Epidural electrical stimulation lower limb rehabilitation system and method based on brain-computer interface

Abstract

The invention relates to the technical field of rehabilitation medical treatment and neural engineering, and particularly discloses an epidural electrical stimulation lower limb rehabilitation system and method based on a brain-computer interface, the epidural electrical stimulation lower limb rehabilitation system comprises a weight reduction suspension device, an epidural electrical stimulation module and a brain-computer interface module, the epidural electrical stimulation module controls left and right electrical stimulation through a program controller, and the brain-computer interface module comprises a wireless electroencephalogram acquisition device and an interaction terminal and is used for acquiring electroencephalogram signals of a movement area of a patient and presenting a virtual gait scene and voice prompt at a computer end. The method comprises the steps of electroencephalogram signal collection and feature extraction, motion intention judgment, stimulation control and feedback backtracking, and by combining a motion intention decoding result with electrical stimulation output, activation of a lower limb target muscle group is achieved, a training closed loop is constructed, and the method is used for postoperative rehabilitation of a spinal cord injury patient.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of rehabilitation medicine and neural engineering, and more particularly, to an epidural electrical stimulation lower limb rehabilitation system and method based on a brain-computer interface. BACKGROUND

[0002] Spinal cord injury (SCI) is a disease that seriously threatens human health. Globally, there are about 900,000 new SCI patients each year, and the total number of existing patients has exceeded 20 million. SCI often causes permanent damage to the neural conduction pathway, leading to various functional disorders, of which motor dysfunction is the most common and has a significant impact on the quality of life of patients. For patients with paraplegia caused by SCI, the recovery of lower limb motor ability is the main goal of rehabilitation treatment.

[0003] In existing treatment methods, neuromodulation therapy combining spinal cord electrical stimulation and rehabilitation training has been proven to promote motor function recovery, and epidural electrical stimulation has shown certain efficacy in improving lower limb muscle strength in patients. However, this therapy still relies on a large amount of long-term leg lifting, stepping, and walking training after surgery, and the rehabilitation process is long and inefficient, with limited efficacy improvement.

[0004] In recent years, research on brain-computer interface (BCI) technology has gradually deepened. This technology can collect scalp electroencephalogram or deep brain electrode signals and decode the motor intention of patients, and then generate external control instructions. BCI can promote neural system plasticity through feedback mechanisms during long-term use, and has potential value for motor function recovery. Current research has applied BCI to the fields of rehabilitation training and external device control, but existing technology has not combined non-invasive electroencephalogram-based BCI with epidural electrical stimulation for postoperative lower limb motor rehabilitation training in SCI patients, thus there is still a technical gap.

[0005] In view of this, the present application provides an epidural electrical stimulation lower limb rehabilitation system and method based on a brain-computer interface to solve the above problems. SUMMARY

[0006] To overcome the problems in the prior art, the present application provides an epidural electrical stimulation lower limb rehabilitation system and method based on a brain-computer interface, aiming to solve the problem of low training efficiency and limited efficacy due to the lack of brain-computer interface assistance during postoperative lower limb rehabilitation training in SCI patients.

[0007] In a first aspect, the present application provides an epidural electrical stimulation lower limb rehabilitation method based on a brain-computer interface, comprising the following steps:

[0008] S101: Collect the electroencephalogram signal of the scalp EEG cap worn by the patient, and perform frequency domain decomposition on the electroencephalogram signal according to the movement attempt paradigm to extract a lower limb movement intention feature vector;

[0009] S102: Nonlinearly match the lower limb movement intention feature vector with a preset gait training instruction library to obtain a movement intention determination result; and when the determination result is a valid movement intention, record the real-time movement parameters of the patient's gait training;

[0010] S103: Control the epidural electrical stimulation program controller based on the movement intention determination result, and generate feedback interaction data in combination with the real-time movement parameters while the program controller outputs a spinal cord stimulation signal;

[0011] S104: Dynamically update the gait scene in the virtual training interface according to the feedback interaction data, and use voice prompts to guide the patient's movement attempt synchronously; wherein when the movement intention determination result is inconsistent with the feedback interaction data, backtrack the lower limb movement intention feature vector to perform nonlinear matching again.

