Artificial intelligence-based brain control gait training method and system for Parkinson's disease

Abstract

The invention discloses a brain control gait training method and system for Parkinson's disease based on artificial intelligence, and belongs to the technical field of medical treatment. The problems that active participation of the brain of a patient is lacked in the existing training process, and the gait training effect and the effect persistence are limited are solved, electroencephalogram signals are analyzed through the brain-computer interface technology, the motion intention of the patient can be recognized and converted into a control instruction, and the training effect is improved. The gait training device executes corresponding actions according to the control instruction so as to help the patient to carry out gait training, gait data of the patient is collected through a sensor to be evaluated, the gait training effect of the patient can be known, meanwhile, electroencephalogram signals, gait data and training effect evaluation data of the patient are continuously accumulated, and the training effect evaluation accuracy is improved. The training scheme can be provided according to the gait characteristics of the patient, the performance of the model is continuously monitored along with the change and update of the data, and the model can be continuously updated and improved according to the change of the data, so that the training scheme can be continuously optimized.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to a brain-controlled gait training method and system for Parkinson's disease based on artificial intelligence. Background Art

[0002] Parkinson's disease is a degenerative disease of the central nervous system characterized by resting tremor, muscle rigidity, bradykinesia and gait disorders. In the early and middle stages of Parkinson's disease, gait disorders are mainly manifested in reduced arm swing, decreased gait speed and stride, and there is no obvious specificity. As the disease progresses, in the late stage of Parkinson's disease, more serious and characteristic gait disorders such as panic gait and freezing gait appear. The interaction between the basal ganglia and the secondary motor cortex of the brain is an important link in motor control. One of the main causes of gait disorders in Parkinson's patients is the disorder of the internal rhythm of the basal ganglia. At present, for the gait training of Parkinson's patients, some studies focus on the detection of abnormal gait, especially late panic gait and freezing gait, and then alleviate abnormal gait and improve walking function through external rhythmic prompts, virtual reality and other means. However, the existing gait training methods or systems lack the active participation of the patient's brain in the training process, which still has certain limitations in the effect and sustainability of gait training; therefore, it does not meet the existing needs. In this regard, we propose a brain-controlled gait training method and system for Parkinson's disease based on artificial intelligence. Summary of the invention

[0003] The purpose of the present invention is to provide a brain-controlled gait training method and system for Parkinson's disease based on artificial intelligence. The patient's EEG signals are collected by brain wave detection equipment, and the patient's movement intentions are identified and converted into control instructions through brain-computer interface technology, so that the gait training equipment performs corresponding actions according to the control instructions, thereby helping the patient to perform gait training. By using sensors to collect the patient's gait data, the patient's gait can be evaluated, and the patient's EEG signals, gait data and training effect evaluation data are continuously accumulated, so that a training plan can be provided according to the patient's gait characteristics, thereby solving the problems raised in the above-mentioned background technology.

[0004] To achieve the above object, the present invention provides the following technical solution: a brain-controlled gait training method for Parkinson's disease based on artificial intelligence, comprising the following steps: Collect the patient's EEG signals through EEG detection equipment and pre-process the collected EEG signals; The pre-processed EEG signals are analyzed in real time through brain-computer interface technology to identify the patient's movement intention and convert the identified movement intention into control instructions; The control instruction is transmitted to the gait training device by wireless or wired means, and the gait training device performs corresponding actions according to the received control instruction; Use sensors to collect the patient's gait data, and analyze the patient's gait data, including gait cycle analysis, gait dynamics analysis, and gait biomechanics analysis, to evaluate the patient's gait; Accumulate patients' EEG signals, gait data, and training effect evaluation data, use artificial intelligence algorithms to process and analyze the accumulated data, identify effective gait patterns through machine learning technology, and then provide training plans based on the patients' gait data to help users improve their gait.

[0005] Furthermore, the pre-processed EEG signals are analyzed in real time through brain-computer interface technology, which specifically includes the following processes: After collecting the patient's EEG signals, the EEG signals are preprocessed and feature extracted; After preprocessing and feature extraction of brain wave signals, data analysis and identification are performed through computer algorithms; The patient's movement intention is identified through analysis and identification, and the movement intention is converted into readable instruction information, and finally the control instruction is obtained.

[0006] Furthermore, the preprocessed EEG signal is analyzed in real time through the brain-computer interface, and after the EEG signal is preprocessed, the data quality of the preprocessed EEG signal is monitored for abnormalities, including: Extract the signal-to-noise ratio and artifacts corresponding to each preprocessed brain wave signal; The data quality evaluation coefficient is obtained according to the signal-to-noise ratio and artifacts corresponding to each brain wave signal after the preprocessing; wherein the data quality evaluation coefficient is obtained by the following formula: Wherein, P represents the signal data evaluation coefficient; SNR represents the signal-to-noise ratio corresponding to each brain wave signal; PSD (f) represents the power spectral density function, which is used to describe the power distribution of the signal at each frequency component; r represents the adjustment factor, and the adjustment factor ranges from 0.37 to 1.32; N f represents the Nyquist frequency; W represents the artifact influence coefficient, and the artifact influence coefficient is obtained by the following formula: Where W represents the artifact influence coefficient; a and b represent the preset weight coefficients; P t Indicates the signal time proportion of the frequency corresponding to the brain wave signal being outside the preset frequency reference range; A0 indicates the preset threshold value of the area occupied by the artifact; A indicates the value of the area occupied by the artifact; f c Indicates the proportion of non-physiological signals in brain wave signals; Comparing the data quality evaluation coefficient with a preset evaluation coefficient threshold; When the data quality evaluation coefficient is not lower than a preset evaluation coefficient threshold, it is determined that the data quality of the preprocessed brain wave signal meets the data quality requirements; When the data quality evaluation coefficient is lower than a preset evaluation coefficient threshold, it is determined that the data quality of the preprocessed brain wave signal does not meet the data quality requirements; When the data quality of the preprocessed brain wave signal does not meet the data quality requirements, a signal quality evaluation is performed on each electrode channel, and the electrode channels whose evaluation quality does not meet the preset requirements are extracted.

