CN121171572A - A method and system for assessing the motor function of both lower limbs in stroke patients - Google Patents
A method and system for assessing the motor function of both lower limbs in stroke patientsInfo
- Publication number
- CN121171572A CN121171572A CN202511207922.4A CN202511207922A CN121171572A CN 121171572 A CN121171572 A CN 121171572A CN 202511207922 A CN202511207922 A CN 202511207922A CN 121171572 A CN121171572 A CN 121171572A
- Authority
- CN
- China
- Prior art keywords
- task
- patient
- motion
- movement
- cognitive
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a method and a system for evaluating the movement of two lower limbs of a cerebral apoplexy patient, and relates to the technical field of movement function evaluation. A brain stroke patient double-lower-limb movement evaluation system comprises a task guiding module, a movement acquisition module, a movement scoring module, an interference evaluation module, a trend analysis module, a combination analysis module and a result output module. According to the invention, the scoring model based on crowd category characteristic matching is constructed, the exercise capacity scoring is carried out by adopting a machine learning algorithm or a fixed weighting combination, personalized evaluation is realized according to different patient group characteristics, the generalization capacity and the prediction accuracy of the scoring model are improved, the grouping modeling is carried out by introducing the characteristics of the age, the sex, the exercise grade and the like of the patients, and the scoring strategy is automatically switched according to the number of samples, so that the scoring mechanism has reliability and clinical consistency in a family scene with limited data volume.
Description
Technical Field
The invention relates to the technical field of movement function evaluation, in particular to a method and a system for evaluating movement of double lower limbs of a cerebral apoplexy patient.
Background
Cerebral apoplexy is one of the main causes of nerve dysfunction, and the problems of lower limb motor function weakening, gait abnormality, cognitive ability decline and the like are often accompanied in the rehabilitation process of patients. The accurate assessment of the lower limb movement ability of the patient is a key link for guiding a rehabilitation scheme, preventing the concurrent risks such as falling and the like.
Traditional rehabilitation assessment means rely mainly on scale scoring, manual observation or clinical gait analysis. These methods have obvious limitations in that evaluation is highly subjective, has low frequency, and is difficult to reflect the actual functional state of a patient in a daily living environment. With the development of wearable sensing technology, partial researches try to quantitatively analyze the motion state of a patient by using sensors such as plantar pressure, an inertia measurement unit and the like, so that the objectivity and the continuity of evaluation are improved.
However, the prior art still lacks a systematic assessment means capable of fusing multisource movement and cognitive data, dynamically identifying ability fluctuation trends and effectively assessing rehabilitation risk states in unstructured environments such as home. In particular, the current scheme is imperfect in terms of data quality control, cognitive interference modeling, time series trend analysis and risk early warning mechanisms, and lacks a personalized scoring model and a risk evolution path identification strategy capable of adapting to different crowd types.
Therefore, there is a need for a brain stroke-oriented dual-lower-limb movement assessment method suitable for a home environment, which realizes high-quality data acquisition, multi-modal feature fusion, dynamic trend modeling and personalized risk assessment so as to support more scientific and accurate rehabilitation intervention and remote medical monitoring.
Disclosure of Invention
The invention aims to provide a brain stroke patient-oriented double-lower-limb movement assessment method and system, which have high robustness data processing capability suitable for a home environment, integrate cognitive task interference and movement performance characteristics, combine ability fluctuation trend and risk evolution path modeling, and achieve individuation, dynamic and predictability assessment of a patient functional state.
A method for assessing movement of two lower limbs of a patient suffering from cerebral apoplexy, comprising:
Guiding a patient to execute evaluation tasks of lower limbs in a home environment, wherein the evaluation tasks comprise a single task and a double task, the single task only comprises a movement task, and the double task comprises a movement task and a cognition task which are performed simultaneously;
in the process of executing a movement task by a patient, acquiring raw data comprising lower limb joint angle, gait cycle, walking acceleration, myoelectric activity and plantar pressure, and performing quality control on the raw data;
extracting features of the original data, acquiring movement function features including gait symmetry, support phase duration, ankle joint movement range, stride and walking speed, matching scoring models of basic features corresponding to crowd categories, and calculating movement capacity scores;
when the assessment task is a double task, the response time and the accuracy rate of the patient in the cognitive task are collected at the same time, the influence of the execution of the cognitive task on the motor function parameters is analyzed, and a cognitive interference index is generated;
The motion ability scores with time stamps in the continuous evaluation period are formed into time series data, and the ability fluctuation rate is calculated by adopting an analysis method of non-uniform sampling adaptation;
Carrying out combined analysis on the cognitive interference index and the ability fluctuation rate, and outputting an abnormality prompt;
And outputting the athletic ability score, the cognitive interference index, the ability fluctuation rate and the abnormality prompt as evaluation results.
As a preferred technical solution of the present invention, the quality control of the raw data includes:
The method comprises the steps of dynamically filtering based on environment perception parameters, wherein the environment perception parameters comprise illumination intensity and ground flatness, correcting signal drift based on historical data, constructing a reference base line based on a historical stable sampling section and executing data recalibration operation, and comparing consistency based on multi-source cross-validation processing, wherein the comparison comprises real-time error comparison between acquisition channels, and is used for identifying and eliminating abnormal data points.
As a preferred embodiment of the present invention, the dynamic filtering and multi-source cross-validation includes:
dynamic filtering, namely when detecting that the illumination intensity change exceeds a set threshold value or the ground flatness fluctuation exceeds a set range, starting a high-pass filter or a weighted moving average filter to carry out filtering processing on the original data so as to inhibit noise interference caused by environmental change;
And performing multi-source cross-validation processing, namely matching and matching acceleration data acquired respectively from two sides of the lower limb in real time, and performing complementary validation by combining synchronous signals of an inertial measurement unit and a myoelectric sensor, and automatically judging the data as abnormal data and rejecting or correcting the abnormal data when detecting that the numerical deviation of a certain channel exceeds a set error tolerance.
As a preferable technical scheme of the invention, the scoring model is constructed based on a machine learning algorithm or a fixed weighted combination;
When the number of the samples corresponding to the crowd category does not meet a preset threshold, a scoring model is built based on fixed weighted combination related to basic features of the crowd category, when the number of the samples meets a threshold condition, a machine learning algorithm is adopted to build the scoring model, the machine learning algorithm is selected according to the sample features of the crowd category and comprises a random forest, a support vector machine or logistic regression, the machine learning algorithm is used for training the scoring model based on the motion function features and outputting corresponding motion ability scores, and the basic features comprise the motion ability grade, age, sex and disease stage of a patient.
