WO2017101621A1 - 闭环的脑控功能性电刺激系统 - Google Patents

闭环的脑控功能性电刺激系统 Download PDF

Info

Publication number
WO2017101621A1
WO2017101621A1 PCT/CN2016/105385 CN2016105385W WO2017101621A1 WO 2017101621 A1 WO2017101621 A1 WO 2017101621A1 CN 2016105385 W CN2016105385 W CN 2016105385W WO 2017101621 A1 WO2017101621 A1 WO 2017101621A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
module
matrix
brain
patient
Prior art date
Application number
PCT/CN2016/105385
Other languages
English (en)
French (fr)
Inventor
李景振
聂泽东
刘宇航
王磊
Original Assignee
深圳先进技术研究院
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳先进技术研究院 filed Critical 深圳先进技术研究院
Publication of WO2017101621A1 publication Critical patent/WO2017101621A1/zh

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
    • A61N1/00—Electrotherapy; Circuits therefor
    • A61N1/18—Applying electric currents by contact electrodes
    • A61N1/32—Applying electric currents by contact electrodes alternating or intermittent currents
    • A61N1/36—Applying electric currents by contact electrodes alternating or intermittent currents for stimulation
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
    • A61N1/00—Electrotherapy; Circuits therefor
    • A61N1/18—Applying electric currents by contact electrodes
    • A61N1/32—Applying electric currents by contact electrodes alternating or intermittent currents
    • A61N1/36—Applying electric currents by contact electrodes alternating or intermittent currents for stimulation
    • A61N1/36003—Applying electric currents by contact electrodes alternating or intermittent currents for stimulation of motor muscles, e.g. for walking assistance
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/1118—Determining activity level
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/369—Electroencephalography [EEG]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7271—Specific aspects of physiological measurement analysis