[0007] Further, when the data quality of the preprocessed brain wave signal does not meet the data quality requirements, a signal quality evaluation is performed on each electrode channel, and the electrode channels whose evaluation quality does not meet the preset requirements are extracted, including: Extract the signal amplitude parameters of each electrode channel in turn; Extracting the highest signal value and the lowest signal value from the signal amplitude parameters of each electrode channel; Obtaining the maximum amplitude difference of each electrode channel according to the highest signal value and the lowest signal value; Comparing the maximum amplitude difference of each electrode channel with a preset amplitude difference threshold; When the maximum amplitude difference of the electrode channel exceeds the preset amplitude difference threshold, it indicates that the electrode channel corresponding to the maximum amplitude difference of the electrode channel exceeds the preset amplitude difference threshold does not meet the preset requirements; When the maximum amplitude difference of the electrode channel does not exceed the preset amplitude difference threshold, the signal data value between the highest signal value and the lowest signal value of the electrode channel corresponding to the maximum amplitude difference of the electrode channel not exceeding the preset amplitude difference threshold is retrieved to obtain the signal data evaluation coefficient; wherein, the signal data evaluation coefficient is obtained by the following formula: Where Q represents the signal data evaluation coefficient; F m Indicates the maximum amplitude difference; F m0 represents the preset amplitude difference threshold; n represents the number of signal data values ​​between the highest signal value and the lowest signal value; F i Indicates the amplitude value corresponding to the i-th signal; F z Indicates the corresponding amplitude midpoint between the highest signal value and the lowest signal value; F b represents the amplitude difference between the highest signal value and the lowest signal value; x represents the adjustment coefficient, and the range of the adjustment coefficient is 0.83-2.18; L represents the Lyapunov exponent corresponding to the brain wave signal analysis; Comparing the signal data evaluation coefficient with a preset signal data evaluation coefficient threshold; When the signal data evaluation coefficient is lower than a preset signal data evaluation coefficient threshold, it is determined that the electrode channel whose signal data evaluation coefficient is lower than the preset signal data evaluation coefficient threshold does not meet the preset requirements; Mark and warn electrode channels that do not meet preset requirements.

[0008] Furthermore, effective gait patterns are identified through machine learning technology, which specifically includes the following processes: Accumulate previous patients’ EEG signals, gait data, and training effect evaluation data; Select a machine learning model based on the accumulated data, for example, a convolutional neural network or a recurrent neural network; Use the selected model and train it with the accumulated data to adjust the model parameters; Deploy the trained model to actual applications, identify effective gait patterns through the model, and then provide training plans based on the patient's gait data; As the data changes and is updated, the performance of the model is continuously monitored, and the model is continuously updated and improved based on the changes and updates of the data.

[0009] The brain-controlled gait training system for Parkinson's disease based on artificial intelligence is used to implement the brain-controlled gait training method for Parkinson's disease based on artificial intelligence, including: The EEG signal acquisition unit, including the EEG detection device, is used to: The patient's EEG signals are collected by EEG detection equipment, and the collected EEG signals are preprocessed and feature extracted. The patient's movement intention is then identified by a computer algorithm, and the movement intention is converted into a control instruction. The converted control instruction is transmitted to the gait training control unit by wireless or wired means; A gait training control unit, including a gait training device, is used for: Receive control instructions transmitted by the EEG signal acquisition unit, and control the gait training device through the control instructions to help the patient perform gait training; Training effectiveness evaluation unit, used for: Use sensors to collect and analyze the patient's gait data to analyze the patient's walking rhythm, stride and stability parameters, and then evaluate the patient's gait; Optimization and improvement unit for: Accumulate patients' EEG signals, gait data, and training effect evaluation data, use artificial intelligence algorithms to process and analyze the accumulated data, identify effective gait patterns through machine learning technology, and then provide training plans based on the patients' gait data.

[0010] Furthermore, the EEG signal acquisition unit comprises: Signal acquisition module for: Obtain the patient's neural information through brain wave detection equipment, and then collect brain wave signals; Signal processing modules for: Perform filtering and denoising preprocessing operations on the collected EEG signals, and perform feature extraction on the preprocessed EEG signals; Signal analysis module for: The EEG signals are analyzed and identified through computer algorithms to recognize the patient's movement intention, and the recognized movement intention is converted into control instructions, which are then transmitted to the gait training control unit wirelessly or wiredly.