[0012] As a preferred technical solution of the first aspect of the application, the acquisition logic of the lower limb movement intention feature vector is:

[0013] During the gait training process of the patient in the movement attempt paradigm, the scalp EEG acquisition device is used to acquire the electroencephalogram signal of the brain motor area, and the electroencephalogram signal is preprocessed through a band-pass filter and an artifact detection algorithm;

[0014] The preprocessed electroencephalogram signal is subjected to frequency domain decomposition to obtain signal components of multiple frequency bands;

[0015] The power spectral density is calculated as an energy feature on the target frequency band, the coherence value between different electrode channels is calculated as a coherence feature, and the sample entropy or approximate entropy is used to quantify the complexity of the target frequency band signal as a nonlinear complexity feature;

[0016] The energy feature, coherence feature and nonlinear complexity feature are combined to form a lower limb movement intention feature vector.

[0017] As a preferred technical solution of the first aspect of the application, the determination logic of the movement intention determination result is:

[0018] The lower limb movement intention feature vector is input into a nonlinear discrimination model to obtain a preliminary determination result of the movement intention; the nonlinear discrimination model adopts a recurrent neural network structure, and outputs the movement intention category and the corresponding confidence value according to the input feature vector;

[0019] And combined with the gait kinematics parameters collected during the training process, the gait kinematics parameters include hip joint angle, knee joint angle and plantar pressure signal, the gait parameters are used as auxiliary determination basis;

[0020] When the output confidence of the nonlinear discrimination model is lower than a preset threshold, the preliminary determination result is secondarily corrected by using the gait parameters combined with an adaptive Bayesian updating mechanism, so that a corrected motion intention category is obtained;

[0021] The corrected motion intention category is used as the final motion intention determination result and is provided for subsequent electric stimulation control steps.

[0022] As a preferred technical solution of the first aspect of the application, the feedback interaction data acquisition logic is:

[0023] According to the motion intention determination result, a corresponding epidural electric stimulation parameter combination is searched in a preset gait training instruction library, and during the epidural electric stimulation output process, the electromyographic signals of the target muscle groups of the lower limbs of the patient and the gait kinematics parameters are synchronously collected;

[0024] The electromyographic signals are band-pass filtered to remove power frequency interference and baseline drift, the kinematics parameters are interpolated and normalized, and the electromyographic signals are time-frequency analyzed by using short-time Fourier transform, so that the instantaneous energy features in the stimulation response window are extracted;

[0025] The instantaneous energy features and the gait kinematics parameters are aligned according to the time stamp, so that a feedback interaction data set containing the electromyographic energy change, gait phase marker and stimulation triggering time is formed;

[0026] According to the feedback interaction data set, the stimulation timing is corrected and output to the epidural electrode through the program control instrument, so that the contraction of the corresponding muscle groups of the lower limbs is triggered.

[0027] As a preferred technical solution of the first aspect of the application, the feedback interaction and backtracking logic of the feedback interaction data is:

[0028] The category of the motion intention determination result and the triggering timing are compared with the electromyographic instantaneous energy features and gait phase features obtained by time-frequency analysis in the feedback interaction data for consistency;

[0029] During the comparison process, if the amplitude, phase or time delay of the electromyographic instantaneous energy deviates from the expected value corresponding to the motion intention determination result by more than a preset threshold, it is determined that there is deviation;

[0030] When the deviation is detected, a backtracking mechanism is triggered to re-input the feedback interaction data into the motion intention determination step, correct the determination result, and adjust the electrical stimulation parameters, including stimulation channels, pulse width, frequency, amplitude, or phase delay, according to the corrected result;

[0031] The consistency comparison result and the backtracking adjustment result are visually presented through the training interface, and real-time feedback is provided to the patient in the form of voice prompts.