Definitions

  • the invention belongs to the field of medical instruments, and in particular relates to a closed-loop brain-controlled functional electrical stimulation system.
  • the principle of traditional functional electrical stimulation is to use a certain intensity of low-frequency pulse current to stimulate one or more groups of muscles through a preset program, induce muscle movement or simulate normal autonomous movement to achieve improvement or recovery of stimulation.
  • the purpose of muscle or muscle function is a passive method of rehabilitation.
  • the invention provides a closed-loop brain-controlled functional electrical stimulation system, the system comprising: a brain motion cortical region signal acquisition module, an information control module, a stimulator, a limb motion acquisition module, and a brain frontal signal acquisition module;
  • the information control module includes: a signal preprocessing sub-module, a pattern recognition sub-module, and a command control sub-module Piece;
  • a signal pre-processing sub-module for filtering and waveform shaping the signal in the motor cortex region and transmitting the signal to the pattern recognition sub-module
  • a pattern recognition sub-module configured to identify a signal by using an initial mapping matrix M 0 and a tunable mapping matrix M 1 to determine a motion intention of the patient;
  • a command control submodule configured to determine a weight of the influence factor according to the BP neural network algorithm, and use a genetic algorithm to optimize the weight of the influence factor, generate a stimulus command corresponding to the factor, and transmit the stimulus command to the stimulator;
  • a limb motion acquisition module configured to collect data of a limb motion state, and transmit data of the limb motion state to the information control module;
  • the brain frontal lobe signal acquisition module is configured to collect data of the frontal lobe signal of the patient's brain and transmit the data to the information control module;
  • the information control module is further configured to evaluate the current stimulation effect according to the feedback data of the limb motion state and the data of the brain frontal lobe signal, and automatically correct the adjustable mapping matrix M 1 of the pattern recognition submodule according to the evaluation result.
  • the signal pre-processing sub-module is specifically configured to: after receiving the signal of the brain motion cortex region, first perform low-pass filtering on the signal of the brain motion cortex region, and then perform waveform shaping on the low-pass filtered signal. Obtaining characteristics of the signal, the characteristics of the signal including: amplitude of the signal, pulse width of the signal, number of peaks of the signal, rise time of the signal, time interval of the signal, and transmission of characteristics of the signal to the pattern identifier Module.
  • the pattern recognition submodule is specifically configured to:
  • the g mn subscript m ranges from 1 to 8, corresponding to 8 different limb movement directions: upper, lower, left, right, upper left, upper right, lower left, lower right; g mn subscript n range is the same From 1 to 8, corresponding to 8 different limb exercise intensities, from level 1 to level 8;
  • command control submodule is specifically configured to:
  • K ⁇ k 1 , k 2 , k 3 , k 4 , k 5 , k 6 , k 7 , k 8 , k 9 , k 10 , k 11 , k 12 , k 13 , k 14 , k 15 , k 16 , k 17 , k 18 , k 19 , k 20 ⁇ , where k 1 is the result of the pattern recognition sub-module determination, k 2 is the male paralysis patient, k 3 is the female paralysis patient, k 4 For infants with paralysis, k 5 is a young paralyzed patient, k 6 is a middle-aged paralysis patient, k 7 is an elderly paralyzed patient, k 8 is the left upper arm paralysis, k 9 is the left forearm paralysis, k 10 is the right upper ankle paralysis, k 10 is the right upper ankle paralysis, k 11 For the right forearm, k 12 is
  • the BP neural network algorithm is used to determine the weight of the input influence factor, including: gender, age, sputum location, sputum duration, patient's willingness to exercise;
  • the genetic algorithm is used to optimize the weight of the input influence factor
  • Y 2 ⁇ y 21 , y 22 , y 23 , y 24 , y 25 ⁇
  • Y 4 ⁇ y 41 ,y 42 ,y 43 ,y 44 ,y 45 ⁇
  • Y 1 is the waveform of the signal
  • y11 is the single-phase waveform
  • y12 is the two-phase charge balance waveform
  • y13 is the two-phase charge imbalance waveform
  • y14 is the two-phase charge delay balance waveform
  • y15 is the two-phase charge delay imbalance waveform
  • 2 is the amplitude of the current corresponding to the Y 1 waveform
  • Y 3 is the pulse width of the Y 1 waveform signal
  • Y 4 is the frequency of the Y 1 waveform signal
  • w jk is the weight value of the corresponding element in the input influence factor set
  • Stimulus command Y K*S ij .
  • the stimulator is configured to generate a corresponding stimulation current according to the stimulation command.
  • the information control module is specifically configured to automatically adjust the adjustable mapping matrix M 1 according to the relationship between the deviation of the limb motion and the satisfaction degree of the patient to the stimulation effect; specifically:
  • u 5 indicates that the direction angle deviation is less than 70°
  • u 6 indicates that the direction angle deviation is less than 90°
  • u 7 indicates that the direction angle deviation is less than 150°
  • u 8 indicates that the direction angle deviation is greater than 150°
  • u 9 indicates the exercise intensity.
  • the level deviation is 0;
  • u 10 indicates that the exercise intensity level deviation is 1,
  • u 11 indicates that the exercise intensity level deviation is 2,
  • u 12 indicates the exercise intensity level deviation is 3
  • u 13 indicates the exercise intensity level deviation is 4, and
  • u 14 indicates exercise intensity
  • the grade deviation is 5, and u 15 indicates that the exercise intensity level deviation is 6 , u 16 indicates that the exercise intensity level deviation is 6 or more;
  • the output evaluation factor is mainly the satisfaction degree of the patient's stimulation effect on the brain-controlled functional electrical stimulation system.
  • the data is mainly derived from the signal obtained by the brain frontal signal acquisition module. By filtering and Fourier transforming the signal, the signal characteristics are extracted to judge the patient's satisfaction with the stimulation effect.
  • mapping matrix derived from f is the mapping matrix M 1 in the pattern recognition sub-module.
  • the technical solution provided by the present invention proposes a closed-loop brain-controlled functional electrical stimulation system, which not only enables the paralyzed patient to complete the corresponding action according to his own wishes, but also the system can also stimulate the effect.
  • the evaluation is performed, and the mapping matrix of the information control module is automatically adjusted to ensure the accuracy of the limb motion recognition and improve the rehabilitation effect of the paralyzed patient.
  • FIG. 1 is a structural diagram of a closed-loop brain-controlled functional electrical stimulation system according to a first preferred embodiment of the present invention
  • FIG. 2 is a schematic diagram of a closed-loop brain-controlled functional electrical stimulation system according to a first preferred embodiment of the present invention
  • FIG. 3 is a flow chart of a genetic algorithm provided by the present invention.
  • the first preferred embodiment of the present invention provides a closed-loop brain-controlled functional electrical stimulation system 100, which is shown in FIG. 1 and includes a brain motion cortex region signal acquisition module 101, an information control module 102, a stimulator 103, a limb motion acquisition module 104 and a brain frontal signal acquisition module 105;
  • the information control module 102 mainly includes: three sub-modules: a signal pre-processing sub-module 1021, a pattern recognition sub-module 1022, and a command control sub-module 1023;
  • the signal pre-processing sub-module 1021 is configured to filter and shape the signal of the brain motion cortex region, and transmit the signal to the pattern recognition sub-module;
  • the pattern recognition sub-module 1022 is configured to identify the signal through the initial mapping matrix M 0 and the adjustable mapping matrix M 1 to determine the motion intention of the patient;
  • the command control sub-module 1023 is configured to determine a weight of the influence factor according to the BP neural network algorithm, and use a genetic algorithm to optimize the weight of the influence factor, generate a response command corresponding to the factor, and transmit the stimulus command to the stimulator.
  • the limb motion collecting module 103 is configured to collect data of the limb motion state, and transmit the data of the limb motion state to the information control module;
  • the brain frontal signal acquisition module 104 is configured to collect data of the patient's brain frontal lobe signal and transmit the data to the information control module;
  • the information control module 102 is configured to evaluate the current stimulation effect according to the feedback data of the limb motion state and the data of the brain frontal lobe signal, and automatically correct the adjustable mapping matrix of the pattern recognition submodule according to the evaluation result.
  • the technical solution provided by the first preferred embodiment of the present invention automatically adjusts the matrix of the adjustable mapping by means of closed-loop control, so that the electrical stimulation can be completed according to the user's wishes, thereby improving the suffering Rehabilitation effect.
  • FIG. 2 is a schematic diagram of a closed-loop brain-based functional electrical stimulation system.
  • the main functions of each module of the brain-controlled functional electrical stimulation system are as follows:
  • the brain motor cortex signal acquisition module is mainly composed of a bioelectrode and a signal transmission circuit.
  • the bioelectrode is first implanted into the brain motor cortex area of the sputum patient, and the signal of the brain motor cortex area is collected by the electrode, and the signal is transmitted to the information control module.
  • the information control module mainly includes three sub-modules: a signal pre-processing sub-module, a pattern recognition sub-module, and a command control sub-module.
  • the main functions of each submodule are as follows:
  • the function of the signal preprocessing sub-module is: filtering and waveform shaping the signal, and transmitting the signal to the pattern recognition sub-module, specifically: when the signal pre-processing sub-module receives the signal of the brain motion cortex area, first Low-pass filtering the signal, the cutoff frequency of the filter is about 50 Hz. Of course, in practical applications, other filter cut-off frequencies may also be used.
  • the specific embodiment of the present invention does not limit the specific value of the above-mentioned cut-off frequency.
  • the method of low-pass filtering can also adopt the method of the prior art, and the specific method of the low-pass filtering of the present invention is not limited.
  • the signal pre-processing sub-module performs waveform shaping on the signal to acquire characteristics of the signal, and the characteristics of the signal include, but are not limited to, the amplitude of the signal, the pulse width of the signal, the number of peaks of the signal, the rise time of the signal, and the time interval of the signal. Etc. and transfer it to the pattern recognition submodule.
  • the function of the pattern recognition sub-module is to accurately recognize the motion intention of the patient's limb according to the characteristics of the signal.