[0011] Furthermore, the signal processing module is specifically: Filtering: used to remove noise to improve the quality and analyzability of EEG signals; Denoising: used to eliminate interfering noise that is not related to EEG activity; Feature extraction: used to extract features from EEG signals. The features in EEG signals are divided into time domain features, frequency domain features, and time-frequency domain features.

[0012] Furthermore, the training effect evaluation unit comprises: Data collection module for: When the patient is undergoing gait training, the patient's gait data is collected through sensors, wherein the gait data includes step length, step speed, and joint angle; Data analysis module for: Analyze gait data, including gait cycle analysis, gait dynamics analysis, and gait biomechanics analysis, to analyze the patient's walking rhythm, stride, and stability; Effect evaluation module, used to: Based on the analysis results of the data analysis module, the patient's gait is evaluated to assess the symmetry, stability and coordination of the patient's gait.

[0013] Furthermore, the optimization and improvement unit includes: Data accumulation module for: Accumulate the patient's previous EEG signals, gait data, and training effect evaluation data; Training program modules for: Selecting a machine learning model based on the accumulated data, using the selected model and training it with the accumulated data, identifying effective gait patterns through the trained model, and then providing a training plan based on the patient's gait data; Continuous improvement module for: As the data changes and is updated, the performance of the model is continuously monitored, and the operations of the data accumulation module and the training scheme module are continuously executed as the data changes, thereby continuously updating and improving the model according to the changes in the data.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention collects the patient's EEG signal through an EEG detection device, and based on the collected EEG signal, identifies the patient's movement intention through brain-computer interface technology and converts it into a control instruction, so that the gait training device performs the corresponding action according to the control instruction, thereby helping the patient to perform gait training, and collects the patient's gait data through a sensor, and analyzes the gait data, so that the patient's gait can be evaluated. Through the patient's gait evaluation, the effect of the patient's gait training can be understood, and a training plan can be provided for the patient based on the evaluation result. At the same time, the patient's EEG signal, gait data and training effect evaluation data are continuously accumulated, and the accumulated data are processed and analyzed using an artificial intelligence algorithm, so that an effective gait pattern can be identified through machine learning technology, and then a training plan is provided to the patient according to the patient's gait data, further helping the user to improve the gait, and, as the data changes and updates, the performance of the model is continuously monitored, and data accumulation and analysis operations need to be continuously performed due to changes in the data, and then the model is continuously updated and improved according to the changes in the data, so that the training plan can be continuously optimized to adapt to the needs and abilities of patients at different stages. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of the brain-controlled gait training method for Parkinson's disease based on artificial intelligence of the present invention; Figure 2 It is a structural schematic diagram of the brain-controlled gait training system for Parkinson's disease based on artificial intelligence of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.

[0017] In order to solve the technical problem that the existing gait training methods or systems lack the active participation of the patient's brain in the training process, which still has certain limitations in the effect and sustainability of gait training, please refer to Figure 1-Figure 2 , this embodiment provides the following technical solutions: The brain-controlled gait training system for Parkinson's disease based on artificial intelligence is used to implement the brain-controlled gait training method for Parkinson's disease based on artificial intelligence, including: The EEG signal acquisition unit, including the EEG detection device, is used to: The patient's EEG signals are collected by EEG detection equipment, and the collected EEG signals are preprocessed and feature extracted. The patient's movement intention is then identified by a computer algorithm, and the movement intention is converted into a control instruction. The converted control instruction is transmitted to the gait training control unit by wireless or wired means; A gait training control unit, including a gait training device, is used for: Receive control instructions transmitted by the EEG signal acquisition unit, and control the gait training device through the control instructions to help the patient perform gait training; Training effectiveness evaluation unit, used for: Use sensors to collect and analyze the patient's gait data to analyze the patient's walking rhythm, stride and stability parameters, and then evaluate the patient's gait; Optimization and improvement unit for: Accumulate patients' EEG signals, gait data, and training effect evaluation data, use artificial intelligence algorithms to process and analyze the accumulated data, identify effective gait patterns through machine learning technology, and then provide training plans based on the patients' gait data.

[0018] The technical effect of the above content is: the patient's EEG signals can be collected by the EEG signal acquisition unit, and the collected EEG signals will be preprocessed and feature extracted, and then the patient's movement intention can be identified through the brain-computer interface technology, and the movement intention can be converted into a control instruction, so that the gait training control unit can control the gait training equipment to perform corresponding actions according to the control instructions, thereby helping the patient to perform gait training. During the patient's training, the training effect evaluation unit will collect and analyze the patient's gait data, so as to evaluate the patient's gait, which is convenient for medical staff to understand the patient's training situation and effect. At the same time, the optimization and improvement unit will continue to accumulate the patient's EEG signals, gait data and training effect evaluation data, so as to identify effective gait patterns through machine learning technology, and then provide training plans based on the patient's gait characteristics.

[0019] The EEG signal acquisition unit comprises: Signal acquisition module for: Obtain the patient's neural information through brain wave detection equipment, and then collect brain wave signals; Signal processing modules for: Perform filtering and denoising preprocessing operations on the collected EEG signals, and perform feature extraction on the preprocessed EEG signals; Signal analysis module for: The EEG signals are analyzed and identified through computer algorithms to recognize the patient's movement intention, and the recognized movement intention is converted into control instructions, which are then transmitted to the gait training control unit wirelessly or wiredly.