As a preferred embodiment of the present invention, the constructing the scoring model by using the fixed weighted combination includes:
And when the basic characteristics of the target patient are matched to a certain crowd category, selecting the fixed weighting combination associated with the crowd category, and carrying out weighting calculation on the motion function characteristic parameters of the patient.
As a preferred embodiment of the present invention, the cognitive interference index includes:
in the process of executing double tasks by a patient, synchronously acquiring the movement function characteristics of the movement tasks, the response time and the accuracy of the cognitive tasks, quantifying the movement performance disturbance caused by the cognitive tasks, acquiring a cognitive interference index, wherein the cognitive interference index comprises a weighted combination of a cognitive response delay rate and a movement parameter deviation rate and is corrected by the accuracy and used for representing the influence degree of the cognitive tasks on the movement capacity of lower limbs;
The response delay rate is the percentage change of the double-task response time relative to the baseline task response time, and the motion parameter deviation rate is the comprehensive change rate of the motion function characteristic under the double-task condition compared with the motion function characteristic under the baseline task condition.
As a preferred embodiment of the present invention, the capability fluctuation rate includes:
And modeling the time sequence data by utilizing time stamp information of the motion capability score, constructing a continuous trend curve by adopting cubic spline interpolation, and taking the first derivative change rate and the local deviation amplitude of the trend curve as quantization indexes of the capability fluctuation rate.
As a preferred embodiment of the present invention, the combinatorial analysis includes:
The method comprises the steps of constructing a two-dimensional risk state vector based on a cognitive interference index and a capability fluctuation rate, analyzing a track trend of the vector in a continuous evaluation period, calculating a track direction change rate and a risk area invasion angle, calculating trend cooperativity indexes of the cognitive interference index and the capability fluctuation rate in a plurality of continuous periods, and identifying a risk cooperativity fluctuation mode, forming evolution features by the risk cooperativity fluctuation mode, the track direction change rate and the risk area invasion angle, inputting the evolution features into a risk evolution model, judging whether the patient accords with a risk evolution path, and outputting an abnormal prompt;
The trend synergy index comprises dynamic time regular similarity of the change rates of the two, a sliding window correlation coefficient or a trend direction consistency score.
As a preferred technical solution of the present invention, the risk evolution model includes:
The method comprises the steps of carrying out cluster analysis based on longitudinal evaluation data of a large-scale cerebral apoplexy rehabilitation patient, constructing a plurality of typical risk evolution paths, setting corresponding abnormal prompts for each type of paths to form a risk evolution model, in the risk evolution model, obtaining an initial evolution path through initial matching of a risk collaborative fluctuation mode, and obtaining a final risk evolution path through similarity matching of the initial evolution path through a track direction change rate and a risk area invasion angle.
A brain stroke patient dual lower limb movement assessment system comprising:
The task guiding module is used for guiding the patient to execute the evaluation task of the lower limb in the home environment;
The motion acquisition module is used for acquiring original data and controlling quality in the process of executing a motion task by a patient;
the motion scoring module is used for extracting features of the original data, acquiring motion function features and calculating motion capability scores through a scoring model;
the interference evaluation module is used for analyzing the influence of the execution of the cognitive task on the motor function parameters and generating a cognitive interference index;
The trend analysis module is used for scoring the motion ability with time stamps in the continuous evaluation period to form time series data, and calculating the ability fluctuation rate by adopting an analysis method of non-uniform sampling adaptation;
the combination analysis module is used for carrying out combination analysis on the cognitive interference index and the ability fluctuation rate and outputting an abnormality prompt;
and the result output module is used for outputting the motor ability score, the cognitive interference index, the ability fluctuation rate and the abnormality prompt as evaluation results.
The invention has the following advantages:
According to the invention, the patient is guided to execute the lower limb assessment tasks including the single task and the double task in the home environment, the assessment is completed in the living scene of the patient at high frequency and low interference, the ecological effectiveness and daily adaptability of the rehabilitation assessment are effectively improved, and compared with the periodic assessment of the traditional clinical scene dependent professional equipment and personnel, the self-help function monitoring is carried out on the patient in the natural environment, and the rehabilitation compliance and long-term tracking continuity are improved.
According to the invention, multisource raw data of lower limb joint angles, gait cycles, walking acceleration, myoelectric activities and plantar pressure are collected, quality control is performed, accuracy and reliability of evaluation data are ensured in an unstructured environment, influence of environmental interference on results is reduced, abnormal data caused by illumination changes, ground materials or wearing errors are effectively identified and removed through dynamic filtering, signal drift correction and multisource cross-validation mechanisms, and data stability and evaluation usability in a family scene are remarkably improved.
According to the invention, the scoring model based on crowd category characteristic matching is constructed, the exercise capacity scoring is carried out by adopting a machine learning algorithm or a fixed weighting combination, personalized evaluation is realized according to different patient group characteristics, the generalization capacity and the prediction accuracy of the scoring model are improved, the grouping modeling is carried out by introducing the characteristics of the age, the sex, the exercise grade and the like of the patients, and the scoring strategy is automatically switched according to the number of samples, so that the scoring mechanism has reliability and clinical consistency in a family scene with limited data volume.
According to the invention, the response time and accuracy of the cognitive task are synchronously acquired under the double-task condition, the cognitive interference index is generated by combining the movement function parameters, the interference effect of the cognitive task on the movement ability is quantized, the sensitivity of evaluation on complications such as falling risks is enhanced, the index not only reflects the cognitive movement coordination load ability, but also reveals the functional weakness of a patient when the complex task is executed, and data support is provided for early identification of falling high risk states and cognitive degradation trends.
According to the invention, the motion ability score in the continuous evaluation period is constructed into a time sequence, the ability change trend and the ability fluctuation rate are identified by adopting an analysis method of nonuniform sampling adaptation, and the fluctuation state of the patient ability in the rehabilitation process is dynamically captured, so that early degradation early warning is realized;
According to the invention, the two-dimensional risk state vector based on the cognitive interference index and the capability fluctuation rate is constructed, the characteristics of track trend, trend cooperativity and the like are analyzed, the cooperative fluctuation mode is identified, the interactive risk mechanism of cognition and movement is accurately modeled, the sensitivity and the specificity of risk identification are enhanced, the cooperative fluctuation mode, the track direction change rate and the risk area invasion angle are input into the risk evolution model, the risk evolution path is matched, abnormal prompt information is output, the state classification of the rehabilitation path and the early identification of the high risk development trend are realized, and the remote rehabilitation monitoring and the individuation intervention are supported.