  • the specific process of pattern recognition is as follows:
  • t 1 is the amplitude 0-2.5 mA
  • t 2 is the amplitude 2.5-5 mA
  • t 3 is the amplitude 5-7.5 mA
  • t 4 is the amplitude 7.5 mA or more
  • t 5 is the pulse width 0-50 us
  • t 6 is the pulse width 50-100us
  • t 7 is pulse width 100-150us
  • t 8 is pulse width 150us or more
  • the limb movement outputted in the pattern recognition sub-module mainly includes the movement direction and the exercise intensity of the limb, and can be specifically divided into:
  • subscript g mn m ranges from 1-8, corresponding to 8 different limb movements: up, down, left, right, upper left, upper right, lower left, bottom right.
  • the g mn subscript n range is also from 1 to 8, corresponding to 8 different limb exercise intensities, from level 1 to level 8, and the higher the level, the stronger the exercise intensity.
  • G is the output limb motion matrix
  • T is the input influence factor matrix
  • M 0 is the initial mapping matrix
  • M 0 ⁇ m 01 , m 02 , m 03 , m 04 &, m 0n ⁇
  • M 1 is an automatically adjustable mapping matrix
  • M 1 ⁇ m 11 , m 12 , m 13 ,m 14 ??,m 1n ⁇ ,order
  • the size of each element in the M 1 matrix is automatically adjusted by the brain-controlled functional electrical stimulation system through closed-loop feedback. The specific process is explained below.
  • the function of the command control sub-module is: according to the result of the pattern recognition sub-module, and according to the actual situation of the patient (such as gender, age, sputum position, length of time, etc.), the command control sub-module generates a corresponding stimulation command, and stimulates Commands include controlling the signal waveform, amplitude, and Pulse width, frequency.
  • the specific working process of the command control submodule is as follows:
  • k 12 is the left calf ⁇
  • k 13 is the left thigh ⁇
  • k 14 is the right calf ⁇
  • k 15 is the right thigh ⁇
  • k 16 is the ⁇ time less than 1 year
  • k 17 is the ⁇ time 1 year
  • k 18 is ⁇ For a period of 2 years
  • k 19 is 3 years long
  • k 20 is longer than 3 years.
  • the weight of each input impact factor needs to be determined.
  • the command control sub-module the BP neural network algorithm is used to determine the weight of the input influence factor. The specific process is: using a three-layer neural network, that is, an input layer, an implicit layer and an output layer. According to the above analysis of the input influence factor, the number of neurons in the input layer is 20, the number of neurons in the hidden layer is 12, the number of neurons in the output layer is 1, and the output layer is the weight value of each impact factor. .
  • the weight of the influence factor needs to be optimized.
  • the genetic influence algorithm is used to optimize the input influence factor in the specific embodiment of the present invention.
  • the genetic algorithm completes the adaptive search process for the optimal solution of the problem through the imitation of the selection, crossover and mutation mechanism in the process of biological inheritance and evolution, so as to achieve the individual's adaptability.
  • the above optimized algorithm can also adopt other optimization algorithms, and the flow chart of the genetic algorithm is shown in FIG. 3, and the specific implementation is as follows:
  • the initial set value is an empirical value determined by the relevant clinical experimental results.
  • C min is the minimum objective function value in the last five iterations.
  • M is the size of the group, set here to 100
  • F i is the fitness of the i-th individual
  • P is the probability that the i-th individual is selected.
  • the main parameters include individual code string length l, population size M, crossover probability p c , mutation probability p m , termination algebra T and generation ditch G.
  • the individual code string length l is a variable length code
  • the population size setting M is 100
  • the crossover probability p c ranges from 0.3 to 0.9
  • the mutation probability p m ranges from 0.000001 to 0.1
  • the termination algebra T is set to 500. .
  • the output stimulation command includes the waveform, amplitude, pulse width and frequency of the stimulus signal, which can be divided into:
  • Y 2 ⁇ y 21 , y 22 , y 23 , y 24 , y 25 ⁇
  • Y 4 ⁇ y 41 ,y 42 ,y 43 ,y 44 ,y 45 ⁇
  • Y 1 is the waveform of the signal, and the five elements in the set correspond to: single-phase waveform, two-phase charge balance waveform, two-phase charge imbalance waveform, two-phase charge delay balance waveform, two-phase charge delay imbalance waveform;
  • Y 2 For the magnitude of the current, the five elements in the set correspond to: 60 mA, 70 mA, 80 mA, 90 mA, 100 mA;
  • Y 3 is the pulse width of the signal, and the five elements in the set correspond to: 0.2 ms, 0.3 ms, 0.4 ms, respectively. , 0.5ms, 0.6ms;
  • Y 4 is the frequency of the signal, and the five elements in the set correspond to: 20Hz, 40Hz, 60Hz, 80Hz, 100Hz.
  • the weight of the input influence factor can be calculated and optimized, but the weight of the input influence factor only reflects The weight assignment of 20 neural sub-modules in the input layer.
  • the input shadow is also needed.
  • the weights between the sets of sound factors are analyzed and processed.
  • the relationship between the input influence factor and the set output command set is mainly characterized by the absolute influence coefficient.
  • w jk is the weight value of the corresponding element in the input influence factor set.
  • the function of the stimulator is to generate stimulating currents that stimulate the muscles.
  • the specific process is: after receiving the signal, the stimulator demodulates the signal to obtain a stimulation command, and generates a corresponding stimulation current (ie, suitable signal waveform, amplitude, pulse width, frequency, time interval) according to the stimulation command to stimulate.
  • a corresponding stimulation current ie, suitable signal waveform, amplitude, pulse width, frequency, time interval
  • the limbs enable the limb to perform the corresponding action according to the wishes of the patient.
  • the main function of the limb motion acquisition module is to collect the motion state of the limb and convert it into a digital signal, and transmit the digital signal to the information control module by means of human body communication.
  • the specific process is: when the stimulator generates current and acts on the limb of the patient, the sensor in the limb motion acquisition module starts to collect the action state of the limb and converts it into a digital signal.
  • the sensor includes a direction sensor and an acceleration sensor.
  • the direction sensor mainly collects the movement direction of the limb when the limb is stimulated by the current; the angular velocity sensor mainly collects the movement change of the limb, thereby reflecting the strength of the limb movement when the limb is stimulated by the current.
  • the limb motion acquisition module transmits the digital signal to the information control module by means of human body communication.
  • the brain frontal signal acquisition module is mainly composed of a bioelectrode and a signal transmission circuit.
  • the bioelectrode is implanted into the frontal surface of the brain of the paralyzed patient, and the electrode is used to collect signals on the surface of the frontal lobe of the brain, and the signal is transmitted to the information control module of the brain-controlled functional electrical stimulation system.
  • the working process of the brain-controlled functional electrical stimulation system based on closed loop is as follows:
  • the brain motor cortex signal acquisition module begins to collect signals from the patient's brain motor cortex and transmits the signals to the brain-controlled functional electrical stimulation system through human communication. Information control module.
  • the signal preprocessing sub-module in the information control module performs waveform shaping on the received brain motor cortex signal, acquires characteristics of the signal (amplitude, pulse width, number of peaks, rise time, time interval), and transmits the characteristics of the signal.
  • the stimulator receives the stimulation command, and generates a corresponding stimulation current according to the stimulation command to act on the limb of the patient, so that the patient can complete the corresponding action according to his or her own desire; meanwhile, the direction sensor and the acceleration sensor in the limb motion acquisition module respectively The direction parameter of the limb movement and the intensity parameter of the muscle movement are collected, and converted into a digital signal, and the digital signal is transmitted to the information control sub-module by means of human body communication.
  • the brain frontal signal acquisition module has also begun to adopt The signals on the frontal surface of the patient's brain are collected and transmitted to the information control sub-module.
  • the information control module analyzes the limb motion data obtained by the limb motion acquisition module and the limb motion data determined by the pattern recognition sub-module in the information control module, and combines the data obtained by the brain frontal signal acquisition module to the brain.
  • current control functional electrical stimulation system to assess the effect of the stimulus, and the correction results of the evaluation in the pattern recognition sub-module mapping matrix M 1, forming a closed loop control system, improve the accuracy of pattern recognition and paralysis patient outcomes.
  • the correction of the mapping matrix M 1 in the recognition sub-module is mainly based on the automatic adjustment of the relationship between the deviation of the limb movement and the patient's satisfaction with the stimulation effect, as follows:
  • the input influence factor is mainly the deviation between the limb motion state data obtained by the limb motion acquisition module and the limb motion state data determined by the pattern recognition sub-module, including the deviation of the motion direction angle and the exercise intensity.
  • the output evaluation factor is mainly the satisfaction degree of the patient's stimulation effect on the brain-controlled functional electrical stimulation system.
  • the data is mainly derived from the signal obtained by the brain frontal signal acquisition module. By filtering and Fourier transforming the signal, the signal characteristics are extracted to judge the patient's satisfaction with the stimulation effect.
  • mapping matrix derived from f is the mapping matrix M 1 in the pattern recognition sub-module.
  • the closed-loop-based brain-controlled functional electrical stimulation system proposed by the invention can not only enable the paralyzed patients to complete the corresponding actions according to their own wishes, but also the system can evaluate the stimulation effect and automatically adjust the mapping matrix of the information control module. To ensure the accuracy of limb movement recognition and improve the rehabilitation effect of paralyzed patients.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • Veterinary Medicine (AREA)
  • Public Health (AREA)
  • General Health & Medical Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • Surgery (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Pathology (AREA)
  • Biophysics (AREA)
  • Physics & Mathematics (AREA)
  • Signal Processing (AREA)
  • Psychiatry (AREA)
  • Physiology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Radiology & Medical Imaging (AREA)
  • Artificial Intelligence (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Physical Education & Sports Medicine (AREA)
  • Psychology (AREA)
  • Dentistry (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Electrotherapy Devices (AREA)