[0020] Signal processing module, specifically: Filtering: used to remove noise to improve the quality and analyzability of EEG signals; Among them, filtering processing achieves the purpose of removing noise from EEG signals by selecting filters and designing filter parameters; Denoising: used to eliminate interfering noise that is not related to EEG activity; Feature extraction: used to extract features from EEG signals. Feature extraction of EEG signals is to further analyze and process the signals in order to extract useful information from EEG activities. The features in EEG signals are divided into time domain features, frequency domain features, and time-frequency domain features. Among them, the time domain features mainly include mean, variance, peak and maximum value, which are used to describe the basic statistical characteristics of EEG signals; the frequency domain features extract the spectral characteristics of the signal, such as power spectral density, energy distribution, etc., by performing Fourier transform or wavelet transform on the signal, which are used to describe the frequency components of the EEG signal; the time-frequency domain features combine the characteristics of the time domain and frequency domain. By using time-frequency domain analysis methods, such as short-time Fourier transform and continuous wavelet transform, the time domain and frequency domain information of the signal can be obtained simultaneously.

[0021] The technical effects of the above content are as follows: brain-computer interface technology is a system that directly connects the brain with external devices. It can convert brain activities into commands that can be understood by machines, and vice versa. The core of brain-computer interface technology is to decode brain activities and realize the conversion of human intentions into machine commands, thereby achieving efficient communication and control between humans and machines. In gait training, the application of brain-computer interface technology can help patients control auxiliary equipment through thoughts, thereby improving and restoring gait. Specifically in the present invention, firstly, the patient's neural information is captured by brain wave detection equipment, so that the patient's brain wave signals can be collected, and the collected The EEG signals received need to be preprocessed for analysis, such as filtering and denoising, to improve the quality and analyzability of the EEG signals, and then feature extraction is performed on the preprocessed EEG signals. After that, the brain wave signals are preprocessed and feature extracted, and then data analysis and identification are performed using computer algorithms to identify the patient's movement intentions. The identified movement intentions will be converted into readable command information to obtain control instructions. Finally, the control instructions will be transmitted to the gait training device wirelessly or wiredly. These control instructions can be used to control the gait training device, thereby helping patients with gait training.

[0022] Training effect evaluation unit, including: Data collection module for: When the patient is undergoing gait training, the patient's gait data is collected through sensors, wherein the gait data includes step length, step speed, and joint angle; Data analysis module for: Analyze gait data, including gait cycle analysis, gait dynamics analysis, and gait biomechanics analysis, to analyze the patient's walking rhythm, stride, and stability; in, Gait cycle analysis: used to analyze the time from when the foot contacts the ground to when the heel of that side touches the ground again during walking. Each walking cycle is divided into two stages: stance phase and stride phase. Gait dynamics analysis: used to analyze the interaction between the patient and the ground during gait training; Gait biomechanical analysis: used to analyze the biomechanical principles and parameters of patients' gait training, and how these principles and parameters affect the normal progress of gait Effect evaluation module, used to: Based on the analysis results of the data analysis module, the patient's gait is evaluated to assess the symmetry, stability and coordination of the patient's gait.

[0023] The technical effect of the above content is: when the patient is performing gait training, the data collection module can collect the patient's gait data through sensors (i.e. wearable sensors), such as the patient's step length, step speed, joint angle, etc. during the training process, and then, the data analysis module can analyze the collected gait data to analyze the patient's walking rhythm, stride, stability and other parameters. Finally, the effect evaluation module can evaluate the patient's gait based on the analysis results, thereby evaluating the symmetry, stability and coordination of the patient's gait. By evaluating the patient's gait data, the patient's training situation and effect can be understood, and a scientific basis can be provided for gait training, which can help patients train better.

[0024] Optimization and improvement units, including: Data accumulation module for: Accumulate the patient's previous EEG signals, gait data, and training effect evaluation data; Training program modules for: Selecting a machine learning model based on the accumulated data, using the selected model and training it with the accumulated data, identifying effective gait patterns through the trained model, and then providing a training plan based on the patient's gait data; Continuous improvement module for: As the data changes and is updated, the performance of the model is continuously monitored, and the operations of the data accumulation module and the training scheme module are continuously executed as the data changes, thereby continuously updating and improving the model according to the changes in the data.

[0025] The technical effect of the above content is: the patient's previous EEG signals, gait data and training effect evaluation data can be accumulated through the data accumulation module, and the accumulated data can be processed and analyzed using artificial intelligence algorithms. First, it is necessary to select a suitable model based on the accumulated data. A convolutional neural network or a recurrent neural network can be selected. After the model is selected, it can be trained through the accumulated data, so that the model parameters can be adjusted to improve the accuracy of the model, which is beneficial to subsequent data analysis. Then, the trained model is deployed and applied, so that effective gait patterns can be identified, so that training plans can be provided to patients. The training plans can be fed back to the gait training equipment, so as to better help users improve their gait. At the same time, the performance of the model is continuously monitored. As the data changes and updates, new EEG signals, gait data and training effect evaluation data need to be continuously accumulated, so that the operations of the data accumulation module and the training plan module need to be continuously executed, and then the model is continuously updated and improved according to the changes in the data, so that the training plan can be continuously optimized to adapt to the needs and abilities of patients at different stages.