[0032] In a second aspect, the application provides an epidural electrical stimulation lower limb rehabilitation training system based on a brain-computer interface, which is an implementation of the brain-computer interface-based epidural electrical stimulation lower limb rehabilitation method according to any one of claims 1-5, based on the implementation of the first aspect, comprising:

[0033] A body weight support system is used to provide a gait training environment for the patient, which can adopt a sky rail suspension device or a body weight support frame to reduce the weight-bearing pressure on the lower limbs of the patient;

[0034] An epidural electrical stimulation implant module includes an epidural electrode and a programmable instrument, which can be operated by a doctor or a patient to control the emission of left and right electrical stimulation signals;

[0035] A brain-computer interface system includes a wireless scalp electroencephalogram acquisition device and an interactive terminal. The electroencephalogram acquisition device is used to acquire motor area electroencephalogram signals when the patient performs lower limb lifting, stepping, and other movement attempts. The interactive terminal provides a virtualized gait training scene in a first-person perspective on a computer interface, and combines picture and voice prompts to guide and feedback to the patient for immersive training interaction.

[0036] As a preferred technical solution of the second aspect of the application, the brain-computer interface module comprises:

[0037] A wireless scalp electroencephalogram cap is used to acquire motor area electroencephalogram signals;

[0038] A signal processing unit is used to perform frequency domain decomposition and feature extraction on the electroencephalogram signals;

[0039] A determination unit is used to determine and correct the motion intention in combination with gait parameters.

[0040] As a preferred technical solution of the second aspect of the application, the feedback interaction module is configured with a time-frequency analysis unit for calculating the instantaneous energy features of the electromyogram signals and aligning and comparing them with the gait phase.

[0041] As a preferred technical solution of the second aspect of the application, the interface interaction includes two forms of a virtual gait training scene and a voice prompt, wherein the virtual scene presents the patient's gait action in a first-person perspective, and the voice prompt is used to guide the patient to perform training actions such as hip flexion, knee extension, or stepping.

[0042] As a preferred technical solution of the second aspect of the application, the program-controlled instrument has a double-end control function, and can be used by the doctor end to preset stimulation parameters, and can also be used by the patient end to call in real time during the training process, so as to realize individualized electrical stimulation adjustment.

[0043] The specific advantages of the application are as follows:

[0044] 1. Improve the accuracy of rehabilitation training: through the brain-computer interface to decode the patient's lower limb movement intention in real time, and link with the epidural electrical stimulation output, so that the electrical stimulation can act on the target nerve area at the right time, and improve the accuracy of lower limb movement training.

[0045] 2. Enhance the efficacy of rehabilitation training: the brain-computer interface provides multi-modal feedback such as vision and hearing, promotes the positive cycle between the patient's movement intention and neural stimulation, accelerates neural function remodeling, and improves rehabilitation efficiency.

[0046] 3. Shorten the rehabilitation period: the application can reduce the dependence on a large number of repetitive gait training in traditional rehabilitation training, shorten the rehabilitation process, and reduce the burden on patients.

[0047] 4. Improve patient participation and compliance: virtual training scenes and real-time feedback enhance patient immersion and initiative, and improve the interest and persistence of the training process.

[0048] 5. Good clinical application prospect: the system and method provide a new type of neuroengineering means for postoperative lower limb rehabilitation training of patients with spinal cord injury, and have strong popularization value and clinical application potential. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 The flowchart of the epidural electrical stimulation lower limb rehabilitation method of the application.

[0050] Figure 2 The framework diagram of the epidural electrical stimulation lower limb rehabilitation training system of the application. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only a part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0052] Embodiment 1

[0053] As Figure 1 shown, the embodiment provides a technical solution: an epidural electrical stimulation-based lower limb rehabilitation method based on a scalp brain-computer interface system with a movement attempt paradigm. The scalp brain-computer interface system is combined with post-implantation training of epidural electrical stimulation. Through the combination of real-time feedback of the brain-computer interface system and output of the epidural electrical stimulation, the accuracy and efficacy of the paralyzed patient in the gait training process are improved, and the neural function reconstruction of "brain-spinal cord-lower limb muscle" is promoted. The method comprises the following steps:

[0054] S101: Collecting the electroencephalogram signal of the scalp brain cap worn by the patient, and performing frequency domain decomposition on the electroencephalogram signal according to the movement attempt paradigm to extract a lower limb movement intention feature vector;

[0055] Specifically, the acquisition logic of the lower limb movement intention feature vector is as follows:

[0056] During the gait training process of the patient with the movement attempt paradigm, the electroencephalogram signal of the brain motor area is acquired by using a scalp electroencephalogram acquisition device, and the electroencephalogram signal is preprocessed by using a band-pass filter and an artifact detection algorithm;

[0057] The preprocessed electroencephalogram signal is subjected to frequency domain decomposition to obtain signal components of multiple frequency bands;

[0058] The power spectral density is calculated on the target frequency band as an energy feature, the coherence value between different electrode channels is calculated as a coherence feature, and the sample entropy or approximate entropy is used to quantify the complexity of the target frequency band signal as a nonlinear complexity feature;

[0059] The energy feature, the coherence feature and the nonlinear complexity feature are combined to form a lower limb movement intention feature vector.

[0060] S102: Nonlinearly matching the lower limb movement intention feature vector with a preset gait training instruction library to obtain a movement intention determination result; and recording real-time movement parameters of the gait training of the patient when the determination result is a valid movement intention;

[0061] Specifically, the determination logic of the movement intention determination result is as follows:

[0062] The lower limb movement intention feature vector is input into a nonlinear discrimination model to obtain a preliminary determination result of the movement intention; the nonlinear discrimination model adopts a recurrent neural network structure, and outputs a movement intention category and a corresponding confidence value according to the input feature vector;

[0063] And combine the gait kinematics parameters collected during the training process, the gait kinematics parameters include hip joint angle, knee joint angle and plantar pressure signal, the gait parameters are used as auxiliary determination basis;

[0064] When the output confidence of the nonlinear discrimination model is lower than the preset threshold, the preliminary determination result is secondarily corrected by using the gait parameters combined with the adaptive Bayesian updating mechanism, and the corrected motion intention category is obtained;

[0065] The corrected motion intention category is used as the final motion intention determination result, and is provided for the subsequent electric stimulation control step.

[0066] S103: Based on the motion intention determination result, an epidural electric stimulation program controller is controlled, and feedback interaction data is generated by combining the real-time motion parameters while the program controller outputs the spinal cord stimulation signal;

[0067] Specifically, the feedback interaction data acquisition logic is:

[0068] According to the motion intention determination result, the corresponding epidural electric stimulation parameter combination is searched in the preset gait training instruction library, and the electromyographic signals of the target muscle groups of the lower limbs of the patient and the gait kinematics parameters are synchronously collected during the epidural electric stimulation output process;

[0069] The electromyographic signals are band-pass filtered to remove power frequency interference and baseline drift, the kinematics parameters are interpolated and normalized, the electromyographic signals are time-frequency analyzed by using short-time Fourier transform, and the instantaneous energy features in the stimulation response window are extracted;

[0070] The instantaneous energy features and the gait kinematics parameters are aligned according to the time stamp to form a feedback interaction data set containing electromyographic energy changes, gait phase markers and stimulation triggering time;

[0071] According to the feedback interaction data set, the stimulation timing is corrected and output to the epidural electrode through the program controller, so as to trigger the contraction of the corresponding muscle groups of the lower limbs.

[0072] S104: The gait scene is dynamically updated in the virtual training interface according to the feedback interaction data, and the patient's movement attempt is guided synchronously by using voice prompts; when the motion intention determination result is inconsistent with the feedback interaction data, the lower limb motion intention feature vector is traced back for nonlinear matching.

[0073] Specifically, the feedback interaction and backtracking logic of the feedback interaction data is:

[0074] The category of the motion intention determination result is consistent with the time-frequency analysis of the electromyography transient energy feature and the gait phase feature in the feedback interaction data, and the trigger timing is consistent with the feedback interaction data;

[0075] In the comparison process, if the amplitude, phase or time delay of the electromyography transient energy deviates from the expected value corresponding to the motion intention determination result by more than a preset threshold, it is determined that there is a deviation;

[0076] When the deviation is detected, a backtracking mechanism is triggered, the feedback interaction data is re-input into the motion intention determination step, the determination result is corrected, and the electric stimulation parameters, including stimulation channel, pulse width, frequency, amplitude or phase delay, are adjusted according to the corrected result;

[0077] The consistency comparison result and the backtracking adjustment result are visually presented through a training interface, and real-time feedback is provided to the patient in the form of voice prompts.

[0078] The patient wears an electroencephalogram cap to perform gait training content such as leg lifting and stepping under suspension, and follows the voice prompts on the brain-computer interface computer interface. If the patient successfully attempts unilateral lower limb movement, the patient receives lower limb muscle movement feedback by operating an epidural electric stimulation program control instrument to deliver spinal cord electric stimulation signals to induce contraction of the corresponding muscle groups in the lower limbs. At the same time, the computer interface provides voice and picture feedback to the patient, prompting the successful movement attempt. If the imagination is not successful, no spinal cord electric stimulation is delivered, and voice and picture feedback are provided to encourage the patient to try again.

[0079] Embodiment 2

[0080] As shown in Figure 2 the embodiment, the present embodiment provides an epidural electric stimulation lower limb rehabilitation training system based on a brain-computer interface, which comprises

[0081] A weight-reducing suspension system is used to provide a gait training environment for the patient. The weight-reducing suspension system can adopt a ceiling suspension device or a weight-reducing suspension frame to reduce the weight-bearing pressure on the patient's lower limbs.

[0082] An epidural electric stimulation implant module comprises an epidural electrode and a program control instrument. The program control instrument can be operated by a doctor or a patient to control the delivery of left and right electric stimulation signals.

[0083] A brain-computer interface system comprises a wireless scalp electroencephalogram acquisition device and an interaction terminal. The electroencephalogram acquisition device is used to acquire motor area electroencephalogram signals when the patient attempts to move the lower limbs. The interaction terminal provides a virtual gait training scene in a first-person perspective in the computer interface, and provides training guidance and feedback to the patient in combination with picture and voice prompts to achieve immersive training interaction.

[0084] Further explanation, epidural electrical stimulation restores limb motor function by activating the remaining spinal nerves, but after implantation, a large amount of long-term suspension under the leg, step and walking training is still needed to achieve true walking. In this process, the paralyzed person needs to constantly try to move the lower limbs, and how to stimulate the correct position of the spinal cord at the right time to achieve precise control of the leg muscles is crucial to accelerate the clinical efficacy after epidural electrical stimulation implantation. The brain-computer interface system based on motor imagery brain signal features applies conditioned reflex and brain plasticity, and through neural feedback, the user's brain activity features, decoding results, and the results of communication or control with external devices are visualized to the user in the form of visual, auditory or tactile feedback, to adjust the user's mental activity and brain signals, and ultimately improve the performance of brain-computer interaction.

[0085] Specifically, after epidural electrical stimulation implantation, the patient is placed under the overhead suspension device or the weight-reducing suspension frame for gait training; during the training process, the patient wears a non-invasive electroencephalogram cap to try to lift the lower limbs and extend the knees, and the virtual scene on the computer screen of the brain-computer interface provides the corresponding lower limb movement scene for the patient, and the patient is guided to perform left and right hip flexion, knee extension or walking training through voice guidance; at the same time, the epidural electrical stimulation serves as an external device of the brain-computer interface, and the synchronous electrical stimulation signal emission serves as the feedback of the movement attempt, and the brain-computer interface system decodes the collected signals to obtain the movement intention information of the patient.

[0086] The brain-computer interface system presents a virtual gait training scene in the first person perspective on the computer side, and guides the patient to perform left and right hip flexion, knee extension or walking training through voice prompts; when the patient's movement intention is correct, the system controls the program-controlled instrument to send corresponding spinal cord stimulation signals to the epidural electrical stimulation implantation device to induce the contraction of the target muscle group of the lower limbs as positive feedback of the movement attempt; when the movement intention is determined to be incorrect, the electrical stimulation is not triggered, and the interface displays corresponding prompts to prompt the patient to try again.