Drawings
For a clearer description of an embodiment of the present invention or a technical solution in the prior art, the following description will briefly introduce the drawings required to be used in the embodiment or the description of the prior art, it is obvious that the drawings in the following description are only schematic views of the present invention, and other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art;
Fig. 1 is a schematic structural diagram of a system for evaluating movement of two lower limbs of a patient with cerebral apoplexy according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail below with reference to the accompanying drawings, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Embodiment 1, a method for evaluating the movement of two lower limbs of a patient suffering from cerebral apoplexy, comprising the following steps:
Step S1, guiding a patient to execute evaluation tasks of lower limbs in a home environment, wherein the evaluation tasks comprise a single task and a double task, the single task only comprises a movement task, and the double task comprises a movement task and a cognition task which are performed simultaneously;
In this embodiment, the home environment refers to a scenario in which a patient evaluates in a living residence of a non-medical institution, and includes typical areas such as bedrooms, living rooms, balconies, etc., where the environment usually has unstructured features such as irregular spatial layout, frequent illumination changes, various background noise, etc., so that the evaluation task needs to have good environmental adaptability and long-range guidance.
The guiding mode comprises text prompt, voice broadcast or visual demonstration (such as App animation or video guidance) and is used for helping a patient to independently complete an action sequence under the condition that no professional exists. The guidance mode is dynamically selected based on patient history,
For example, when it is detected that the accuracy of the completion of the last evaluation task by the patient is insufficient, the guidance mode with the video presentation is preferentially selected.
The evaluation tasks are divided into a single task type and a double task type according to task complexity;
The single task refers to a lower limb motion sequence only comprising a motion task, and the single task comprises standardized rehabilitation motions such as in-situ high leg lifting, slow linear walking, standing and turning, alternating forward and backward stepping and the like, and is mainly used for evaluating the basic lower limb motion capability of a patient under the condition of no cognitive interference.
The double tasks are to superimpose a mild cognitive task while executing the above-mentioned exercise task so as to simulate the multitask situations such as 'walking+thinking' or 'walking+listening to voice' which are needed by the patient in daily life. Cognitive tasks include, but are not limited to, continuous reverse numeric counting (e.g., counting down from 100), instruction recognition and reaction tasks (e.g., quickly judging and speaking a color or object name after hearing speech), simple word memory tasks (e.g., memorizing 3 words during walking), and dual task assessment to significantly increase sensitivity to "cognitive-motor" coupling capability, which is a key element in judging potential fall risk.
To ensure action standardization and result consistency, the assessment tasks should be uniformly performed after calibration. For example, in a high leg lifting task, the maximum knee lifting angle needs to be collected each time, and the action is prompted to be completed or repeated. The start-stop time of the cognitive task and the movement task are synchronously triggered, so that the real-time interference of the cognitive load on the movement behavior can be accurately captured.
Step S2, acquiring original data comprising lower limb joint angles, gait cycles, walking acceleration, myoelectric activities and plantar pressure in the process of executing a movement task by a patient, and performing quality control on the original data;
in this embodiment, the raw data are lower limb movement physiological and kinetic data synchronously collected by the multi-mode sensor during the movement task completion process of the patient, and the data form the basis of subsequent feature extraction, capability scoring and risk modeling.
The original data acquisition mode and source include:
the lower limb joint angle is obtained by an Inertial Measurement Unit (IMU) arranged at the positions of knee joint, ankle joint and the like, and comprises dual-source fusion of a gyroscope and an accelerometer. The output is the angular change in degrees (degrees) at the sagittal and coronal planes for analysis of flexion-extension amplitude and gait periodicity. For example, the maximum right knee flexion angle is about 65 degrees, and is a determination threshold value for high leg lifting operation.
Gait cycle the start-stop time period of a complete gait is calculated from the plantar pressure sensor array, comprising the duration of the support phase and the swing phase in seconds or milliseconds. For example, the left foot has a full gait cycle of 1.1 seconds and a support phase of 64%.
The three-axis acceleration signal is obtained by the wearable IMU, the sensing points of the legs, the hips and the feet are covered, the data frequency is 50-200 Hz, and the unit is m/s2. The variation trend is used for calculating the step frequency, symmetry and identifying abnormal gait rhythms. For example, left and right lateral x-axis acceleration deviations >20%, exist as gait asymmetry.
Myoelectric activity (EMG) is that myoelectric signals of main lower limb muscle groups are collected by using surface myoelectric electrodes, common muscles comprise quadriceps femoris, popliteal cord muscle and tibialis anterior, sampling frequency is 500-1000 Hz, unit is mV, and muscle contraction strength and coordination are reflected. For example, there is a decrease in knee extension force with a lower right quadriceps EMGRMS value.
Plantar pressure, namely, acquiring pressure values of unit areas of the forefoot, heel and midfoot areas by using a flexible pressure sensing pad (distributed array), wherein the unit is kPa, and the pressure values are used for detecting a support load transfer mode and gait stability. For example, the difference in maximum values of the bipedal pressure is >20kPa, with the risk of shifting the center of gravity.
In order to ensure the accuracy and stability of the acquired data in the home environment, three quality control mechanisms are executed:
The dynamic filtering (participation of environment sensing parameters) comprises the steps of configuring an environment sensing unit, collecting the current illumination intensity (such as Lux value) and the ground flatness (estimated by a millimeter wave sensor or a visual depth camera), and when the detected ambient light change exceeds a set threshold (such as DeltaLux > 200) or the ground flatness fluctuation exceeds a threshold (such as inclination angle >5 degrees), automatically starting a high-pass filter (removing low-frequency drift) or weighting a sliding average filter (suppressing high-frequency noise) so as to ensure the stability of data, for example, the acceleration signal fluctuation is amplified by the ambient change and the normal waveform is restored after filtering.
Signal drift correction (history contrast modeling) is to record a stable motion segment under a user history normal state to construct a multidimensional reference base line, and if a current signal continuously deviates from a trend (such as angle base line drift is more than 10 degrees), a dynamic recalibration operation is executed, for example, the ankle joint angle gradually deviates in flat ground walking, and the ankle joint angle is returned to the stable base line after being detected.
Multisource cross-validation (real-time consistency comparison) by real-time comparison of output data from multiple sensor channels (e.g., two-leg IMU, EMG, pressure pad), and if abnormal differences (e.g., 25% of deviation rate) between the single-side channel data and the opposite-side and auxiliary channels (e.g., EMG) are detected, abnormal rejection or signal correction is performed, e.g., right foot acceleration signal distortion is identified by myoelectricity-left leg acceleration cross-validation, and automatic rejection is performed.