Abstract

一种医疗器械领域闭环的脑控功能性电刺激系统，所述系统包括：大脑运动皮层区信号采集模块（101）、信息控制模块（102）、刺激器（103）、肢体动作采集模块（104）和大脑额叶信号采集模块（105）；其中，信息控制模块（102）包括：信号预处理子模块（1021），模式识别子模块（1022），命令控制子模块（1023）。该系统具有确保肢体动作识别的准确性，提高瘫痪患者的康复治疗效果的优点。

Description

闭环的脑控功能性电刺激系统 技术领域
本发明属于医疗器械领域，尤其涉及一种闭环的脑控功能性电刺激系统。
背景技术
随着交通事故、跌倒等意外伤害的不断增加，脊髓损伤发病人数不断上升。据估计，全球现有250万脊髓损伤患者，且每年新发生近13万例。功能性电刺激是一种临床上用于恢复由脊椎损伤造成神经功能障碍的主要治疗方法，其通过电流刺激患者的肢体，使失去运动功能的肢体能完成相应的动作。然而，目前，该方法只能通过预先设定的电刺激流程对患者进行治疗，若治疗时间过长，可能会引起患者的不适，甚至出现肌肉劳损现象，并且患者无法根据自己的意愿来完成相应的动作。
传统的功能性电刺激的原理是：利用一定强度的低频脉冲电流，通过预先设定的程序来刺激一组或多组肌肉，诱发肌肉运动或模拟正常的自主运动，以达到改善或恢复被刺激肌肉或肌群功能的目的，属于一种被动的康复方法。
所以现有技术提供的电刺激系统无法根据用户的意愿来完成电刺激。
发明内容
本发明实施例的目的在于提供一种闭环的脑控功能性电刺激系统，该系统能够依据用户的意愿来完成电刺激。
本发明提供一种闭环的脑控功能性电刺激系统，所述系统包括：大脑运动皮层区信号采集模块、信息控制模块、刺激器、肢体动作采集模块和大脑额叶信号采集模块；其中，
信息控制模块包括：信号预处理子模块，模式识别子模块，命令控制子模 块；
信号预处理子模块，用于对大脑运动皮层区的信号进行滤波和波形整形，并将信号传输到模式识别子模块；
模式识别子模块，用于通过初始映射矩阵M0和可调映射矩阵M1对信号进行识别，判断患者的运动意图；
命令控制子模块，用于根据BP神经网络算法确定影响因子的权重，并利用遗传算法优化影响因子的权重，生成该因子对应的刺激命令，将该刺激命令传输到刺激器中；
肢体动作采集模块，用于采集肢体运动状态的数据，并将肢体运动状态的数据传送到信息控制模块；
大脑额叶信号采集模块，用于采集患者大脑额叶信号的数据，并将数据传输到信息控制模块；
信息控制模块，还用于依据反馈的肢体运动状态的数据和大脑额叶信号的数据评价电流刺激效果，并根据评价结果自动修正模式识别子模块的可调映射矩阵M1。
可选的，所述信号预处理子模块，具体用于当接收到大脑运动皮层区的信号后，首先对大脑运动皮层区的信号进行低通滤波，然后对低通滤波后的信号进行波形整形，获取信号的特性，所述信号的特性包括：信号的幅度、信号的脉宽、信号的尖峰数目，信号的上升时间、信号的时间间隔，将所述信号的特性传输到所述模式识别子模块。
可选的，所述模式识别子模块，具体用于，
设定影响因子矩阵 T＝{t1,t2,t3,t4,t5,t6,t7,t8,t9,t10,t11,t12,t13,t14,t15,t16,t17,t18,t19,t20}；其中，t1为幅度0-2.5mA，t2为幅度2.5-5mA，t3为幅度5-7.5mA,t4为幅度7.5mA以上,t5为脉宽0-50us,t6为脉宽50-100us,t7为脉宽100-150us，,t8为脉宽150us以上，t9为尖峰数目1-2个，t10为尖峰数目3-4个，t11为尖峰数目5-6个，t12为尖峰数目6个以上，t13为上升时间0-5us,t14为上升时间5-10us,t15为上升时间11-15us，t16为上升时间15us以上，t17为时间间隔400-420us,t18为时间间隔420-440us,t19为时间间隔440-460us,t20为时间间隔460us以上；
设定输出的肢体运动矩阵G：
Figure PCTCN2016105385-appb-000001
其中gmn下标m范围为从1到8，分别对应8种不同的肢体运动方向：上、下、左、右、上左、上右、下左、下右；gmn下标n范围同样为从1到8，分别对应8种不同的肢体运动强度，从等级1到等级8；
确定输入的影响因子矩阵与输出的肢体运动矩阵的映射关系：两者的映 射关系可表示为：G＝M0×M1×T，其中G为输出的肢体运动矩阵，T为输入的影响因子矩阵，M0为初始的映射矩阵，M0＝{m01,m02,m03,m04......,m0n}，M1为可自动调整的映射矩阵,且M1＝{m11,m12,m13,m14......,m1n}，令
Figure PCTCN2016105385-appb-000002
根据映射关系G＝M0×M1×T计算G矩阵中每个元素的大小，如果
Figure PCTCN2016105385-appb-000003
则将G矩阵中每个元素进行归一化处理，并根据最大隶属原则，gi中的最大值则为模式识别子模块对患者肢体运动的识别结果。
可选的，所述命令控制子模块具体用于，
设定输入的影响因子集：K＝{k1,k2,k3,k4,k5,k6,k7,k8,k9,k10,k11,k12,k13,k14,k15,k16,k17,k18,k19,k20}，其中k1为模式识别子模块判定的结果，k2为男性瘫痪患者，k3为女性瘫痪患者，k4为幼儿瘫痪患者，k5为青年瘫痪患者，k6为中年瘫痪患者，k7为老年瘫痪患者，k8为左上臂瘫痪，k9为左前臂瘫痪，k10为右上臂瘫痪，k11为右前臂瘫痪，k12为左小腿瘫痪，k13为左大腿瘫痪，k14为右小腿瘫痪，k15为右大腿瘫痪，k16为瘫痪时长小于1年，k17为瘫痪时长1年，k18为瘫痪时长2年，k19为瘫痪时长3年，k20为瘫痪时长大于3年；
采用BP神经网络算法来确定输入影响因子的权重，所述影响因子包括：性别、年龄、瘫痪部位、瘫痪时长、患者运动意愿；
采用遗传算法对输入影响因子的权重进行优化；
设定输出命令集Y；
其中，
Y＝Y1×Y2×Y3×Y4
其中
Figure PCTCN2016105385-appb-000004
Y2＝{y21,y22,y23,y24,y25}，
Figure PCTCN2016105385-appb-000005
Y4＝{y41,y42,y43,y44,y45}
Y1为信号的波形，y11为单相波形，y12为双相电荷平衡波形，y13为双相电荷不平衡波形，y14为双相电荷延迟平衡波形，y15为双相电荷延迟不平衡波形；Y2为Y1波形对应的电流的幅值，Y3为Y1波形信号的脉宽；Y4为Y1波形信号的频率；
确定输入影响因子集与输出命令集的关系中绝对影响系数；
Figure PCTCN2016105385-appb-000006
其中
Figure PCTCN2016105385-appb-000007
wjk为输入影响因子集中相应元素的权重值；
刺激命令Y＝K*Sij。
可选的，所述刺激器，用于根据刺激命令产生相应的刺激电流。
可选的，所述信息控制模块，具体用于根据肢体动作的偏差与患者对刺激效果的满意程度之间的关系自动调整可调映射矩阵M1；具体为：
确定输入影响因子：输入影响因子主要是肢体动作采集模块所得到的肢体运动状态数据与模式识别子模块所判定的肢体运动状态数据之间的偏差，包括运动方向角度的偏差和运动强度的偏差，具体可表示为：U＝{u1,u2,u3,u4,u5,u6,u7,u8,u9,u10,u11,u12,u13,u14,u15,u16}，其中u1表示方向角度偏差为0，u2表示方向角度偏差为小于10°，u3表示表示方向角度偏差为小于30°，u4表示方向角度偏差为小于50°，u5表示方向角度偏差为小于70°，u6表示方向角度偏差为小于90°，u7表示方向角度偏差为小于150°，u8表示方向角度偏差为大于150°，u9表示运动强度等级偏差为0；u10表示运动强度等级偏差为1，u11表示运动强度等级偏差为2，u12表示运动强度等级偏差为3，u13表示运动强度等级偏差为4，u14表示运动强度等级偏差为5，u15表示运动强度等级偏差为6，u16表示运动强度等级偏差为6以上；
确定输出评价因子：输出评价因子主要是瘫痪患者对脑控功能性电刺激系统的刺激效果的满意程度。数据主要来源于大脑额叶信号采集模块所得到 的信号，通过对该信号进行滤波和傅里叶变换，提取信号特性，从而判断患者对刺激效果的满意程度。满意程度具体划分为：V＝{v1,v2,v3,v4,v5}，其中v1表示非常满意，v2表示满意，v3表示基本满意，v4表示不满意，v5表示非常不满意；