[0026] Specifically, this embodiment also proposes a brain-controlled gait training method for Parkinson's disease based on artificial intelligence, comprising the following steps: The patient's EEG signals are collected through EEG detection equipment, and the collected EEG signals are pre-processed to improve the quality and analyzability of the EEG signals; The pre-processed EEG signals are analyzed in real time through brain-computer interface technology to identify the patient's movement intentions, such as walking, stopping, turning, etc., and the identified movement intentions are converted into control instructions, which will be used to control the gait training equipment; The control instruction is transmitted to the gait training device by wireless or wired means, and the gait training device performs corresponding actions according to the received control instruction to realize gait training; Use sensors to collect the patient's gait data, such as step length, gait speed, joint angle, etc., and analyze the patient's gait data, including gait cycle analysis, gait dynamics analysis, and gait biomechanics analysis, to analyze the patient's walking rhythm, stride, stability and other parameters, thereby providing a scientific basis for gait training, and then evaluate the patient's gait, including gait symmetry, stability, coordination and other aspects; Accumulate patients' EEG signals, gait data, and training effect evaluation data, use artificial intelligence algorithms to process and analyze the accumulated data, identify effective gait patterns through machine learning technology, and then provide training plans based on the patients' gait data to help users improve their gait.

[0027] Among them, the pre-processed EEG signals are analyzed in real time through brain-computer interface technology, which specifically includes the following processes: After collecting the patient's EEG signals, the EEG signals are preprocessed and feature extracted; After preprocessing and feature extraction of brain wave signals, data analysis and identification are performed through computer algorithms; The patient's movement intention is identified through analysis and identification, and the movement intention is converted into readable instruction information, and finally the control instruction is obtained.

[0028] Among them, the effective gait pattern is identified through machine learning technology, which specifically includes the following processes: Accumulate previous patients’ EEG signals, gait data, and training effect evaluation data; Select a machine learning model based on the accumulated data, for example, a convolutional neural network or a recurrent neural network; Use the selected model and train it with the accumulated data to adjust the model parameters to improve the accuracy and generalization ability of the model; Deploy the trained model to actual applications, identify effective gait patterns through the model, and then provide training plans based on the patient's gait data; As the data changes and is updated, the performance of the model is continuously monitored, and the model is continuously updated and improved based on the changes and updates of the data.

[0029] Working principle: The patient's EEG signals are collected through the EEG detection equipment, and the EEG signals are preprocessed and feature extracted. The EEG signals are then analyzed through the brain-computer interface technology, so that the patient's movement intention can be identified and converted into readable instruction information, thereby obtaining control instructions, so that the gait training equipment can control the gait training equipment to perform corresponding actions according to the control instructions, thereby helping the patient to perform gait training. By using sensors to collect and analyze the patient's gait data, the patient's gait can be evaluated, so as to understand the effect of the patient's gait training, and provide a scientific basis for gait training, which can help patients to train better. At the same time, by continuously accumulating patients' EEG signals, gait data and training effect evaluation data, effective gait patterns can be identified through machine learning technology, and then training plans can be provided to patients based on their gait data to further help users improve their gait. In addition, as the data changes and updates, it is necessary to continuously accumulate new EEG signals, gait data and training effect evaluation data. Due to changes in data, data accumulation and analysis operations need to be continuously performed, and thus operations of the data accumulation module and the training plan module need to be continuously performed, so that the model can be continuously updated and improved, and the training plan can be continuously optimized to adapt to the needs and abilities of patients at different stages.

[0030] Specifically, the preprocessed EEG signal is analyzed in real time through the brain-computer interface, and after the EEG signal is preprocessed, the data quality of the preprocessed EEG signal is monitored for abnormalities, including: Extract the signal-to-noise ratio and artifacts corresponding to each preprocessed brain wave signal; The data quality evaluation coefficient is obtained according to the signal-to-noise ratio and artifacts corresponding to each brain wave signal after the preprocessing; wherein the data quality evaluation coefficient is obtained by the following formula: Wherein, P represents the signal data evaluation coefficient; SNR represents the signal-to-noise ratio corresponding to each brain wave signal; PSD (f) represents the power spectral density function, which is used to describe the power distribution of the signal at each frequency component; r represents the adjustment factor, and the adjustment factor ranges from 0.37 to 1.32; N f represents the Nyquist frequency; W represents the artifact influence coefficient, and the artifact influence coefficient is obtained by the following formula: Where W represents the artifact influence coefficient; a and b represent the preset weight coefficients; P t Indicates the signal time proportion of the frequency corresponding to the brain wave signal being outside the preset frequency reference range; A0 indicates the preset threshold value of the area occupied by the artifact; A indicates the value of the area occupied by the artifact; f c Indicates the proportion of non-physiological signals in brain wave signals; Comparing the data quality evaluation coefficient with a preset evaluation coefficient threshold; When the data quality evaluation coefficient is not lower than a preset evaluation coefficient threshold, it is determined that the data quality of the preprocessed brain wave signal meets the data quality requirements; When the data quality evaluation coefficient is lower than a preset evaluation coefficient threshold, it is determined that the data quality of the preprocessed brain wave signal does not meet the data quality requirements; When the data quality of the preprocessed brain wave signal does not meet the data quality requirements, a signal quality evaluation is performed on each electrode channel, and the electrode channels whose evaluation quality does not meet the preset requirements are extracted.