Step S3, extracting features of the original data, obtaining movement function features including gait symmetry, support phase duration time, ankle joint movement range, stride and walking speed, matching scoring models of basic features corresponding to crowd types, and calculating movement capacity scores;
the scoring model is constructed based on a machine learning algorithm or a fixed weighted combination;
When the number of the samples corresponding to the crowd category does not meet a preset threshold, a scoring model is built based on fixed weighted combination related to basic features of the crowd category, when the number of the samples meets a threshold condition, a machine learning algorithm is adopted to build the scoring model, the machine learning algorithm is selected according to the sample features of the crowd category and comprises a random forest, a support vector machine or logistic regression, the machine learning algorithm is used for training the scoring model based on the motion function features and outputting corresponding motion ability scores, and the basic features comprise the motion ability grade, age, sex and disease stage of a patient.
The fixed weighted combination construction scoring model comprises:
And when the basic characteristics of the target patient are matched to a certain crowd category, selecting the fixed weighting combination associated with the crowd category, and carrying out weighting calculation on the motion function characteristic parameters of the patient.
In this embodiment, the purpose is to quantify the core lower limb motor performance of a stroke patient by performing motor function feature extraction on the multimodal raw data with the data quality control completed, and map these performances into structured motor ability scores by a personalized scoring model.
The description and extraction modes of the motion function features comprise:
Gait symmetry, which is to represent the consistency of the left and right lower limbs in terms of motion rhythm, amplitude and frequency, is calculated by using a Symmetry Index (SI), and a formula is shown as SI= |left step size-right step size|/(0.5× (left step size+right step size)) ×100% in units of%. High symmetry represents good nerve control, low symmetry may suggest hemiplegia or motor incompatibility.
The support phase duration time is obtained by detecting the contact start-stop time of the foot by a plantar pressure sensor, wherein the support phase duration time is the proportion of the foot landing bearing time to the whole gait cycle. For example, the left foot support phase time is 0.72 seconds, accounting for 66% of the gait cycle.
Ankle joint range of motion, which represents the maximum Qu Shenjiao degree difference in gait of the ankle joint, is obtained by integrating the IMU angular velocity signals in degrees. Normal values are about 20-30 deg., with restricted movement indicating stiffness or movement disorders.
Stride, representing the horizontal distance in cm between two consecutive identical foot falls, is obtained by IMU trajectory estimation or plantar pressure space solution, and stride shortening is related to cognitive deterioration or fatigue.
The walking speed is obtained by dividing the total distance by the total time, and the unit is m/s, and is the most representative single index in rehabilitation evaluation.
To improve the accuracy of scoring and the suitability of individuals, a scoring model dynamic matching mechanism is designed, namely a scoring strategy is selected according to the group of people to which the patient belongs, comprising:
the basic features of the population categories are formulated by clinical professionals, including exercise capacity levels (based on FMA scores, etc.), age intervals (e.g. <45, middle aged 45-65, elderly > 65), gender (considering differences in muscle strength), disease stages (acute, subacute, chronic, etc.). The basic features are used to construct matching logic between the patient and the historical samples.
If the historical training sample number of the crowd class of the target patient is more than or equal to a preset threshold (such as 50), a machine learning scoring model is used, and if the historical training sample number is less than the threshold, a fixed weighting combination scoring model matched with the crowd class is selected.
The scoring model constructed by the machine learning algorithm comprises 5 types of motion function features which are input and extracted, model types which are automatically selected according to training sample features, and motion capability scores (for example, 0-100 points) which are used for tracking recovery trend or evaluating intervention effect.
The method comprises the steps of fixing a weighted combination model, presetting a weight combination for each group of people, analyzing the historical feature sensitivity of the weight from the corresponding group of people, inputting patient features, executing weighted summation to output scores, and automatically selecting a weight template with the highest matching degree with the basic features of a target patient to execute the scores.
The dynamic switching mechanism ensures that an available scoring strategy is provided at an early stage of sample accumulation, learning capability is provided after the sample is sufficient, and meanwhile, the interpretive performance of cross-crowd comparison is ensured.
S4, when the assessment task is a double task, the response time and the accuracy rate of the patient in the cognitive task are collected at the same time, the influence of the execution of the cognitive task on the motor function parameters is analyzed, and a cognitive interference index is generated;
The cognitive interference index comprises:
in the process of executing double tasks by a patient, synchronously acquiring the movement function characteristics of the movement tasks, the response time and the accuracy of the cognitive tasks, quantifying the movement performance disturbance caused by the cognitive tasks, acquiring a cognitive interference index, wherein the cognitive interference index comprises a weighted combination of a cognitive response delay rate and a movement parameter deviation rate and is corrected by the accuracy and used for representing the influence degree of the cognitive tasks on the movement capacity of lower limbs;
The response delay rate is the percentage change of the double-task response time relative to the baseline task response time, and the motion parameter deviation rate is the comprehensive change rate of the motion function characteristic under the double-task condition compared with the motion function characteristic under the baseline task condition.
In order to quantify the actual interference degree of cognitive tasks on the lower limb movement ability of a cerebral apoplexy patient when the cerebral apoplexy patient executes the double-task movement, the embodiment constructs a cognitive interference index, and reflects the movement-cognitive resource allocation ability and the stability of nerve control through multi-mode signal collaborative analysis.
The dual task definition, the dual task assessment in this embodiment, requires the patient to perform a cognitive task while performing standard lower limb exercise tasks (e.g., walking, leg lifting, turning).
And synchronously acquiring cognitive task data (response time, unit ms, accuracy and unit percent), motor function data (five parameters same as step S3), control data such as time stamp, execution stage mark (start/end/answer point) and the like in each task process.
The baseline task definition requires the patient to complete a pure motion task on the premise of not loading a cognitive task, so as to construct an individualized motion and cognitive response baseline for comparison analysis.
The cognitive interference index is used for evaluating the compound influence on the movement capacity after the cognitive task is loaded, and specifically comprises the following components:
Cognitive response delay rate: Wherein For the response time of the dual task,Reflecting processing delay caused by cognitive load rise for response time of a baseline task;
Calculating the motion parameter offset rate by comparing the motion function feature change rates under the double-task condition and the single-task condition: In the present embodiment Representing 5 movement function characteristics,As a subscript to the feature of the athletic functionality,AndRespectively under the dual task and the baseline taskAnd the motion function characteristic parameters.
And (3) an accuracy correction factor, namely considering that if the accuracy of the cognitive task is obviously reduced, the task cannot be effectively executed, the weight needs to improve the interference index, and otherwise, the interference evaluation is weakened in proportion.
