CN-122005276-A - Rehabilitation training method and system based on mixed-paradigm brain-computer interface

CN122005276ACN 122005276 ACN122005276 ACN 122005276ACN-122005276-A

Abstract

The application relates to the technical field of brain-computer interfaces, and provides a rehabilitation training method and system based on a mixed-paradigm brain-computer interface. The method comprises the steps of collecting brain electrical signals of forehead leaves or motor cortex of a patient, carrying out cooperative decoding on the brain electrical signals according to a preset mixed paradigm decoding model, identifying signal characteristics, if the brain electrical signals are forehead leaf brain electrical signals, the signal characteristics are intention intensity scores of imagination walking and movement, if the brain electrical signals are motor cortex brain electrical signals, the signal characteristics are motor imagination types and intention intensity scores of movement, generating lower limb rehabilitation training instructions which are matched with rehabilitation stages of the patient according to the signal characteristics, wherein the training instructions at least comprise any one of walking, knee bending, leg lifting, left leg lifting or right leg lifting, and transmitting the training instructions to an execution module of a rehabilitation robot to drive the rehabilitation robot to finish corresponding training actions.

Inventors

XIE LONGHAN
CHEN YAN
HUANG GUOWEI
LI JUNCHENG
HUANG SHUANGYUAN
LIANG BINGHONG

Assignees

力之智能科技(广州)有限公司

Dates

Publication Date: 20260512
Application Date: 20260413

Claims (10)

1. A rehabilitation training method based on a mixed-paradigm brain-computer interface is characterized by comprising the following steps: collecting brain electrical signals of forehead leaves or sports cortex of a patient; The electroencephalogram signal is cooperatively decoded according to a preset mixed-paradigm decoding model to identify signal characteristics, wherein if the electroencephalogram signal is a forehead lobe electroencephalogram signal, the signal characteristics are the intension intensity scores of whether walking and movement are imagined, and if the electroencephalogram signal is a motor cortex electroencephalogram signal, the signal characteristics are the intension intensity scores of motor imagination type and movement; Generating a lower limb rehabilitation training instruction which is adapted to the rehabilitation stage of the patient according to the signal characteristics, wherein the training instruction at least comprises any one of walking, knee bending, leg lifting, left leg lifting or right leg lifting, transmitting the training instruction to an execution module of the rehabilitation robot, and driving the rehabilitation robot to complete corresponding training actions.
2. The method of claim 1, further comprising, prior to the co-decoding of the electroencephalogram signal according to the preset mixed-paradigm decoding model: and (3) carrying out validity check on the acquired brain electrical signals, and eliminating invalid signals.
3. The method according to claim 2, wherein the performing validity check on the acquired electroencephalogram signals and removing invalid signals includes: Judging whether the electrode of the electroencephalogram acquisition equipment falls off or not through a preset falling-off detection mechanism; judging whether the impedance value of the electroencephalogram signal is in an effective range or not through a preset impedance detection mechanism, judging the electroencephalogram signal with the detected electrode falling off or the impedance value exceeding the effective range as an ineffective signal, and rejecting the ineffective signal.
4. The method of claim 1, wherein the collaborative decoding of the electroencephalogram signal according to a preset mixed-paradigm decoding model to identify signal features comprises: if the electroencephalogram signal is a forehead lobe electroencephalogram signal, judging whether a patient generates a walking motor imagery or not through the mixed paradigm decoding model, and outputting a corresponding motor intention intensity score; If the electroencephalogram signal is a motion cortex electroencephalogram signal, identifying a specific motion type imagined by a patient through the mixed-paradigm decoding model, wherein the motion type comprises at least one of left leg lifting and right leg lifting, and outputting a corresponding motion intention intensity score.
5. The method of claim 4, wherein generating lower limb rehabilitation training instructions adapted to a patient rehabilitation session from the signal characteristics comprises: When judging that a patient generates a motor imagery meeting preset requirements, determining gait parameters corresponding to the motor imagery according to a preset mapping relation according to the motor intention intensity score, wherein the gait parameters comprise at least one of stride, step frequency and step number; generating a training instruction containing the motor imagery type and gait parameters, and generating a static instruction when the motor imagery meeting the preset requirements is not judged to be generated by the patient.
6. The method of claim 1, wherein transmitting the training instructions to an execution module of a rehabilitation robot to drive the rehabilitation robot to perform corresponding training actions comprises: Transmitting the motor imagery type and gait parameters contained in the training instruction to an execution module of the rehabilitation robot, wherein the execution module controls the rehabilitation robot to execute the actions of walking, knee bending, leg lifting, left leg lifting or right leg lifting according to the motor imagery type and the gait parameters, and adjusts the action amplitude, the action frequency and the action times according to the gait parameters.
7. The method of claim 1, wherein the acquiring the brain electrical signals of the forehead lobe or the motor cortex of the patient comprises: The brain electrical signal acquisition is carried out at preset points of forehead leaves or sports cortex by using the double-lug clip anti-interference structure and the magnetic suction female buckle electrode through the full-silica gel flexible brain electrical acquisition equipment, and the preset points are 6-point polar profile acquisition points.
8. The method of claim 5, wherein the method further comprises: And displaying the movement intention strength score, the action type corresponding to the training instruction and the training completion condition through a preset feedback interface so as to realize visual monitoring of the rehabilitation training process.
9. The method of claim 1, further comprising, after the driving the rehabilitation robot completes a corresponding training action: And acquiring a corresponding motion intention quantification result and training completion condition feedback, and dynamically adjusting the fusion proportion of the motor imagery and the attention and the training difficulty information corresponding to the mixed-pattern decoding model according to the motion intention quantification result and the training completion condition feedback to form a personalized rehabilitation training closed loop.
10. A hybrid-paradigm brain-computer interface based rehabilitation training system, characterized by being applied to the method of any one of claims 1-9, comprising: the signal acquisition unit is used for acquiring brain electrical signals of forehead leaves or sports cortex of a patient; The collaborative decoding unit is used for collaborative decoding the electroencephalogram signal according to a preset mixed-paradigm decoding model and identifying signal characteristics, wherein if the electroencephalogram signal is a forehead lobe electroencephalogram signal, the signal characteristics are the intention intensity scores of imagination walking and movement, and if the electroencephalogram signal is a motor cortex electroencephalogram signal, the signal characteristics are the motor imagination type and the intention intensity score of movement; The training completion unit is used for generating lower limb rehabilitation training instructions adapting to the rehabilitation stage of the patient according to the signal characteristics, wherein the training instructions at least comprise any one of walking, knee bending, leg lifting, left leg lifting or right leg lifting, and the training instructions are transmitted to an execution module of the rehabilitation robot to drive the rehabilitation robot to complete corresponding training actions.

