain energy distribution and the stability of the time domain fluctuation amplitude; According to the difference between the dynamic balance index and a preset dynamic balance benchmark of healthy people, latent behavioral pattern characteristics are extracted; the latent behavioral pattern characteristics include cognitive load overload characteristics and physiological stress response characteristics.

4. The method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics according to claim 1, characterized in that: The calculation of the behavior pattern migration index includes: quantifying the attenuation trend of the compensatory behavioral characteristics to obtain the attenuation rate of the compensatory behavioral characteristics; quantifying the enhancement trend of the latent behavior pattern feature to obtain the enhancement rate of the latent behavior pattern feature; The behavior pattern migration index is calculated based on the decay rate of the compensatory behavior characteristics and the enhancement rate of the implicit behavior pattern characteristics; the behavior pattern migration index is a weighted ratio of the decay rate to the enhancement rate.

5. The method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics according to claim 1, characterized in that: The construction of a dynamic assessment model for patient rehabilitation ability includes: Constructing a behavior pattern migration state space according to the behavior pattern migration index; the behavior pattern migration state space includes a compensatory behavior dominant state, a transition state, and a normal behavior dominant state; Calculating a dynamic score of rehabilitation ability based on the distance between the patient's current state in the behavioral pattern migration state space and the healthy population behavioral pattern benchmark; the distance is calculated using Euclidean distance; According to the time series changes of the dynamic rehabilitation ability score, a patient rehabilitation ability dynamic evaluation model is constructed; the patient rehabilitation ability dynamic evaluation model uses a long short-term memory network to model the time series changes.

6. The method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics according to claim 1, characterized in that: The adaptive adjustment strategy for designing a virtual driving scene includes: determining, based on the real-time rehabilitation ability score, a direction for adjusting complexity parameters of the simulated driving scenario; the complexity parameters including road curvature, obstacle density, and traffic flow; determining an adjustment range of a complexity parameter of a simulated driving scenario according to the behavior pattern migration index; wherein the adjustment range is positively correlated with an absolute value of the behavior pattern migration index; An adaptive adjustment strategy for a virtual driving scene is generated according to the adjustment direction and the adjustment amplitude; the adaptive adjustment strategy includes prioritizing reduction of road curvature and obstacle density when the behavior pattern migration index is lower than a preset migration threshold.

7. The method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics according to claim 1, characterized in that: The constructing of the behavior pattern migration trend curve includes: Performing cluster analysis on the adaptive behavior data to obtain the patient's behavior pattern migration subsets under different complexity parameters; the behavior pattern migration subsets include an attenuation subset of compensatory behavior characteristics and an enhancement subset of implicit behavior pattern characteristics; Fitting a behavior pattern migration trend curve according to the time series changes of the behavior pattern migration subset; the behavior pattern migration trend curve is fitted using a polynomial regression model; The inflection point of the behavior pattern migration is determined according to the curvature change of the behavior pattern migration trend curve.

8. The method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics according to claim 1, characterized in that: The generating of the rehabilitation effect prediction model comprises: According to the behavior pattern migration trend curve, extracting the long-term trend characteristics and short-term fluctuation characteristics of the trend curve; Constructing a rehabilitation effect prediction model based on the long-term trend characteristics and the short-term fluctuation characteristics; the rehabilitation effect prediction model uses an autoregressive integrated moving average model to jointly model the long-term trend characteristics and the short-term fluctuation characteristics; Based on the output of the rehabilitation effect prediction model, the patient's behavior pattern migration index and rehabilitation ability score in the future training cycle are predicted.

9. The method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics according to claim 1, characterized in that: The optimization of the training program parameters for the next stage includes: Determining the optimization direction of the training program parameters based on the deficiencies of the behavioral pattern transfer index in the personalized rehabilitation training report; the deficiencies include insufficient attenuation of compensatory behavioral characteristics and insufficient enhancement of implicit behavioral pattern characteristics; Determining the optimization range of training program parameters according to the prediction results of the rehabilitation effect prediction model; According to the optimization direction and the optimization amplitude, the training program parameters for the next stage are adjusted; the training program parameters include complexity parameters of the simulated driving scene, training duration, and training frequency.

10. The method for dynamic evaluation of rehabilitation training effects driven by work behavior characteristics according to claim 2, characterized in that: The determination of compensatory behavioral characteristics includes: Performing kernel density estimation on the distribution density of the abnormal fluctuation feature to obtain a probability density function of the abnormal fluctuation feature; Determining the pattern type of the compensatory behavior characteristics according to the number of peaks and the peak spacing of the probability density function; the pattern type includes single-mode compensation and multi-mode compensation; A significance score of the compensatory behavioral feature is calculated based on the pattern type and the duration of the abnormal fluctuation feature.

Images