The final cognitive interference index is calculated by the following formula, CII=alpha×delay rate+beta×offset rate (accuracy), wherein alpha and beta are preset weighting coefficients (such as 0.5/0.5 or adjusted according to crowd characteristics);
according to historical data and expert settings, setting a personalized interference threshold (for example, 25% of the CII is judged to be significant interference);
The disturbance index is used for constructing a subsequent 'risk state vector', identifying a 'cognitive degradation' rehabilitation mode, assisting in evaluating the cognitive load bearing capacity of a patient, and guiding training rhythm and task arrangement.
The Cognitive Interference Index (CII) is the average disturbance intensity of cognitive task loading on the motion performance under the condition of double tasks, reflects the nerve integration efficiency of a cerebral apoplexy patient under the condition of resource conflict, and is one of important innovation indexes provided by the invention, so as to bridge the gap between traditional motion evaluation and multi-task challenges in actual life.
S5, scoring the motion ability with time stamps in a continuous evaluation period to form time series data, and calculating the ability fluctuation rate by adopting an analysis method of non-uniform sampling adaptation;
The capability fluctuation rate includes:
And modeling the time sequence data by utilizing time stamp information of the motion capability score, constructing a continuous trend curve by adopting cubic spline interpolation, and taking the first derivative change rate and the local deviation amplitude of the trend curve as quantization indexes of the capability fluctuation rate.
The method is used for identifying the stability and the variation amplitude of the exercise capacity of the patient by carrying out trend modeling and dynamic analysis on the continuous exercise capacity scores of the patient in a period of time, so that fluctuation characteristic input is provided for subsequent risk assessment.
After each scoring task is executed in the step S3, a scoring result of the athletic ability with a time stamp is output, wherein the recording format is as follows [ scoring numerical value, time stamp, task type (single task/double task), completion mark ];
in a home evaluation scenario, the task execution frequency of the patient is not fixed (e.g., once a day or once three days), and the time stamp interval has irregularity and sparsity, which results in the failure of the conventional equidistant time series analysis algorithm.
To accommodate such time interval variations, the present invention employs the step of ordering all scoring results in a time-stamped order, creating a non-equidistant time sequence. A cubic spline interpolation method is introduced to fit the continuous curve of the sequence in a set interval (such as 7 days and 14 days), and the method ensures that the fitted curve is continuous at each sampling point, has continuous first-order and second-order derivatives at nodes and avoids oscillation artifacts. And outputting a continuous function curve f (t), wherein t is a time axis, and the function value is a scoring result.
The ability to change rate (FVR), which represents the severity of the trend change, is constructed by calculating the variance of the derivative value (i.e., slope) change per unit time, in "score/day 2", based on the trend curve. Local deviation amplitude by setting a sliding window (such as 3 days) and comparing the local score with the residual error of the fitting value, reflecting the fluctuation degree of the patient in a short period, and identifying the early unstable sign. The maximum descent rate and duration (optional index) are recorded as the most significant slope of the descent of the score and the number of days the trend is sustained in a period of time, and the information is used for carrying out reasoning modeling by combining with the subsequent risk evolution path.
The capacity fluctuation rate is finally expressed as fvr=w 1 slope variance+w 2 local bias+w 3 maximum downlink rate (optional), where w 1、w2、w3 is the weight adjustment parameter;
The method is particularly suitable for data characteristics in a home environment, and has high sensitivity recognition capability for the clinical long-period mild fluctuation trend.
S6, carrying out combined analysis on the cognitive interference index and the ability fluctuation rate, and outputting an abnormality prompt;
The combinatorial analysis includes:
The method comprises the steps of constructing a two-dimensional risk state vector based on a cognitive interference index and a capability fluctuation rate, analyzing a track trend of the vector in a continuous evaluation period, calculating a track direction change rate and a risk area invasion angle, calculating trend cooperativity indexes of the cognitive interference index and the capability fluctuation rate in a plurality of continuous periods, and identifying a risk cooperativity fluctuation mode, forming evolution features by the risk cooperativity fluctuation mode, the track direction change rate and the risk area invasion angle, inputting the evolution features into a risk evolution model, judging whether the patient accords with a risk evolution path, and outputting an abnormal prompt;
The trend synergy index comprises dynamic time regular similarity of the change rates of the two, a sliding window correlation coefficient or a trend direction consistency score.
The risk evolution model comprises:
The method comprises the steps of carrying out cluster analysis based on longitudinal evaluation data of a large-scale cerebral apoplexy rehabilitation patient, constructing a plurality of typical risk evolution paths, setting corresponding abnormal prompts for each type of paths to form a risk evolution model, in the risk evolution model, obtaining an initial evolution path through initial matching of a risk collaborative fluctuation mode, and obtaining a final risk evolution path through similarity matching of the initial evolution path through a track direction change rate and a risk area invasion angle.
In the application, the aim of the current step is to jointly model a Cognitive Interference Index (CII) and a capacity fluctuation rate (FVR) to form a dynamic patient risk state evolution process, and establish a risk evolution model based on real rehabilitation big data for identifying the trend of the patient to evolve to a high risk state in advance.
Constructing a two-dimensional vector space using CII and FVR generated by the patient during each evaluation cycle as the abscissa: riskVecotr t=[CIIt,FVRt ];
The coordinate position of the vector represents the current state of the patient in two dimensions of cognitive influence and motion stability, the set high-risk area boundary (such as high interference and high fluctuation corresponding to the upper right quadrant) is combined to realize coarse judgment of the risk level of a single state point, and vector sequences under a plurality of evaluation periods form a track curve for further trend judgment.
Trend analysis is carried out on RiskVector sequences, and the following three evolution characteristics are extracted:
track Direction Change Rate (DCR) is used for representing the continuity of the track in the change direction in different evaluation periods, the direction fluctuation frequently reflects the instability of the recovery of a patient or the disturbance of the patient by various factors, and the vector included angle change root mean square value is used for calculation.
The risk area invasion angle (REA) is used for measuring the included angle between the track and the vertical line of the boundary when the track enters the high risk image limit for the first time, and the closer to the vertical entering, the risk area invasion angle (REA) shows quicker risk accumulation and has burst property, and is used for distinguishing two evolution trends of slow landslide and sudden drop.
The trend cooperativity index (TCL) is used for showing whether the cognitive interference index and the ability fluctuation rate show the same trend or not, the higher the trend cooperativity is, the synchronous worsening of the cognitive burden and the movement instability is represented, and the risk improvement is more clinically warned, and the TCL comprises three calculation modes including dynamic time regular similarity (DTW similarity) used for capturing nonlinear similar trends, sliding window correlation coefficients used for identifying the synchronous change degree in a short period, and direction consistency scores used for counting the synchronous ascending or descending proportion of the cognitive interference index and the movement instability.
And integrating the DCR, REA and TCI into a three-dimensional evolution feature vector, and sending the three-dimensional evolution feature vector into a risk evolution model as input.