Description

Rehabilitation training method and system based on mixed-paradigm brain-computer interface Technical Field The application relates to the technical field of brain-computer interfaces, in particular to a rehabilitation training method and system based on a mixed-paradigm brain-computer interface. Background In the technical scheme of combining the existing brain-computer interface and the lower limb rehabilitation robot, the technical bottleneck of single rehabilitation training paradigm exists generally, most systems only adopt single Motor Imagery (MI) paradigm or attention paradigm, and double requirements of capturing the active motion intention of a patient and evaluating the training concentration degree are difficult to be considered. The scheme of partly claiming "multiple paradigm" also can only select a certain fixed paradigm before training, can't be according to patient's real-time state dynamic adjustment paradigm fusion strategy, leads to rehabilitation training to lack individualized adaptation, is difficult to effectively encourage the patient to take part in initiatively, especially can't satisfy the different demands of stroke hemiplegia patient's different rehabilitation stages to training intensity and type. For example, the conventional single paradigm cannot quantify the intensity of exercise intent and concentration of attention at the same time, resulting in a stiff training instruction generation mechanism, which is difficult to accurately reflect the actual rehabilitation progress of the patient. Accordingly, a need exists for a method that addresses at least one of the problems described above. Disclosure of Invention The application provides a rehabilitation training method and system based on a mixed-paradigm brain-computer interface, and aims to solve the problem. In a first aspect, an embodiment of the present application provides a rehabilitation training method based on a mixed-paradigm brain-computer interface, where the method includes: collecting brain electrical signals of forehead leaves or sports cortex of a patient; The electroencephalogram signal is cooperatively decoded according to a preset mixed-paradigm decoding model to identify signal characteristics, wherein if the electroencephalogram signal is a forehead lobe electroencephalogram signal, the signal characteristics are the intension intensity scores of whether walking and movement are imagined, and if the electroencephalogram signal is a motor cortex electroencephalogram signal, the signal characteristics are the intension intensity scores of motor imagination type and movement; Generating a lower limb rehabilitation training instruction which is adapted to the rehabilitation stage of the patient according to the signal characteristics, wherein the training instruction at least comprises any one of walking, knee bending, leg lifting, left leg lifting or right leg lifting, transmitting the training instruction to an execution module of the rehabilitation robot, and driving the rehabilitation robot to complete corresponding training actions. In some embodiments, before the collaborative decoding of the electroencephalogram signals according to the preset mixed-paradigm decoding model, the method further comprises the steps of performing validity check on the acquired electroencephalogram signals and eliminating invalid signals. In some embodiments, the verifying the validity of the collected electroencephalogram signals and eliminating invalid signals comprises judging whether electrodes of the electroencephalogram collection device fall off or not through a preset fall-off detection mechanism, judging whether the impedance value of the electroencephalogram signals is in an effective range or not through a preset impedance detection mechanism, and judging the electroencephalogram signals with detected electrode fall-off or impedance values exceeding the effective range as invalid signals and eliminating the invalid signals. In some embodiments, the collaborative decoding of the electroencephalogram according to the preset mixed-paradigm decoding model and the identifying of signal characteristics include judging whether a patient generates a motor imagery of walking and outputting a corresponding motor intention intensity score if the electroencephalogram is a forehead lobe electroencephalogram, and identifying a specific motor type of the patient imagery by the mixed-paradigm decoding model if the electroencephalogram is a motor cortex electroencephalogram, wherein the motor type includes at least one of lifting a left leg and lifting a right leg and outputting a corresponding motor intention intensity score. In some embodiments, the generating lower limb rehabilitation training instructions adapted to the rehabilitation stage of the patient according to the signal features comprises determining gait parameters corresponding to the motor imagery according to a preset mapping relation according to the motor inte