[0031] The technical effect of the above technical solution is: by performing real-time analysis on the pre-processed EEG signals through the brain-computer interface, EEG data can be quickly acquired and processed, thereby improving the real-time and efficiency of data analysis.

[0032] The signal-to-noise ratio and artifacts of each brain wave signal are accurately extracted, providing accurate basic data for subsequent data quality evaluation. The solution has established a complete data quality evaluation system, which can comprehensively evaluate the data quality of brain wave signals after preprocessing by calculating the data quality evaluation coefficient. Real-time monitoring of data quality can promptly detect and handle data anomalies to ensure the accuracy and reliability of analysis results.

[0033] When the data quality evaluation coefficient is lower than the preset threshold, the problem can be quickly located, that is, the signal quality of each electrode channel is evaluated, and the electrode channels that do not meet the preset requirements are extracted. This helps to manage the electrode channels in a refined manner and adjust or optimize the electrode settings in time to improve data quality.

[0034] Through real-time analysis, monitoring and exception handling, the stability and reliability of the brain-computer interface system in various application scenarios can be ensured. This is of great significance for brain-computer interface applications that require high precision and high reliability (such as medical diagnosis, neuroscience research, etc.). This technical solution provides new ideas and methods for data processing and analysis of brain-computer interface technology, which will help promote the further development and application of brain-computer interface technology. By continuously optimizing the data quality evaluation system and exception handling methods, the overall performance and user experience of the brain-computer interface system can be improved.

[0035] In summary, the above-mentioned technical solution improves the real-time, accuracy and reliability of brain-computer interface system data processing through real-time analysis, monitoring and exception handling, providing strong support for the further development and application of brain-computer interface technology.

[0036] Specifically, when the data quality of the preprocessed brain wave signal does not meet the data quality requirements, a signal quality evaluation is performed on each electrode channel, and the electrode channels whose evaluation quality does not meet the preset requirements are extracted, including: Extract the signal amplitude parameters of each electrode channel in turn; Extracting the highest signal value and the lowest signal value from the signal amplitude parameters of each electrode channel; Obtaining the maximum amplitude difference of each electrode channel according to the highest signal value and the lowest signal value; Comparing the maximum amplitude difference of each electrode channel with a preset amplitude difference threshold; When the maximum amplitude difference of the electrode channel exceeds the preset amplitude difference threshold, it indicates that the electrode channel corresponding to the maximum amplitude difference of the electrode channel exceeds the preset amplitude difference threshold does not meet the preset requirements; When the maximum amplitude difference of the electrode channel does not exceed the preset amplitude difference threshold, the signal data value between the highest signal value and the lowest signal value of the electrode channel corresponding to the maximum amplitude difference of the electrode channel not exceeding the preset amplitude difference threshold is retrieved to obtain the signal data evaluation coefficient; wherein, the signal data evaluation coefficient is obtained by the following formula: Where Q represents the signal data evaluation coefficient; F m Indicates the maximum amplitude difference; F m0 represents the preset amplitude difference threshold; n represents the number of signal data values ​​between the highest signal value and the lowest signal value; F i Indicates the amplitude value corresponding to the i-th signal; F z Indicates the corresponding amplitude midpoint between the highest signal value and the lowest signal value; F b represents the amplitude difference between the highest signal value and the lowest signal value; x represents the adjustment coefficient, and the range of the adjustment coefficient is 0.83-2.18; L represents the Lyapunov exponent corresponding to the brain wave signal analysis; Comparing the signal data evaluation coefficient with a preset signal data evaluation coefficient threshold; When the signal data evaluation coefficient is lower than a preset signal data evaluation coefficient threshold, it is determined that the electrode channel whose signal data evaluation coefficient is lower than the preset signal data evaluation coefficient threshold does not meet the preset requirements; Mark and warn electrode channels that do not meet preset requirements.

[0037] The technical effect of the above technical solution is: by extracting the signal amplitude parameters of each electrode channel, especially the highest signal value and the lowest signal value, the maximum amplitude difference of each electrode channel can be accurately calculated. This step helps to gain a deeper understanding of the signal fluctuations of each electrode channel. By comparing the maximum amplitude difference of each electrode channel with the preset amplitude difference threshold, those electrode channels with excessive amplitude fluctuations can be quickly screened out. These electrode channels may have poor signal quality due to poor contact, damage, or external interference, thereby affecting the overall data quality.

[0038] For electrode channels whose amplitude difference does not exceed the preset threshold, the signal data quality is further evaluated by calculating the signal data evaluation coefficient. This step takes into account multiple factors such as the distribution of signal amplitude values, the amplitude median value, the amplitude difference value, and combines complex dynamic characteristics such as the Lyapunov exponent to provide a more comprehensive and accurate signal quality assessment. When the signal data evaluation coefficient is lower than the preset threshold, the system can automatically mark and warn these electrode channels that do not meet the preset requirements. This helps researchers or operators to promptly discover and deal with problematic electrode channels to avoid adverse effects on the overall experimental results or application effects.

[0039] Through the above technical solution, it can be ensured that only electrode channels with signal quality that meets the requirements are used for subsequent data analysis and processing. This helps to improve the stability and reliability of the brain-computer interface system and reduce errors and interference caused by electrode channel quality problems. This technical solution provides a new idea and method for the quality assessment of electrode channels in brain-computer interface technology. With the continuous optimization and improvement of technology, the signal acquisition quality of electrode channels can be further improved, promoting the application and development of brain-computer interface technology in medical treatment, rehabilitation, neuroscience research and other fields.

[0040] In summary, this technical solution effectively improves the stability and reliability of the brain-computer interface system through refined electrode channel quality assessment, multi-dimensional signal data quality evaluation, and real-time early warning and problem tracking mechanism, providing strong support for the optimization and development of brain-computer interface technology.

[0041] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0042] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A brain-controlled gait training method for Parkinson's disease based on artificial intelligence, characterized in that: The following steps are involved: Collect the patient's EEG signals through EEG detection equipment and pre-process the collected EEG signals; The pre-processed EEG signals are analyzed in real time through brain-computer interface technology to identify the patient's movement intention and convert the identified movement intention into control instructions; The control instruction is transmitted to the gait training device by wireless or wired means, and the gait training device performs corresponding actions according to the received control instruction; Use sensors to collect the patient's gait data, and analyze the patient's gait data, including gait cycle analysis, gait dynamics analysis, and gait biomechanics analysis, to evaluate the patient's gait; Accumulate patients' EEG signals, gait data, and training effect evaluation data, use artificial intelligence algorithms to process and analyze the accumulated data, identify effective gait patterns through machine learning, and provide training plans based on patients' gait data.

2. The method for brain-controlled gait training for Parkinson's disease based on artificial intelligence according to claim 1, characterized in that: The pre-processed EEG signals are analyzed in real time through the brain-computer interface, which specifically includes the following processes: After collecting the patient's EEG signals, the EEG signals are preprocessed and feature extracted; After preprocessing and feature extraction of brain wave signals, data analysis and identification are performed through computer algorithms; The patient's movement intention is identified through analysis and identification, and the movement intention is converted into readable instruction information, and finally the control instruction is obtained.

3. The method for brain-controlled gait training for Parkinson's disease based on artificial intelligence according to claim 1, characterized in that: The pre-processed EEG signal is analyzed in real time through the brain-computer interface, and after the pre-processing of the EEG signal, the data quality of the pre-processed EEG signal is monitored for abnormalities, including: Extract the signal-to-noise ratio and artifacts corresponding to each preprocessed brain wave signal; The data quality evaluation coefficient is obtained according to the signal-to-noise ratio and artifacts corresponding to each brain wave signal after the preprocessing; wherein the data quality evaluation coefficient is obtained by the following formula: Wherein, P represents the signal data evaluation coefficient; SNR represents the signal-to-noise ratio corresponding to each brain wave signal; PSD (f) represents the power spectrum density function, which is used to describe the power distribution of the signal at each frequency component; r represents the adjustment factor, and the adjustment factor ranges from 0.37 to 1.32; N f represents the Nyquist frequency; W represents the artifact influence coefficient, and the artifact influence coefficient is obtained by the following formula: Where W represents the artifact influence coefficient; a and b represent the preset weight coefficients; P t Indicates the signal time proportion of the frequency corresponding to the brain wave signal being outside the preset frequency reference range; A0 indicates the preset threshold value of the area occupied by the artifact; A indicates the value of the area occupied by the artifact; f c Indicates the proportion of non-physiological signals in the brain wave signal; compares the data quality evaluation coefficient with a preset evaluation coefficient threshold; When the data quality evaluation coefficient is not lower than a preset evaluation coefficient threshold, it is determined that the data quality of the preprocessed brain wave signal meets the data quality requirements; When the data quality evaluation coefficient is lower than a preset evaluation coefficient threshold, it is determined that the data quality of the preprocessed brain wave signal does not meet the data quality requirements; When the data quality of the preprocessed brain wave signal does not meet the data quality requirements, a signal quality evaluation is performed on each electrode channel, and the electrode channels whose evaluation quality does not meet the preset requirements are extracted.