The risk evolution model is a core intelligent analysis component constructed based on large-scale rehabilitation patient data and comprises the steps of extracting a track sequence with obvious statistical characteristics from historical data, constructing a typical evolution path by using a time sequence clustering algorithm (such as K-SpectralClustering), and binding specific abnormal prompt tags such as high falling risk, cognitive degradation leading, fluctuation unbalance, rehabilitation hysteresis and the like to each path.
And (3) performing double-stage path matching, namely judging whether to trigger model matching according to whether a risk collaborative fluctuation mode (TCI exceeds a threshold value) occurs, and performing similarity sorting on paths by using a track Direction Change Rate (DCR) and a risk area invasion angle (REA) to select the most probable evolution direction. And if the matching path is of a high risk type and the similarity exceeds a set threshold, outputting abnormal prompt information.
And in the risk cooperative fluctuation mode, the interference index and the ability fluctuation rate are identified to synchronously change (e.g. synchronously rise) in a plurality of continuous evaluation periods, so that the cognitive-motor regulation mechanism is dysregulated, and the clinical risk directivity is higher.
The evolution feature vector comprises three-dimensional indexes [ TCI, DCR, REA ] which are used for quantifying the risk state trend of the current evaluation stage and are core inputs of path identification.
And the risk evolution path is a historical experience model reflecting how the risk indexes of typical rehabilitation crowd evolve in a specific stage, and provides basis for systematically predicting future risk trend of patients.
And S7, outputting the athletic ability score, the cognitive interference index, the ability fluctuation rate and the abnormality prompt as evaluation results.
In the application, the current step aims at carrying out unified summarization and structural expression on the previous analysis results and outputting the analysis results in a standardized format for feeding back to a user terminal, a medical worker platform or a clinical evaluation system so as to realize the support of personalized rehabilitation management.
The evaluation results are output in a structured JSON or HL7 data format for easy integration with third party medical information systems (e.g., EMR).
Embodiment 2, a system for evaluating the movement of two lower limbs of a patient suffering from cerebral apoplexy, see fig. 1, comprises the following modules:
The task guiding module is used for guiding the patient to execute the evaluation task of the lower limb in the home environment;
The motion acquisition module is used for acquiring original data and controlling quality in the process of executing a motion task by a patient;
the motion scoring module is used for extracting features of the original data, acquiring motion function features and calculating motion capability scores through a scoring model;
the interference evaluation module is used for analyzing the influence of the execution of the cognitive task on the motor function parameters and generating a cognitive interference index;
The trend analysis module is used for scoring the motion ability with time stamps in the continuous evaluation period to form time series data, and calculating the ability fluctuation rate by adopting an analysis method of non-uniform sampling adaptation;
the combination analysis module is used for carrying out combination analysis on the cognitive interference index and the ability fluctuation rate and outputting an abnormality prompt;
and the result output module is used for outputting the motor ability score, the cognitive interference index, the ability fluctuation rate and the abnormality prompt as evaluation results.
Embodiment 3 provides a brain stroke patient double-lower-limb movement evaluation scheme realized by combining a sole rehabilitation device, which is based on the evaluation method provided by the invention, and based on the existing 'simple dynamic controllable sole multi-region independent load test standing posture training rehabilitation device', the comprehensive evaluation of the brain stroke patient lower-limb functional state in a home or unstructured environment is realized through the cooperative expansion of software and hardware.
In the actual deployment process, the patient stands on the plantar pedal set by the device, and the system guides the patient to sequentially complete the action sequences of single tasks (such as in-situ stepping and static load transfer) and double tasks (such as stepping and reverse counting). Gait cycle and plantar stress distribution data are obtained through a multi-region plantar pressure sensor arranged in the device, and the angle, acceleration and myoelectric activity of the lower limb joint of a patient are collected by combining an external Inertial Measurement Unit (IMU) and a surface myoelectric sensor (sEMG).
The acquired original data enters a feature extraction module after dynamic filtering and multi-channel cross verification of the system, and the motion function features such as gait symmetry, support phase duration time and the like are acquired. The system automatically matches the scoring model based on the patient's basic characteristics (e.g., age, motor ability level), outputting motor ability scores. Meanwhile, the system processes the cognitive response time and accuracy rate in the double-task period to generate a cognitive interference index.
Finally, the continuous periodic motor ability score and the cognitive interference index are used for risk combination analysis, a two-dimensional risk state vector is constructed, evolution features such as track trend and trend cooperativity are extracted, the evolution features are input into a preset risk evolution model, whether a patient is in a high risk evolution path is judged, and if necessary, fall early warning or rehabilitation strategy adjustment suggestions are output.
The embodiment shows the landing capability and expansibility of the invention on the basis of the existing rehabilitation device, and proves that the invention has practical application feasibility and precision guarantee capability in a home scene.
The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the invention, and is not meant to limit the scope of the invention, but to limit the invention to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (10)
1. A method for assessing movement of two lower limbs of a patient suffering from cerebral apoplexy, comprising the steps of:
Guiding a patient to execute evaluation tasks of lower limbs in a home environment, wherein the evaluation tasks comprise a single task and a double task, the single task only comprises a movement task, and the double task comprises a movement task and a cognition task which are performed simultaneously;
in the process of executing a movement task by a patient, acquiring raw data comprising lower limb joint angle, gait cycle, walking acceleration, myoelectric activity and plantar pressure, and performing quality control on the raw data;
extracting features of the original data, acquiring movement function features including gait symmetry, support phase duration, ankle joint movement range, stride and walking speed, matching scoring models of basic features corresponding to crowd categories, and calculating movement capacity scores;
when the assessment task is a double task, the response time and the accuracy rate of the patient in the cognitive task are collected at the same time, the influence of the execution of the cognitive task on the motor function parameters is analyzed, and a cognitive interference index is generated;
The motion ability scores with time stamps in the continuous evaluation period are formed into time series data, and the ability fluctuation rate is calculated by adopting an analysis method of non-uniform sampling adaptation;
Carrying out combined analysis on the cognitive interference index and the ability fluctuation rate, and outputting an abnormality prompt;
And outputting the athletic ability score, the cognitive interference index, the ability fluctuation rate and the abnormality prompt as evaluation results.
2. The method for evaluating the motion of two lower limbs of a patient suffering from cerebral apoplexy according to claim 1, wherein said quality control of the raw data comprises:
The method comprises the steps of dynamically filtering based on environment perception parameters, wherein the environment perception parameters comprise illumination intensity and ground flatness, correcting signal drift based on historical data, constructing a reference base line based on a historical stable sampling section and executing data recalibration operation, and comparing consistency based on multi-source cross-validation processing, wherein the comparison comprises real-time error comparison between acquisition channels, and is used for identifying and eliminating abnormal data points.