[0087] Through the above interaction mode, the movement intention decoding of the brain-computer interface and the real-time linkage of the epidural electrical stimulation are realized, the brain-computer interface system provides virtual scene and voice prompt real-time feedback for the brain movement intention, and according to the correct / incorrect determination of the patient's lower limb movement intention, the epidural electrical stimulator is controlled to send spinal cord positive feedback stimulation signals, which can assist the epidural electrical stimulation treatment to more accurately position the correct lower limb muscle group corresponding to the nerve control, improve the success rate of the treatment, shorten the patient's rehabilitation process, and thus assist the patient to complete the precise lower limb motor rehabilitation training.

[0088] The above merely describes specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any modification or substitution within the technical range disclosed by the present application can be easily thought of by those skilled in the art, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0089] Finally, the above merely describes preferred embodiments of the present application, and is not used to limit the present application, and any modification, equivalent substitution, improvement, etc. within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. An epidural electrical stimulation lower limb rehabilitation method based on a brain-computer interface, characterized in that, Comprise: S101: Collect the electroencephalogram signal of the scalp EEG cap worn by the patient, and perform frequency domain decomposition on the electroencephalogram signal according to the movement attempt paradigm to extract the lower limb movement intention feature vector; S102: Nonlinearly match the lower limb movement intention feature vector with the preset gait training instruction library to obtain a movement intention determination result; And when the determination result is an effective movement intention, record the real-time movement parameters of the patient's gait training; S103: Control the epidural electrical stimulation program controller based on the movement intention determination result, and generate feedback interaction data in combination with the real-time movement parameters while the program controller outputs the spinal cord stimulation signal; S104: Dynamically update the gait scene in the virtual training interface according to the feedback interaction data, and guide the patient to move by voice prompt; When the movement intention determination result is inconsistent with the feedback interaction data, backtrack the lower limb movement intention feature vector and re-nonlinearly match.

2. The epidural electrical stimulation lower limb rehabilitation method based on brain-computer interface according to claim 1, characterized in that, The acquisition logic of the lower limb movement intention feature vector: During the gait training process of the patient in the movement attempt paradigm, the scalp EEG acquisition device is used to acquire the electroencephalogram signal of the brain motor area, and the electroencephalogram signal is preprocessed by band-pass filtering and artifact detection algorithm; The preprocessed electroencephalogram signal is subjected to frequency domain decomposition to obtain signal components of multiple frequency bands; The power spectral density is calculated on the target frequency band as the energy feature, the coherence value between different electrode channels is calculated as the coherence feature, and the sample entropy or approximate entropy is used to quantify the complexity of the target frequency band signal as the nonlinear complexity feature; The energy feature, coherence feature and nonlinear complexity feature are combined to form the lower limb movement intention feature vector.

3. The epidural electrical stimulation lower limb rehabilitation method based on brain-computer interface according to claim 2, characterized in that, The determination logic of the movement intention determination result is: Input the lower limb movement intention feature vector into a nonlinear discrimination model to obtain a preliminary determination result of the movement intention; The nonlinear discrimination model adopts a recurrent neural network structure, and outputs the movement intention category and the corresponding confidence value according to the input feature vector; And combine the gait kinematics parameters collected during the training process, the gait kinematics parameters include hip joint angle, knee joint angle and foot pressure signal, and use the gait parameters as auxiliary determination basis; When the output confidence of the nonlinear discrimination model is lower than a preset threshold, use the gait parameters in combination with an adaptive Bayesian update mechanism to make a secondary correction to the preliminary determination result to obtain a corrected movement intention category; The corrected movement intention category is used as the final movement intention determination result, which is provided for subsequent electrical stimulation control steps.

4. The epidural electrical stimulation lower limb rehabilitation method based on brain-computer interface according to claim 3, characterized in that, The acquisition logic of the feedback interaction data: According to the movement intention determination result, search for the corresponding epidural electrical stimulation parameter combination in the preset gait training instruction library, and synchronously collect the electromyogram of the target muscle group of the patient's lower limb and the gait kinematics parameters during the output process of the epidural electrical stimulation; The electromyogram is subjected to band-pass filtering to remove power frequency interference and baseline drift, the kinematics parameters are subjected to interpolation and normalization processing, and the short-time Fourier transform is used to perform time-frequency analysis on the electromyogram to extract the instantaneous energy feature in the stimulation response window; The transient energy features are aligned with the gait kinematic parameters according to timestamps to form a feedback interaction dataset containing the changes in myoelectric energy, gait phase markers, and stimulation trigger time; The stimulation timing is corrected according to the feedback interaction dataset and output to the epidural electrode through the program controller, thereby triggering the contraction of the corresponding muscle groups in the lower limbs.

5. The epidural electrical stimulation lower limb rehabilitation method based on brain-computer interface according to claim 4, characterized in that, The feedback interaction of the feedback interaction data and the backtracking logic are: The consistency of the category of the movement intention determination result and the trigger timing with the myoelectric transient energy features and gait phase features obtained through time-frequency analysis in the feedback interaction data is compared; During the comparison process, if the amplitude, phase, or time delay of the myoelectric transient energy deviates from the expected value corresponding to the movement intention determination result by more than a preset threshold, it is determined that there is a deviation; When the deviation is detected, the backtracking mechanism is triggered, the feedback interaction data is re-input into the movement intention determination step, the determination result is corrected, and the electrical stimulation parameters, including the stimulation channel, pulse width, frequency, amplitude, or phase delay, are adjusted according to the corrected result; The consistency comparison result and the backtracking adjustment result are visually presented through the training interface, and real-time feedback is provided to the patient in the form of voice prompts.

6. An epidural electrical stimulation lower limb rehabilitation training system based on brain-computer interface, based on the implementation of the brain-computer interface-based epidural electrical stimulation lower limb rehabilitation method of any one of claims 1-5, characterized in that, It includes: A body weight support system: used to provide a gait training environment for patients, which can use a ceiling suspension device or a body weight support frame to reduce the weight-bearing pressure on the lower limbs of patients; An epidural electrical stimulation implant module: including an epidural electrode and a program controller, which can be operated by the doctor side or the patient side to control the emission of left and right electrical stimulation signals; A brain-computer interface system: including a wireless scalp EEG acquisition device and an interaction terminal, the EEG acquisition device is used to acquire motor area EEG signals when the patient performs lower limb lifting, stepping, and other movement attempts; The interaction terminal provides a virtual gait training scene in a first-person perspective in the computer interface, and combines picture and voice prompts to guide and feedback to the patient, to achieve immersive training interaction.

7. The epidural electrical stimulation lower limb rehabilitation training system based on brain-computer interface according to claim 6, characterized in that, The brain-computer interface module includes: A wireless scalp EEG cap for acquiring motor area EEG signals; A signal processing unit for frequency domain decomposition and feature extraction of EEG signals; A determination unit for determining and correcting movement intention in combination with gait parameters.

8. The epidural electrical stimulation lower limb rehabilitation training system based on brain-computer interface according to claim 6, characterized in that, The feedback interaction module is configured with a time-frequency analysis unit for calculating the transient energy features of myoelectric signals and aligning and comparing with gait phases.

9. The epidural electrical stimulation lower limb rehabilitation training system based on brain-computer interface according to claim 6, characterized in that, The interface interaction includes virtual gait training scenes and voice prompts, where the virtual scene presents the patient's gait movements in a first-person perspective, and the voice prompts are used to guide the patient to perform hip flexion, knee extension, or stepping training movements.

10. The epidural electrical stimulation lower limb rehabilitation training system based on brain-computer interface according to claim 6, characterized in that, The program controller has a double-end control function, which can be used by the doctor side to preset stimulation parameters, or by the patient side to call in real time during the training process, to achieve individualized electrical stimulation adjustment.

Images