建立从输入影响因子到输出评价因子的评判矩阵，即得到一个从U→F(V)的映射矩阵；
Figure PCTCN2016105385-appb-000008
Figure PCTCN2016105385-appb-000009
根据f推导出的映射矩阵，即为模式识别子模块中的映射矩阵M1。
在本发明实施例中，本发明提供的技术方案提出一种基于闭环的脑控功能性电刺激系统，不但可以使瘫痪患者能根据自己的意愿完成相应的动作，同时，系统还能对刺激效果进行评估，并自动调整信息控制模块的映射矩阵，确保肢体动作识别的准确性，提高瘫痪患者的康复治疗效果。
附图说明
图1为本发明第一较佳实施方式提供的一种闭环的脑控功能性电刺激系统的结构图；
图2为本发明第一较佳实施方式提供的闭环的脑控功能性电刺激系统的示意图；
图3为本发明提供的遗传算法的流程图。
具体实施方式
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图及实 施例，对本发明进行进一步详细说明。应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。
本发明第一较佳实施方式提供一种闭环的脑控功能性电刺激系统100，该系统如图1所示，包括：大脑运动皮层区信号采集模块101、信息控制模块102、刺激器103、肢体动作采集模块104和大脑额叶信号采集模块105；其中，
信息控制模块102主要包括：三个子模块：信号预处理子模块1021，模式识别子模块1022，命令控制子模块1023；
信号预处理子模块1021，用于对大脑运动皮层区的信号进行滤波和波形整形，并将信号传输到模式识别子模块；
模式识别子模块1022，用于通过初始映射矩阵M0和可调映射矩阵M1对信号进行识别，判断患者的运动意图；
命令控制子模块1023，用于根据BP神经网络算法确定影响因子的权重，并利用遗传算法优化影响因子的权重，生成该因子对应刺激命令，将该刺激命令传输到刺激器中。
肢体动作采集模块103，用于采集肢体运动状态的数据，并将肢体运动状态的数据传送到信息控制模块；
大脑额叶信号采集模块104，用于采集患者大脑额叶信号的数据，并将数据传输到信息控制模块；
信息控制模块102，用于依据反馈的肢体运动状态的数据和大脑额叶信号的数据评价电流刺激效果，并根据评价结果自动修正模式识别子模块的可调映射矩阵。
本发明第一较佳实施方式提供的技术方案通过闭环控制的方式自动调整可调映射的矩阵，这样就能够根据用户的意愿来完成电刺激，实现提高瘫痪患 者的康复治疗效果。
图2为基于闭环的脑控功能性电刺激系统的示意图，脑控功能性电刺激系统各模块主要功能如下：
(一)大脑运动皮层信号采集模块
研究表明，人们的运动意图是由大脑运动皮层区的信号决定的。大脑运动皮层信号采集模块主要由生物电极和信号传输电路组成。为了减少环境噪声对脑控功能性电刺激系统的影响，首先将生物电极植入到瘫痪患者的大脑运动皮层区，并利用该电极采集大脑运动皮层区的信号，并将信号传输信息控制模块。
(二)信息控制模块
信息控制模块主要包括三个子模块：信号预处理子模块、模式识别子模块、命令控制子模块。各子模块的主要功能如下所示：
(1)信号预处理子模块的功能是：对信号进行滤波和波形整形，并将信号传输到模式识别子模块，具体为：当信号预处理子模块接收到大脑运动皮层区的信号后，首先对信号进行低通滤波，滤波器的截止频率约为50Hz，当然在实际应用中，也可以采用其他的滤波器截至频率，本发明具体实施方式并不限制上述截至频率的具体数值，另外，上述低通滤波的方法也可以采用现有技术的方法，本发明对低通滤波的具体方法并不限定。然后，信号预处理子模块对信号进行波形整形，获取信号的特性，上述信号的特性包括但不限于，信号的幅度、信号的脉宽、信号的尖峰数目，信号的上升时间、信号的时间间隔等，并将其传输到模式识别子模块。
(2)模式识别子模块的功能是：根据信号的特性，准确地识别出患者肢体的运动意图。模式识别的具体过程如下所示：
a、设定输入的影响因子矩阵：输入的影响因子主要包括信号的幅度、脉宽、尖峰数目，上升时间、时间间隔，具体可划分为：T＝{t1,t2,t3,t4,t5,t6,t7,t8,t9,t10,t11,t12,t13,t14,t15,t16,t17,t18,t19,t20},其中t1为幅度0-2.5mA，t2为幅度2.5-5mA，t3为幅度5-7.5mA,t4为幅度7.5mA以上,t5为脉宽0-50us,t6为脉宽50-100us,t7为脉宽100-150us，,t8为脉宽150us以上，t9为尖峰数目1-2个，t10为尖峰数目3-4个，t11为尖峰数目5-6个，t12为尖峰数目6个以上，t13为上升时间0-5us,t14为上升时间5-10us,t15为上升时间11-15us，t16为上升时间15us以上，t17为时间间隔400-420us,t18为时间间隔420-440us,t19为时间间隔440-460us,t20为时间间隔460us以上。
b、设定输出的肢体运动矩阵：模式识别子模块中输出的肢体运动主要包括肢体的运动方向和运动强度，具体可划分为：
Figure PCTCN2016105385-appb-000010
其中gmn下标m范围为从1到8，分别对应8种不同的肢体运动方向：上、下、左、右、上左、上右、下左、下右。gmn下标n范围同样为从1到8，分别对应8种不同的肢体运动强度，从等级1到等级8，且等级越高，运动强度越强。
c、确定输入的影响因子矩阵与输出的肢体运动矩阵的映射关系：在模式识别子模块中，需要根据信号的特性识别出患者的肢体运动意图，两者的映射关系可表示为：G＝M0×M1×T。其中G为输出的肢体运动矩阵，T为输入的影响因子矩阵，M0为初始的映射矩阵，M0＝{m01,m02,m03,m04......,m0n}，且矩阵中每个元素的大小根据临床的数据和经验来确定，一旦确定后，则保持不变；M1为可自动调整的映射矩阵,且M1＝{m11,m12,m13,m14......,m1n}，令
Figure PCTCN2016105385-appb-000011
M1矩阵中每个元素的大小由脑控功能电刺激系统经闭环反馈后实现自动调整，具体的过程在下面会进行解释。
M1的大小确定后，根据映射关系G＝M0×M1×T计算G矩阵中每个元素的大小，如果
Figure PCTCN2016105385-appb-000012
则将G矩阵中每个元素进行归一化处理，并根据最大隶属原则，gi中的最大值则为模式识别子模块对患者肢体运动的识别结果。
(3)命令控制子模块的功能是：根据模式识别子模块的结果，并根据患者的实际情况(如性别、年龄、瘫痪部位、瘫痪时长等)，命令控制子模块生成相应的刺激命令，刺激命令包括控制刺激器输出的信号波形、幅值、 脉宽、频率。命令控制子模块具体的工作过程如下：
a、设定输入的影响因子集：输入的影响因子包括模式识别子模块判定的肢体运动结果、患者的性别、年龄、瘫痪部位、瘫痪时长。具体划分为：K＝{k1,k2,k3,k4,k5,k6,k7,k8,k9,k10,k11,k12,k13,k14,k15,k16,k17,k18,k19,k20}，其中k1为模式识别子模块判定的结果，k2为男性瘫痪患者，k3为女性瘫痪患者，k4为幼儿瘫痪患者，k5为青年瘫痪患者，k6为中年瘫痪患者，k7为老年瘫痪患者，k8为左上臂瘫痪，k9为左前臂瘫痪，k10为右上臂瘫痪，k11为右前臂瘫痪，k12为左小腿瘫痪，k13为左大腿瘫痪，k14为右小腿瘫痪，k15为右大腿瘫痪，k16为瘫痪时长小于1年，k17为瘫痪时长1年，k18为瘫痪时长2年，k19为瘫痪时长3年，k20为瘫痪时长大于3年。
b、确定输入影响因子的权重：由于每一个瘫痪患者的实际情况都存在一定的区别，如性别、年龄、瘫痪部位、瘫痪时长、患者运动意愿等。在不同的情况下，每个影响因子对治疗效果的影响会有所不同。为了取得最佳的治疗效果，需要确定每一个输入影响因子的权重。在命令控制子模块中，采用BP神经网络算法来确定输入影响因子的权重。具体过程为：采用三层的神经网络，即一个输入层，一个隐含层和一个输出层。根据上述的输入影响因子的分析，输入层的神经元数目为20个，隐含层的神经元数目为12个，输出层的神经元数目为1个，输出层即为各影响因子的权重值。
表1神经网络算法的具体参数
Figure PCTCN2016105385-appb-000013