3. The method for assessing the motion of a double lower limb of a stroke patient according to claim 2, wherein the dynamic filtering and multi-source cross-validation comprises:
dynamic filtering, namely when detecting that the illumination intensity change exceeds a set threshold value or the ground flatness fluctuation exceeds a set range, starting a high-pass filter or a weighted moving average filter to carry out filtering processing on the original data so as to inhibit noise interference caused by environmental change;
And performing multi-source cross-validation processing, namely matching and matching acceleration data acquired respectively from two sides of the lower limb in real time, and performing complementary validation by combining synchronous signals of an inertial measurement unit and a myoelectric sensor, and automatically judging the data as abnormal data and rejecting or correcting the abnormal data when detecting that the numerical deviation of a certain channel exceeds a set error tolerance.
4. The method for evaluating the motion of two lower limbs of a patient suffering from cerebral apoplexy according to claim 1, wherein the scoring model is constructed based on a machine learning algorithm or a fixed weighted combination;
When the number of the samples corresponding to the crowd category does not meet a preset threshold, a scoring model is built based on fixed weighted combination related to basic features of the crowd category, when the number of the samples meets a threshold condition, a machine learning algorithm is adopted to build the scoring model, the machine learning algorithm is selected according to the sample features of the crowd category and comprises a random forest, a support vector machine or logistic regression, the machine learning algorithm is used for training the scoring model based on the motion function features and outputting corresponding motion ability scores, and the basic features comprise the motion ability grade, age, sex and disease stage of a patient.
5. The method for assessing the motion of two lower extremities of a stroke patient according to claim 4, wherein said constructing a scoring model from said fixed weighted combination comprises:
And when the basic characteristics of the target patient are matched to a certain crowd category, selecting the fixed weighting combination associated with the crowd category, and carrying out weighting calculation on the motion function characteristic parameters of the patient.
6. The method for evaluating the movement of two lower limbs in a stroke patient according to claim 1, wherein the cognitive interference index comprises:
in the process of executing double tasks by a patient, synchronously acquiring the movement function characteristics of the movement tasks, the response time and the accuracy of the cognitive tasks, quantifying the movement performance disturbance caused by the cognitive tasks, acquiring a cognitive interference index, wherein the cognitive interference index comprises a weighted combination of a cognitive response delay rate and a movement parameter deviation rate and is corrected by the accuracy and used for representing the influence degree of the cognitive tasks on the movement capacity of lower limbs;
The response delay rate is the percentage change of the double-task response time relative to the baseline task response time, and the motion parameter deviation rate is the comprehensive change rate of the motion function characteristic under the double-task condition compared with the motion function characteristic under the baseline task condition.
7. The method for evaluating the motion of two lower limbs of a patient suffering from cerebral apoplexy according to claim 1, wherein the capability fluctuation rate comprises:
And modeling the time sequence data by utilizing time stamp information of the motion capability score, constructing a continuous trend curve by adopting cubic spline interpolation, and taking the first derivative change rate and the local deviation amplitude of the trend curve as quantization indexes of the capability fluctuation rate.
8. The method for assessing the movement of two lower limbs in a stroke patient according to claim 1, wherein said combined analysis comprises:
The method comprises the steps of constructing a two-dimensional risk state vector based on a cognitive interference index and a capability fluctuation rate, analyzing a track trend of the vector in a continuous evaluation period, calculating a track direction change rate and a risk area invasion angle, calculating trend cooperativity indexes of the cognitive interference index and the capability fluctuation rate in a plurality of continuous periods, and identifying a risk cooperativity fluctuation mode, forming evolution features by the risk cooperativity fluctuation mode, the track direction change rate and the risk area invasion angle, inputting the evolution features into a risk evolution model, judging whether the patient accords with a risk evolution path, and outputting an abnormal prompt;
The trend synergy index comprises dynamic time regular similarity of the change rates of the two, a sliding window correlation coefficient or a trend direction consistency score.
9. The method for assessing the motion of two lower limbs of a stroke patient according to claim 8, wherein said risk evolution model comprises:
The method comprises the steps of carrying out cluster analysis based on longitudinal evaluation data of a large-scale cerebral apoplexy rehabilitation patient, constructing a plurality of typical risk evolution paths, setting corresponding abnormal prompts for each type of paths to form a risk evolution model, in the risk evolution model, obtaining an initial evolution path through initial matching of a risk collaborative fluctuation mode, and obtaining a final risk evolution path through similarity matching of the initial evolution path through a track direction change rate and a risk area invasion angle.
10. A brain stroke patient double lower limb movement assessment system, characterized in that the system applies any one of the brain stroke patient double lower limb movement assessment methods of claims 1 to 9, comprising:
The task guiding module is used for guiding the patient to execute the evaluation task of the lower limb in the home environment;
The motion acquisition module is used for acquiring original data and controlling quality in the process of executing a motion task by a patient;
the motion scoring module is used for extracting features of the original data, acquiring motion function features and calculating motion capability scores through a scoring model;
the interference evaluation module is used for analyzing the influence of the execution of the cognitive task on the motor function parameters and generating a cognitive interference index;
The trend analysis module is used for scoring the motion ability with time stamps in the continuous evaluation period to form time series data, and calculating the ability fluctuation rate by adopting an analysis method of non-uniform sampling adaptation;
the combination analysis module is used for carrying out combination analysis on the cognitive interference index and the ability fluctuation rate and outputting an abnormality prompt;