4. The method for brain-controlled gait training for Parkinson's disease based on artificial intelligence according to claim 3, characterized in that: When the data quality of the preprocessed brain wave signal does not meet the data quality requirements, a signal quality evaluation is performed on each electrode channel, and the electrode channels whose evaluation quality does not meet the preset requirements are extracted, including: Extract the signal amplitude parameters of each electrode channel in turn; Extracting the highest signal value and the lowest signal value from the signal amplitude parameters of each electrode channel; Obtaining the maximum amplitude difference of each electrode channel according to the highest signal value and the lowest signal value; Comparing the maximum amplitude difference of each electrode channel with a preset amplitude difference threshold; When the maximum amplitude difference of the electrode channel exceeds the preset amplitude difference threshold, it indicates that the electrode channel corresponding to the maximum amplitude difference of the electrode channel exceeds the preset amplitude difference threshold does not meet the preset requirements; When the maximum amplitude difference of the electrode channel does not exceed the preset amplitude difference threshold, the signal data value between the highest signal value and the lowest signal value of the electrode channel corresponding to the maximum amplitude difference of the electrode channel not exceeding the preset amplitude difference threshold is retrieved to obtain the signal data evaluation coefficient; wherein, the signal data evaluation coefficient is obtained by the following formula: Where Q represents the signal data evaluation coefficient; F m Indicates the maximum amplitude difference; F m0 represents the preset amplitude difference threshold; n represents the number of signal data values ​​between the highest signal value and the lowest signal value; F i Indicates the amplitude value corresponding to the i-th signal; F z Indicates the corresponding amplitude midpoint between the highest signal value and the lowest signal value; F b represents the amplitude difference between the highest signal value and the lowest signal value; x represents the adjustment coefficient, and the range of the adjustment coefficient is 0.83-2.18; L represents the Lyapunov exponent corresponding to the brain wave signal analysis; Comparing the signal data evaluation coefficient with a preset signal data evaluation coefficient threshold; When the signal data evaluation coefficient is lower than a preset signal data evaluation coefficient threshold, it is determined that the electrode channel whose signal data evaluation coefficient is lower than the preset signal data evaluation coefficient threshold does not meet the preset requirements; Mark and warn electrode channels that do not meet preset requirements.

5. The brain-controlled gait training method for Parkinson's disease based on artificial intelligence according to claim 1, characterized in that: Identifying effective gait patterns through machine learning includes the following processes: Accumulate previous patients’ EEG signals, gait data, and training effect evaluation data; Select a machine learning model based on the accumulated data; Use the selected model and train it with the accumulated data to adjust the model parameters; Deploy the trained model to actual applications, identify effective gait patterns through the model, and then provide training plans based on the patient's gait data; Monitor the performance of the model based on changes and updates to the data, and update and improve the model based on changes and updates to the data.

6. An artificial intelligence-based brain-controlled gait training system for Parkinson's disease, used to implement the artificial intelligence-based brain-controlled gait training method for Parkinson's disease according to any one of claims 1 to 5, characterized in that: include: The EEG signal acquisition unit, including the EEG detection device, is used to: The patient's EEG signals are collected by EEG detection equipment, and the collected EEG signals are preprocessed and feature extracted. The patient's movement intention is then identified by a computer algorithm, and the movement intention is converted into a control instruction. The converted control instruction is transmitted to the gait training control unit by wireless or wired means; A gait training control unit, including a gait training device, is used for: Receive control instructions transmitted by the EEG signal acquisition unit, and control the gait training device through the control instructions to help the patient perform gait training; Training effectiveness evaluation unit, used for: Use sensors to collect and analyze the patient's gait data to analyze the patient's walking rhythm, stride and stability parameters, and then evaluate the patient's gait; Optimization and improvement unit for: Accumulate patients' EEG signals, gait data, and training effect evaluation data, use artificial intelligence algorithms to process and analyze the accumulated data, identify effective gait patterns through machine learning technology, and then provide training plans based on the patients' gait data.

7. The artificial intelligence-based brain-controlled gait training system for Parkinson's disease according to claim 6, characterized in that: The electroencephalogram signal acquisition unit comprises: Signal acquisition module for: Obtain the patient's neural information through brain wave detection equipment, and then collect brain wave signals; Signal processing modules for: Perform filtering and denoising preprocessing operations on the collected EEG signals, and perform feature extraction on the preprocessed EEG signals; Signal analysis module for: The EEG signals are analyzed and identified through computer algorithms to recognize the patient's movement intention, and the recognized movement intention is converted into control instructions, which are then transmitted to the gait training control unit wirelessly or wiredly.

8. The artificial intelligence-based brain-controlled gait training system for Parkinson's disease according to claim 6, characterized in that: The signal processing module is specifically: Filtering: used to remove noise to improve the quality and analyzability of EEG signals; Denoising: used to eliminate interfering noise that is not related to EEG activity; Feature extraction: used to extract features from EEG signals. The features in EEG signals are divided into time domain features, frequency domain features, and time-frequency domain features.

9. The artificial intelligence-based brain-controlled gait training system for Parkinson's disease according to claim 6, characterized in that: The training effect evaluation unit comprises: Data collection module for: When the patient is undergoing gait training, the patient's gait data is collected through sensors, wherein the gait data includes step length, step speed, and joint angle; Data analysis module for: Analyze gait data, including gait cycle analysis, gait dynamics analysis, and gait biomechanics analysis, to analyze the patient's walking rhythm, stride, and stability; Effect evaluation module, used to: Based on the analysis results of the data analysis module, the patient's gait is evaluated to assess the symmetry, stability and coordination of the patient's gait.

10. The artificial intelligence-based brain-controlled gait training system for Parkinson's disease according to claim 6, characterized in that: The optimization and improvement unit comprises: Data accumulation module for: Accumulate the patient's previous EEG signals, gait data, and training effect evaluation data; Training program modules for: Selecting a machine learning model based on the accumulated data, using the selected model and training it with the accumulated data, identifying effective gait patterns through the trained model, and then providing a training plan based on the patient's gait data; Continuous improvement module for: Execute the operations of the data accumulation module and the training scheme module to update and improve the model according to the changes in the data.

Images