c、输入影响因子的权重的优化：采用BP神经网络算法确定输入影响因子的权重后，还需对影响因子的权重进行优化。为了取得最佳的优化效果，在本发明具体实施方式中采用遗传算法对输入影响因子进行优化。遗传算法借助生物遗传学的观点，通过对生物遗传和进化过程中的选择、交叉、变异机理的模仿，来完成对问题最优解的自适应搜索过程，以实现个体的适应性的提高。当然在实际应用中，上述优化的算法还可以采用其他的优化算法，遗传算法的流程图如图3所示，具体实现如下：
1)设定初始的群体：在本发明具体实施方式中，初始的设定值为经验值，由相关的临床实验结果确定。
2)确定群体的多样性：群体规模越大，群体中个体的多样性越高，算法陷入局部解的危险就越小，但群体规模太大会带来计算量增加的弊病，因此，在本专利中，群体的规模设为100较为合适。
3)确定适应度函数：根据最优化问题的类型，由目标函数值按一定的转换规则求目标函数的全局最大值，具体计算为：
Figure PCTCN2016105385-appb-000014
其中Cmin为当最近五次迭代中的最小目标函数值。
4)确定选择算子：采用比例选择算子进行选择，，具体计算为：
Figure PCTCN2016105385-appb-000015
其中M为群体的大小，在这里设置为100，Fi为第i个个体的适应度，Pis为第i个个体被选中的概率。
5)设定遗传算法的运行参数：主要参数包括个体编码串长度l，群体大小M，交叉概率pc，变异概率pm、终止代数T和代沟G等。其中个体编码串长度l采用的是可变长度的编码，群体大小设置M为100，交叉概率pc的范围为0.3-0.9，变异概率pm的范围为0.000001-0.1；终止代数T设为500。
d、设定输出命令集：输出的刺激命令包括刺激信号的波形，幅值、脉宽、频率，具体可划分为：
Y＝Y1×Y2×Y3×Y4
其中
Figure PCTCN2016105385-appb-000016
Y2＝{y21,y22,y23,y24,y25}，
Figure PCTCN2016105385-appb-000017
Y4＝{y41,y42,y43,y44,y45}
Y1为信号的波形，集合中五个元素分别对应于：单相波形，双相电荷平衡波形，双相电荷不平衡波形，双相电荷延迟平衡波形，双相电荷延迟不平衡波形；Y2为电流的幅值，集合中五个元素分别对应于：60mA,70mA,80mA,90mA,100mA；Y3为信号的脉宽，集合中五个元素分别对应于：0.2ms,0.3ms,0.4ms,0.5ms,0.6ms；Y4为信号的频率，集合中五个元素分别对应于：20Hz,40Hz,60Hz,80Hz,100Hz。
e、确定输入影响因子集与输出命令集的关系：在上述的阐述中，根据BP神经网络算法和遗传算法，可以计算和优化输入影响因子的权重，但该输入影响因子的权重只是反映了在输入层中20个神经子模块的权重分配情况。为了得到输入影响因子集和输出命令集的之间的关系，还需要对输入影 响因子集之间的权重加以分析处理。主要是通过绝对影响系数来表征输入影响因子与集合输出命令集的关系。
绝对影响系数
Figure PCTCN2016105385-appb-000018
其中
Figure PCTCN2016105385-appb-000019
wjk为输入影响因子集中相应元素的权重值。
(三)刺激器
刺激器的功能是产生刺激电流，刺激肌肉。具体过程为：刺激器接收到信号后，对信号进行解调，得到刺激命令，并根据刺激命令产生相应的刺激电流(即合适的信号波形、幅值、脉宽、频率、时间间隔)来刺激肢体，使肢体能够根据瘫痪患者的意愿完成相应的动作。
(四)肢体动作采集模块
肢体动作采集模块的主要功能是采集肢体的运动状态，并转换成数字信号，通过人体通信的方式，将数字信号传输到信息控制模块。具体过程为：当刺激器产生电流，并作用于患者的肢体时，肢体动作采集模块中的传感器则开始采集肢体的动作状态，并转换成数字信号。其中，传感器包括方向传感器和加速度传感器。方向传感器主要采集当肢体受到电流刺激时，肢体的运动方向；角速度传感器则主要采集肢体的运动变化，从而反映出肢体受到电流刺激时，肢体运动的强度。采集完毕后，肢体动作采集模块通过人体通信的方式将数字信号传输到信息控制模块。
(五)大脑额叶信号采集模块
大脑额叶信号采集模块主要由生物电极和信号传输电路组成。首先将生物电极植入到瘫痪患者的大脑额叶表面，并利用该电极采集大脑额叶表面的信号，并将信号传输到脑控功能性电刺激系统的信息控制模块。
基于闭环的脑控功能性电刺激系统的工作过程如下：
首先，当脑控功能性电刺激系统处于工作状态时，大脑运动皮层信号采集模块开始采集瘫痪患者的大脑运动皮层区的信号，并将信号通过人体通信方式传输到脑控功能性电刺激系统的信息控制模块。
其次，信息控制模块中的信号预处理子模块对接收到的大脑运动皮层信号进行波形整形，获取信号的特性(幅度、脉宽、尖峰数目，上升时间、时间间隔)，并将信号的特性传输到模式识别子模块；模式识别子模块根据从输入的影响因子映射到输出的肢体运动状态的初始映射矩阵M0和可调的映射矩阵M1来计算肢体运动意图，其中矩阵M0由经验值确定，矩阵M1则根据闭环反馈自动调整，并根据最大隶属原则确定瘫痪患者的肢体运动意图；命令控制子模块根据模式识别子模块的结果，并结合患者性别、年龄、瘫痪部位、瘫痪时长的因素，根据BP神经网络算法确定影响因子的权重，利用遗传算法对影响因子进行优化，并通过绝对影响系数来确定输入参数与输出刺激命令的对应关系，生成相应的刺激命令，并将刺激命令通过人体通信的方式将刺激命令传输到刺激器。
再次，刺激器接收到刺激命令，并根据刺激命令产生相应的刺激电流作用于患者的肢体，使患者能根据自己的意愿完成相应的动作；同时，肢体动作采集模块中的方向传感器和加速度传感器分别采集肢体运动时的方向参数和肌肉运动时的强度参数，并将其转换为数字信号，通过人体通信的方式将数字信号传输到信息控制子模块。此外，大脑额叶信号采集模块也开始采 集患者的大脑额叶表面的信号，并将信号传输到信息控制子模块。
最后，信息控制模块通过分析肢体动作采集模块所得到的肢体运动数据、以及信息控制模块中的模式识别子模块所判定的肢体运动数据，并结合大脑额叶信号采集模块所得到的数据，对脑控功能性电刺激系统的电流刺激效果进行评估，并根据评估结果修正模式识别子模块中的映射矩阵M1，形成一个闭环的控制系统，提高模式识别的正确率和瘫痪患者的治疗效果。识别子模块中的映射矩阵M1的修正主要是根据肢体动作的偏差与患者对刺激效果的满意程度之间的关系进行自动调整，具体如下：
1)确定输入影响因子：输入影响因子主要是肢体动作采集模块所得到的肢体运动状态数据与模式识别子模块所判定的肢体运动状态数据之间的偏差，包括运动方向角度的偏差和运动强度的偏差，具体可表示为：U＝{u1,u2,u3,u4,u5,u6,u7,u8,u9,u10,u11,u12,u13,u14,u15,u16}，其中u1表示方向角度偏差为0，u2表示方向角度偏差为小于10°，u3表示表示方向角度偏差为小于30°，u4表示方向角度偏差为小于50°，u5表示方向角度偏差为小于70°，u6表示方向角度偏差为小于90°，u7表示方向角度偏差为小于150°，u8表示方向角度偏差为大于150°，u9表示运动强度等级偏差为0；u10表示运动强度等级偏差为1，u11表示运动强度等级偏差为2，u12表示运动强度等级偏差为3，u13表示运动强度等级偏差为4，u14表示运动强度等级偏差为5，u15表示运动强度等级偏差为6，u16表示运动强度等级偏差 为6以上。
2)确定输出评价因子：输出评价因子主要是瘫痪患者对脑控功能性电刺激系统的刺激效果的满意程度。数据主要来源于大脑额叶信号采集模块所得到的信号，通过对该信号进行滤波和傅里叶变换，提取信号特性，从而判断患者对刺激效果的满意程度。满意程度具体划分为：V＝{v1,v2,v3,v4,v5}，其中v1表示非常满意，v2表示满意，v3表示基本满意，v4表示不满意，v5表示非常不满意。
3)建立从输入影响因子到输出评价因子的评判矩阵，即得到一个从U→F(V)的映射矩阵。
Figure PCTCN2016105385-appb-000020
Figure PCTCN2016105385-appb-000021
根据f推导出的映射矩阵，即为模式识别子模块中的映射矩阵M1。
本发明提出的基于闭环的脑控功能性电刺激系统，不但可以使瘫痪患者能根据自己的意愿完成相应的动作，同时，系统还能对刺激效果进行评估，并自动调整信息控制模块的映射矩阵，确保肢体动作识别的准确性，提高瘫痪患者的康复治疗效果。
以上所述仅为本发明的较佳实施例而已，并不用以限制本发明，凡在本发明的精神和原则之内所作的任何修改、等同替换和改进等，均应包含在本发明的保护范围之内。

Claims (6)

  1. 一种闭环的脑控功能性电刺激系统，其特征在于，所述系统包括：大脑运动皮层区信号采集模块、信息控制模块、刺激器、肢体动作采集模块和大脑额叶信号采集模块；其中，
    信息控制模块包括：信号预处理子模块，模式识别子模块，命令控制子模块；
    信号预处理子模块，用于对大脑运动皮层区的信号进行滤波和波形整形，并将信号传输到模式识别子模块；
    模式识别子模块，用于通过初始映射矩阵M0和可调映射矩阵M1对信号进行识别，判断患者的运动意图；
    命令控制子模块，用于根据BP神经网络算法确定影响因子的权重，并利用遗传算法优化影响因子的权重，生成该因子对应的刺激命令，将该刺激命令传输到刺激器中；
    肢体动作采集模块，用于采集肢体运动状态的数据，并将肢体运动状态的数据传送到信息控制模块；
    大脑额叶信号采集模块，用于采集患者大脑额叶信号的数据，并将数据传输到信息控制模块；
    信息控制模块，还用于依据反馈的肢体运动状态的数据和大脑额叶信号的数据评价电流刺激效果，并根据评价结果自动修正模式识别子模块的可调映射矩阵M1。
  2. 根据权利要求1所述的系统，其特征在于，所述信号预处理子模块，具体用于当接收到大脑运动皮层区的信号后，首先对大脑运动皮层区的信号进行 低通滤波，然后对低通滤波后的信号进行波形整形，获取信号的特性，所述信号的特性包括：信号的幅度、信号的脉宽、信号的尖峰数目，信号的上升时间、信号的时间间隔，将所述信号的特性传输到所述模式识别子模块。
  3. 根据权利要求2所述的系统，其特征在于，所述模式识别子模块，具体用于，
    设定影响因子矩阵T＝{t1,t2,t3,t4,t5,t6,t7,t8,t9,t10,t11,t12,t13,t14,t15,t16,t17,t18,t19,t20}；其中，t1为幅度0-2.5mA，t2为幅度2.5-5mA，t3为幅度5-7.5mA,t4为幅度7.5mA以上,t5为脉宽0-50us,t6为脉宽50-100us,t7为脉宽100-150us，,t8为脉宽150us以上，t9为尖峰数目1-2个，t10为尖峰数目3-4个，t11为尖峰数目5-6个，t12为尖峰数目6个以上，t13为上升时间0-5us,t14为上升时间5-10us,t15为上升时间11-15us，t16为上升时间15us以上，t17为时间间隔400-420us,t18为时间间隔420-440us,t19为时间间隔440-460us,t20为时间间隔460us以上；
    设定输出的肢体运动矩阵G：
    Figure PCTCN2016105385-appb-100001
    其中gmn下标m范围为从1到8，分别对应8种不同的肢体运动方向：上、下、左、右、上左、上右、下左、下右；gmn下标n范围同样为从1到8，分别对应8种不同的肢体运动强度，从等级1到等级8；
    确定输入的影响因子矩阵与输出的肢体运动矩阵的映射关系：两者的映射关系可表示为：G＝M0×M1×T，其中G为输出的肢体运动矩阵，T为输入的影响因子矩阵，M0为初始的映射矩阵，M0＝{m01,m02,m03,m04......,m0n}，M1为可自动调整的映射矩阵,且M1＝{m11,m12,m13,m14......,m1n}，令
    Figure PCTCN2016105385-appb-100002
    根据映射关系G＝M0×M1×T计算G矩阵中每个元素的大小，如果
    Figure PCTCN2016105385-appb-100003
    则将G矩阵中每个元素进行归一化处理，并根据最大隶属原则，gi中的最大值则为模式识别子模块对患者肢体运动的识别结果。
  4. 根据权利要求3所述的系统，其特征在于，所述命令控制子模块具体 用于，
    设定输入的影响因子集：K＝{k1,k2,k3,k4,k5,k6,k7,k8,k9,k10,k11,k12,k13,k14,k15,k16,k17,k18,k19,k20}，其中k1为模式识别子模块判定的结果，k2为男性瘫痪患者，k3为女性瘫痪患者，k4为幼儿瘫痪患者，k5为青年瘫痪患者，k6为中年瘫痪患者，k7为老年瘫痪患者，k8为左上臂瘫痪，k9为左前臂瘫痪，k10为右上臂瘫痪，k11为右前臂瘫痪，k12为左小腿瘫痪，k13为左大腿瘫痪，k14为右小腿瘫痪，k15为右大腿瘫痪，k16为瘫痪时长小于1年，k17为瘫痪时长1年，k18为瘫痪时长2年，k19为瘫痪时长3年，k20为瘫痪时长大于3年；
    采用BP神经网络算法来确定输入影响因子的权重，所述的影响因子包括：性别、年龄、瘫痪部位、瘫痪时长、患者运动意愿；
    采用遗传算法对输入影响因子的权重进行优化；
    设定输出命令集Y；
    其中，
    Y＝Y1×Y2×Y3×Y4
    其中
    Figure PCTCN2016105385-appb-100004
    Y2＝{y21,y22,y23,y24,y25}，
    Figure PCTCN2016105385-appb-100005
    Y4＝{y41,y42,y43,y44,y45}
    Y1为信号的波形，y11为单相波形，y12为双相电荷平衡波形，y13为双相电荷不平衡波形，y14为双相电荷延迟平衡波形，y15为双相电荷延迟不平衡波形；Y2为Y1波形对应的电流的幅值，Y3为Y1波形信号的脉宽；Y4为Y1波形信号的频率；
    确定输入影响因子集与输出命令集的关系中绝对影响系数；
    Figure PCTCN2016105385-appb-100006
    其中
    Figure PCTCN2016105385-appb-100007
    wjk为输入影响因子集中相应元素的权重值；
    刺激命令Y＝K*Sij。
  5. 根据权利要求4所述的系统，其特征在于，
    所述刺激器，用于根据刺激命令产生相应的刺激电流。
  6. 根据权利要求1所述的系统，其特征在于，所述信息控制模块，具体用于根据肢体动作的偏差与患者对刺激效果的满意程度之间的关系自动调整可调映射矩阵M1。
PCT/CN2016/105385 2015-12-16 2016-11-11 闭环的脑控功能性电刺激系统 WO2017101621A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510953497.3A CN105727442B (zh) 2015-12-16 2015-12-16 闭环的脑控功能性电刺激系统
CN201510953497.3 2015-12-16

Publications (1)

Publication Number Publication Date
WO2017101621A1 true WO2017101621A1 (zh) 2017-06-22

Family

ID=56296030

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/105385 WO2017101621A1 (zh) 2015-12-16 2016-11-11 闭环的脑控功能性电刺激系统

Country Status (2)

Country Link
CN (1) CN105727442B (zh)
WO (1) WO2017101621A1 (zh)

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105727442B (zh) * 2015-12-16 2018-11-06 深圳先进技术研究院 闭环的脑控功能性电刺激系统
CN106264520A (zh) * 2016-07-27 2017-01-04 深圳先进技术研究院 一种神经反馈运动训练系统及方法
CN108392734A (zh) * 2018-01-30 2018-08-14 深圳市前海未来无限投资管理有限公司 一种电刺激保健方法、装置及可穿戴设备
CN108310633A (zh) * 2018-02-01 2018-07-24 核工业四六医院 一种智能型瘫痪病人辅助康复系统
CN109011141B (zh) * 2018-05-02 2022-06-07 中国人民解放军军事科学院军事医学研究院 用于大鼠脑部的单双相恒流电刺激器
CN108926766A (zh) * 2018-08-06 2018-12-04 南通大学 一种微刺激器及微刺激器系统
RU2738197C2 (ru) * 2018-09-24 2020-12-09 "Ай-Брэйн Тех ЛТД" Система и способ формирования команд управления на основании биоэлектрических данных оператора
CN109124590A (zh) * 2018-10-19 2019-01-04 南通大学 大鼠脑初级运动皮层神经在脊髓中的传导功能检测系统
DE202019102592U1 (de) * 2019-05-08 2020-08-11 Albert-Ludwigs-Universität Freiburg Adaptive Tiefenhirnstimulation des superolateralen medialen Vorderhirnbündels
CN110694169A (zh) * 2019-09-16 2020-01-17 浙江大学 基于运动意图诱发中枢神经系统微电刺激的运动功能障碍神经桥接系统
CN113197585B (zh) * 2021-04-01 2022-02-18 燕山大学 一种神经肌肉信息交互模型构建及参数辨识优化方法
CN116549843A (zh) * 2023-07-11 2023-08-08 杭州般意科技有限公司 经颅直流电刺激的干预电流控制方法、装置及终端设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101947356A (zh) * 2010-10-22 2011-01-19 上海交通大学 基于脑机交互的损伤脑功能康复装置
US20140277582A1 (en) * 2013-03-15 2014-09-18 Neurolutions, Inc. Brain-controlled body movement assistance devices and methods
CN104173124A (zh) * 2014-08-29 2014-12-03 电子科技大学 一种基于生物信号的上肢康复系统
CN104548347A (zh) * 2015-01-22 2015-04-29 天津大学 一种纯意念神经肌肉电刺激控制与神经功能评价系统
CN105054927A (zh) * 2015-07-16 2015-11-18 西安交通大学 一种下肢康复系统中主动参与度的生物量化评估方法
CN105727442A (zh) * 2015-12-16 2016-07-06 深圳先进技术研究院 闭环的脑控功能性电刺激系统

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102309365A (zh) * 2011-08-30 2012-01-11 西安交通大学苏州研究院 一种可穿戴的脑控智能假肢
CN104951082B (zh) * 2015-07-09 2018-01-12 浙江大学 一种利用随机共振增强脑电信号的脑机接口方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101947356A (zh) * 2010-10-22 2011-01-19 上海交通大学 基于脑机交互的损伤脑功能康复装置
US20140277582A1 (en) * 2013-03-15 2014-09-18 Neurolutions, Inc. Brain-controlled body movement assistance devices and methods
CN104173124A (zh) * 2014-08-29 2014-12-03 电子科技大学 一种基于生物信号的上肢康复系统
CN104548347A (zh) * 2015-01-22 2015-04-29 天津大学 一种纯意念神经肌肉电刺激控制与神经功能评价系统
CN105054927A (zh) * 2015-07-16 2015-11-18 西安交通大学 一种下肢康复系统中主动参与度的生物量化评估方法
CN105727442A (zh) * 2015-12-16 2016-07-06 深圳先进技术研究院 闭环的脑控功能性电刺激系统

Also Published As

Publication number Publication date
CN105727442B (zh) 2018-11-06
CN105727442A (zh) 2016-07-06

Similar Documents

Publication Publication Date Title
WO2017101621A1 (zh) 闭环的脑控功能性电刺激系统
US11883175B2 (en) Paretic limb rehabilitation methods and systems
CN110355761B (zh) 一种基于关节刚度和肌肉疲劳的康复机器人控制方法
Graupe EMG pattern analysis for patient-responsive control of FES in paraplegics for walker-supported walking
JP4695605B2 (ja) 神経筋刺激
JP7141205B2 (ja) 能動閉ループ医療システム
CN106377837A (zh) 基于步态识别的功能性肌肉电刺激助行装置及控制方法
CN106691770A (zh) 产生并执行训练课程程序
CN110248693B (zh) 电刺激设备
US20200147382A1 (en) Control system for movement reconstruction and/or restoration for a patient
KR20130044234A (ko) 트레이닝 장치
US11752325B2 (en) Functional electrical stimulation system offering coordinated and natural movements
US20200254260A1 (en) System for neuromodulation
Hu et al. EEG-based classification of upper-limb ADL using SNN for active robotic rehabilitation
CN110339024A (zh) 下肢外骨骼机器人及其实时步态切换方法及存储装置
CN109106339A (zh) 一种功能性电刺激下肘关节力矩的在线估计方法
CN106345055A (zh) 一种基于步态识别的功能性肌肉电刺激助行装置
CN106693178A (zh) 一种基于自主意愿的上肢康复功能性电刺激闭环控制方法
CN115177864A (zh) 基于肌肉激活度与lstm的功能性电刺激闭环调控方法
CN108543216A (zh) 一种基于主从控制的手部功能重建装置及其实现方法
US11580877B2 (en) Movement reconstruction control system
CN102697496B (zh) 一种功能性电刺激表面肌电信号的滤波方法
CN107596560A (zh) 一种基于角速度信号的足下垂助行仪的控制方法
Serea et al. Preliminary tests on a hybrid upper arm exoskeleton for upper arm rehabilitation for disabled patients
WO2019173084A1 (en) Non-invasive nerve stimulation

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16874678

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16874678

Country of ref document: EP

Kind code of ref document: A1