and the result output module is used for outputting the motor ability score, the cognitive interference index, the ability fluctuation rate and the abnormality prompt as evaluation results.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202511207922.4A CN121171572A (en) | 2025-08-27 | 2025-08-27 | A method and system for assessing the motor function of both lower limbs in stroke patients |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202511207922.4A CN121171572A (en) | 2025-08-27 | 2025-08-27 | A method and system for assessing the motor function of both lower limbs in stroke patients |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN121171572A true CN121171572A (en) | 2025-12-19 |
Family
ID=98040932
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202511207922.4A Pending CN121171572A (en) | 2025-08-27 | 2025-08-27 | A method and system for assessing the motor function of both lower limbs in stroke patients |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN121171572A (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120196257A1 (en) * | 2009-08-27 | 2012-08-02 | Joe Verghese | Cognitive function training to improve motor ability |
| US20140276130A1 (en) * | 2011-10-09 | 2014-09-18 | The Medical Research, Infrastructure and Health Services Fund of the Tel Aviv Medical Center | Virtual reality for movement disorder diagnosis and/or treatment |
| CN112869743A (en) * | 2021-01-15 | 2021-06-01 | 北京理工大学 | Exercise initiation intention neural analysis method considering cognitive distraction |
| CN117064374A (en) * | 2023-07-07 | 2023-11-17 | 北京清华长庚医院 | Intelligent evaluation equipment and evaluation method for cerebral apoplexy patient |
| CN117297583A (en) * | 2023-11-29 | 2023-12-29 | 首都医科大学附属北京天坛医院 | Multimodal proprioceptive perception disorder assessment system and lower limb movement model construction system |
| CN117357059A (en) * | 2023-10-16 | 2024-01-09 | 北京清华长庚医院 | System, method, device and storage medium for evaluating dual tasks of motor cognition |
| CN120280157A (en) * | 2025-06-06 | 2025-07-08 | 成都布鲁奥森信息科技有限责任公司 | Intelligent evaluation method for cognitive impairment after stroke based on brain network dynamic reconstruction |
| WO2025150332A1 (en) * | 2024-01-12 | 2025-07-17 | 国立大学法人大阪大学 | Cognitive function evaluation system |
| CN120473130A (en) * | 2025-04-09 | 2025-08-12 | 华中科技大学同济医学院附属同济医院 | Intelligent assessment method and device for physical rehabilitation of stroke patients |
| CN120526989A (en) * | 2025-05-15 | 2025-08-22 | 东莞市滨海湾中心医院 | A method and system for analyzing lower limb motor function of the elderly based on Internet of Things technology |
-
2025
- 2025-08-27 CN CN202511207922.4A patent/CN121171572A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120196257A1 (en) * | 2009-08-27 | 2012-08-02 | Joe Verghese | Cognitive function training to improve motor ability |
| US20140276130A1 (en) * | 2011-10-09 | 2014-09-18 | The Medical Research, Infrastructure and Health Services Fund of the Tel Aviv Medical Center | Virtual reality for movement disorder diagnosis and/or treatment |
| CN112869743A (en) * | 2021-01-15 | 2021-06-01 | 北京理工大学 | Exercise initiation intention neural analysis method considering cognitive distraction |
| CN117064374A (en) * | 2023-07-07 | 2023-11-17 | 北京清华长庚医院 | Intelligent evaluation equipment and evaluation method for cerebral apoplexy patient |
| CN117357059A (en) * | 2023-10-16 | 2024-01-09 | 北京清华长庚医院 | System, method, device and storage medium for evaluating dual tasks of motor cognition |
| CN117297583A (en) * | 2023-11-29 | 2023-12-29 | 首都医科大学附属北京天坛医院 | Multimodal proprioceptive perception disorder assessment system and lower limb movement model construction system |
| WO2025150332A1 (en) * | 2024-01-12 | 2025-07-17 | 国立大学法人大阪大学 | Cognitive function evaluation system |
| CN120473130A (en) * | 2025-04-09 | 2025-08-12 | 华中科技大学同济医学院附属同济医院 | Intelligent assessment method and device for physical rehabilitation of stroke patients |
| CN120526989A (en) * | 2025-05-15 | 2025-08-22 | 东莞市滨海湾中心医院 | A method and system for analyzing lower limb motor function of the elderly based on Internet of Things technology |
| CN120280157A (en) * | 2025-06-06 | 2025-07-08 | 成都布鲁奥森信息科技有限责任公司 | Intelligent evaluation method for cognitive impairment after stroke based on brain network dynamic reconstruction |
Non-Patent Citations (2)
| Title |
|---|
| R. TISSERAND ET AL.: "Cognitive-motor dual-task interference modulates mediolateral dynamic stability during gait in post-stroke individuals", HUMAN MOVEMENT SCIENCE, 12 March 2018 (2018-03-12), pages 175 - 184 * |
| 向丽莎 等: "认知-运动双任务训练在脑卒中患者康复中的研究进展", 医学研究与教育, vol. 41, no. 3, 30 June 2024 (2024-06-30), pages 8 - 17 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hannink et al. | Sensor-based gait parameter extraction with deep convolutional neural networks | |
| Zhao et al. | Dual channel LSTM based multi-feature extraction in gait for diagnosis of Neurodegenerative diseases | |
| Potluri et al. | Deep learning based gait abnormality detection using wearable sensor system | |
| Woelfle et al. | Wearable sensor technologies to assess motor functions in people with multiple sclerosis: systematic scoping review and perspective | |
| KR20200074951A (en) | Machine learning-based system for identification and monitoring of nervous system disorders | |
| CN116312947B (en) | Immersive ankle and foot rehabilitation training method based on upper limb movement signals and electronic equipment | |
| WO2015199258A1 (en) | Diagnosis model generation system and method | |
| Moeller et al. | Use of lower limb exoskeletons as an assessment tool for human motor performance: A systematic review | |
| KR20210114012A (en) | Diagnosis and Effectiveness of Attention Deficit Hyperactivity Disorder Monitoring | |
| US20230329587A1 (en) | System And Method For Assessing Neuro Muscular Disorder By Generating Biomarkers From The Analysis Of Gait | |
| US20220111257A1 (en) | System, Method and Computer Program Product Configured for Sensor-Based Enhancement of Physical Rehabilitation | |
| US20240324907A1 (en) | Systems and methods for evaluating gait | |
| D’arco et al. | Application of smart insoles for recognition of activities of daily living: A systematic review | |
| CN119626449A (en) | A method, device, electronic device and storage medium for generating exercise prescription | |
| Duong et al. | Ecological validation of machine learning models for spatiotemporal gait analysis in free-living environments using instrumented insoles | |
| Mohapatra et al. | Wearable network for multilevel physical fatigue prediction in manufacturing workers | |
| KR102357041B1 (en) | Method for Analyzing and Predicting Disease by Using Artificial Intelligence | |
| Meigal et al. | Ambient Assisted Living At-Home Laboratory for Motor Status Diagnostics in Parkinson's Disease Patients and Aged People | |
| Cai et al. | Closed-looped sensing and stimulation system for Parkinson’s disease early diagnosis and rehabilitation | |
| CN121171572A (en) | A method and system for assessing the motor function of both lower limbs in stroke patients | |
| Hermez et al. | Gait deviation and neurological diseases: a comparative study of quantitative measures | |
| CN119517398A (en) | Medical indicator risk warning method and device, electronic equipment and storage medium | |
| Ballesteros et al. | Online estimation of rollator user condition using spatiotemporal gait parameters | |
| Rozanski et al. | Recording context matters: Differences in gait parameters collected by the OneStep smartphone application | |
| Matey-Sanz et al. | Implementing and Evaluating the Timed Up and Go Test Automation Using Smartphones and Smartwatches |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |