CN104000586A - Stroke patient rehabilitation training system and method based on brain myoelectricity and virtual scene - Google Patents

Abstract

Provided are a stroke patient rehabilitation training system and method based on brain myoelectricity and a virtual scene. Control over the virtual rehabilitation scene is achieved through myoelectric signals, and rehabilitation training intensity is adjusted in a self-adaptation mode with a brain myoelectricity fatigue index combined. The design of the virtual rehabilitation scene is completed with the needs of stroke patient rehabilitation training and the advice of a rehabilitation physician combined, the brain fatigue index is provided, and quantitative evaluation on brain region fatigue is achieved. The surface myoelectric signal features under different motion modes of an arm are extracted, the motion intention of a patient is obtained, and control over the virtual rehabilitation scene is achieved. The muscle fatigue and brain fatigue index comprehensive features are extracted, the fatigue state of a rehabilitation patient is obtained, self-adaptation rehabilitation training scene adjusting is achieved, rehabilitation training intensity is relieved or enhanced, and secondary damage caused by improper training is avoided. The system and method have the advantages of being high in safety, high in intelligence and scientific in training, and damage cannot happen easily.

Description

Patients with cerebral apoplexy rehabilitation training system and method based on brain myoelectricity and virtual scene

Technical field

The present invention relates to rehabilitation medicine equipment technical field, especially a kind of rehabilitation training system for patients with cerebral apoplexy and method.

Background technology

Apoplexy is the known apoplexy of people, and it has, and sickness rate is high, mortality rate is high, disability rate is high, relapse rate high, so medical circle is listed as its same coronary heart disease, cancer one of three large diseases that threaten human health.Clinical research shows, by timely, positive rehabilitation training, most of paralytic can recover simple limb motion ability, even recovery from illness.Traditional stroke at convalescence Therapeutic Method be reflect or graded movement to control be theoretical basis, be mainly to rely on manually auxiliary Rehabilitation training of physiatrician, but cannot be protected the time of rehabilitation training and the course for the treatment of, thereby affect rehabilitation efficacy.

Along with the development of robotics, the training of robot assisted motor function is arisen at the historic moment, and is applied to the training of patients with cerebral apoplexy late rehabilitation.The variation that robot could control and quantize training strength, measure objectively kinesiology and strength in training process, provides that patients with cerebral apoplexy is repeatable, task orientation and interactively treatment.The people such as Gerdienk think that robot assisted technology is more better than conventional exercises method to the effect of improving of motor control, can effectively improve convalescent period patients with cerebral apoplexy limb motion control and functional level.Yet, during current patients with cerebral apoplexy application healing robot auxiliary rehabilitation exercise, mostly accept passive treatment, patient's motion intention is seldom embodied in rehabilitation course, and the enthusiasm and the initiative that make patient participate in rehabilitation training are short of to some extent.

In recent years, scientific research institution's Combining with technology of virtual reality has realized new Therapeutic Method both at home and abroad---computer assisted motor function training, and be applied to patients with cerebral apoplexy late rehabilitation training, for example: EBRSR (Evidence-Based Review of Stroke Rehabilitation) guide in 2008 is recommended in apoplexy and uses virtual reality technology to improve patient moving function, and recommendation intensity is A.Computer assisted motor function training, is to make patient attention be concentrated on to result rather than the motion itself of motion by game, can strengthen to a certain extent the interest of rehabilitation.But above-mentioned computer aided training system can only assist patient to complete relatively simple rehabilitation training, is still difficult to transfer patient's active sense of participation and self-confidence.While, due to the Evaluation Strategy of shortcoming to patient physiological state, produces fatigue in Rehabilitation process, be easy to occur surprisingly causing secondary injured, and then limited the clinical application of virtual reality technology.

Summary of the invention

For mentioning the deficiency of current patients with cerebral apoplexy in rehabilitation training in above-mentioned technology, the object of the invention is to provide a kind of virtual scene that utilizes to train, in training process, understand patient's fatigue strength, also can increase patients with cerebral apoplexy rehabilitation training system and the method based on brain myoelectricity and virtual scene of training interest and safety.

For achieving the above object, adopted following technical scheme: system of the present invention comprises signals collecting part, data preprocessing part, motion index Extraction parts, exercise fatigue index extraction part, virtual scene design part and rehabilitation training experimental section; Wherein,

Described signals collecting part is extracted patient's EEG signals and electromyographic signal;

Described data preprocessing part carries out Filtering Processing to the EEG signals collecting and electromyographic signal;

Described motion index Extraction parts is that the quantitative analysis of discharging during to patient's muscle movement is extracted;

Described exercise fatigue index extraction is partly that patient's EEG signals and electromyographic signal are analyzed respectively to extraction, and then judgement exercise fatigue or brain fag;

Described virtual scene design be partly under computer environment based on Visual C#2010 development and Design, generating virtual scene show output, sets up man-machine interaction feedback mechanism on computers;

Described rehabilitation training experimental section be patient according to rehabilitation training requirement in virtual scene, by arm, bend, stretch with left and right waving and control virtual palm in virtual scene and complete appointed task.

In described signals collecting part, the collection of electromyographic signal adopts the differential input of bikini, two differential input ends that be myoelectricity wherein, and another one be reference, difference input electrode, along muscle fiber direction, is placed on belly of muscle place; Eeg signal acquisition adopts 8 passage brain myoelectricity synchronous acquisition instrument to gather, 10～20 electrodes of adopting international standards are placed standards, by electrode cap, electrode is connected with scalp, adopt the single-stage method of leading, reference electrode leads and is connected respectively to mastoid process place after the ear of left and right, and ground electrode arrangement overhead hits exactly.

Described data preprocessing part, is used self adaptation high pass filter and self adaptation 50Hz notch filter wave filter respectively EEG signals, electromyographic signal to be carried out to Filtering Processing, and baseline drift and the power frequency removed in signal are disturbed; Re-use the logical FIR wave filter of Butterworth three rank bands EEG signals, electromyographic signal are processed, according to effective frequency range feature of signal, the cut-off frequency of choosing electromyographic signal is 2Hz～200Hz, and the cut-off frequency of choosing EEG signals is 2Hz～50Hz.

Described motion index Extraction parts, electromyographic signal motion index is as follows,

$\mathrm{iEMG}=\underset{t}{\overset{t+T}{\∫}}\left|\mathrm{EMG}\left(t\right)\right|\mathrm{dt}$

In formula, iEMG is integration myoelectricity value, the quantity of moving cell and the discharge magnitude of each moving cell during reflection muscle movement; T is for gathering the time of electromyographic signal; T is for analyzing the cycle of the electromyographic signal collecting; The electromyographic signal of the respective muscle motion that EMG (t) collects constantly for t.

Described exercise fatigue index extraction partly comprises electromyographic signal fatigue index and brain fag index;

(1) electromyographic signal fatigue index as shown in the formula,

$\mathrm{MPF}=\underset{0}{\overset{\∞}{\∫}}f\·P\left(f\right)\mathrm{df}/\underset{0}{\overset{\∞}{\∫}}P\left(f\right)\mathrm{df}$

In formula, MPF is frequency of average power, is the frequency of power spectrum curve position of centre of gravity, and the spectral change of underload motion is had compared with hypersensitivity; F is the frequency of electromyographic signal; P (f) is power spectrum function;

(2) brain fag index

Based on wavelet packet decomposition algorithm, the conversion of employing Binary Scale, EEG signals f (t) is decomposed into 4 layers, obtain EEG signals low frequency sub-band, by wavelet package reconstruction, obtaining slow wave is the frequency band rhythm and pace of moving things at 4～8Hz place, obtaining fast wave is the frequency band rhythm and pace of moving things at 12～32Hz place, wherein slow wave is θ ripple, and fast wave is β ripple, further tries to achieve the energy Ratios of θ ripple and β ripple;

Concrete steps are as follows:

1. WAVELET PACKET DECOMPOSITION

$f\left(t\right)=\underset{i=0}{\overset{2^j-1}{\mathrm{\Σ}}}{f}_{j,i}\left(t_i\right)=f_{j,0}\left(t_0\right)+f_{j,1}\left(t_1\right)+...+f_{j,2^j-1}\left(t_{2^j-1}\right)$

In formula, i=0,1,2 ..., 2 ^j-1, f _j,i(t _i) be the reconstruct EEG signals of WAVELET PACKET DECOMPOSITION on j node layer (j, i);

2. wavelet packet decomposition computation formula by Parseval theorem and 1., the energy spectrum that can calculate EEG signals f (t) WAVELET PACKET DECOMPOSITION is:

$E_{j,i}\left(t_i\right)=\∫{\left|f_{j,i}\left(t_i\right)\right|}^2\mathrm{dt}=\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{i,\mathrm{\π}}\right|}^2$

In formula, E _j,i(t _i) be that EEG signals f (t) WAVELET PACKET DECOMPOSITION arrives the frequency band energy on node (j, i); x _{i, π}(i=0,2 ..., 2 ^j-1; π=1,2 ..., n) be reconstruct EEG signals f _j,i(t _i) discrete point amplitude; N is that signal sampling is counted;

3. ask for brain fag index:

Table 1

According to table 1, to wavelet packet subband (4,1) reconstruct, obtain 4～8Hz rhythm and pace of moving things, be θ ripple, defining its energy is E _θ, by $E_{j,i}\left(t_i\right)=\∫{\left|f_{j,i}\left(t_i\right)\right|}^2\mathrm{dt}=\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{i,\mathrm{\π}}\right|}^2$ Known:

$E_{\mathrm{\θ}}=E_{\mathrm{4,1}}\left(t_1\right)=\∫{\left|f_{\mathrm{4,1}}\left(t_1\right)\right|}^2\mathrm{dt}=\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{1,\mathrm{\π}}\right|}^2$

In like manner, to wavelet packet subband (4,3), (4,4), (4,5), (4,6), (4,7) reconstruct, obtain 12～32Hz rhythm and pace of moving things, be β ripple, defining its energy is E _β, equally by known:

$E_{\mathrm{\β}}=\underset{i=3}{\overset{7}{\mathrm{\Σ}}}{E}_{4,i}\left(t_i\right)=\underset{i=3}{\overset{7}{\mathrm{\Σ}}}\∫{\left|f_{\mathrm{4,1}}\left(t_1\right)\right|}^2\mathrm{dt}=\underset{i=3}{\overset{7}{\mathrm{\Σ}}}\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{i,\mathrm{\π}}\right|}^2$

Definition brain fag index is F _θ/β,

By analyzing the brain fag index F of certain patients with cerebral apoplexy convalescent period C3 passage _θ/βwith the situation of change of rehabilitation training time, the known increase along with movement time, F _θ/βpresent ascendant trend, the process that this and adult change from normal condition to fatigue state, the slow wave of EEG signals (θ ripple) increases gradually, and it is consistent that fast wave (β ripple) reduces gradually.Therefore, by F _θ/βquantitative assessment for brain fag state, realizes the difficulty or ease adjustment of rehabilitation training scene.

Described virtual scene design part, under computer environment, based on Visual C#2010 design virtual scene; In virtual scene display window in computer, be provided with control knob, described control knob comprises START button, " rehabilitation training " button, " physical signs " button, " preservation " button, the Close button; START button is brain myoelectricity synchronous acquisition button; " rehabilitation training " button is rehabilitation training start button; " physical signs " button is myoelectricity motion index iEMG, electromyographic signal fatigue index MPF and brain fag index F _θ/βreal-time the Show Button; " preservation " button is myoelectricity motion index iEMG, electromyographic signal fatigue index MPF and brain fag index F _θ/βsave button; The Close button is rehabilitation training conclusion button; The task platform that the rehabilitation training scene of virtual scene is rehabilitation training.

Described rehabilitation training experimental section, patient, according to the requirement of rehabilitation training scene in virtual scene design, completes appointed task by arm bending, stretching, extension and the left and right virtual palm that waves to control in rehabilitation training scene.

The present invention also provides a kind of patients with cerebral apoplexy recovery training method based on brain myoelectricity and virtual scene, first described method gathers patient's EEG signals and electromyographic signal and carries out signal processing, adopt electromyographic signal motion index to be intended to for identifying patient's motion as characteristic vector, in support vector machines-1 training before electromyographic signal motion index iEMG is sent into, the bending of identification rehabilitation arm, stretch and a left side, the action that wave on the right side, according to recognition result, drive the virtual palm of rehabilitation training scene the inside, make it drag corresponding article (all kinds of fruit under virtual environment) and arrive assigned address (fruits basket under virtual environment), complete the rehabilitation training project of appointment.

During rehabilitation training, consider MPF and F _θ/βcan reflect respectively patient moving muscle fatigue degree and motion brain district degree of fatigue, by the corresponding MPF of rehabilitation and F _θ/βmaximum average D the grade (get D=6, D can regulate according to the different rehabilitation stage) of dividing respectively, the fatigue state that different brackets correspondence is different, higher grade degree of fatigue is larger, otherwise degree of fatigue is less.Every a cycle of training, obtain electromyographic signal fatigue index and brain fag index, support vector machines-2 that train before sending into equally, identify the fatigue state of this cycle of training, according to the tired grade of identification, trigger rehabilitation training scene difficulty level control, self adaptation regulates the complexity of rehabilitation training, thereby realizes best rehabilitation training effect.

Work process is roughly as follows:

First, in conjunction with the needs of patients with cerebral apoplexy late rehabilitation training and physiatrician's suggestion, based on Visual C#2010 Independent Development Design the rehabilitation training scene of applicable patients with cerebral apoplexy late rehabilitation training; Secondly, brain fag index F is proposed _θ/β, realize the quantitative assessment of patient moving brain district fatigue in during rehabilitation training; Again, the motion of obtaining patient by extracting arm different motion pattern lower surface electromyographic signal feature is intended to, and realizes the control to virtual rehabilitation scene, completes appointment training program; Finally, by extracting the comprehensive characteristics of the caused muscle fatigue of rehabilitation training, brain fag, obtain the fatigue state of current rehabilitation, and adaptive adjusting rehabilitation training scene, slow down or strengthen rehabilitation training intensity, the secondary damage of avoiding improper training to cause, makes that rehabilitation training is more intelligent, hommization.

Compared with prior art, tool of the present invention has the following advantages:

1, the action trend that shows patient on computer display that can be visual and clear, has good interest, realizes the man-machine interaction feedback mechanism of virtual scene, improves initiative and the self-confidence of Rehabilitation training;

2, can monitor at any time patient's muscle fatigue and brain fag situation, automatically regulate the complexity of rehabilitation training according to tired classification results, slow down or strengthen the intensity of rehabilitation training, the secondary damage of avoiding improper training to cause, realizes best training effect;

3, make that rehabilitation training is more intelligent, hommization, safe, promote the clinical practice process of virtual reality technology, alleviate the present situation of physiatrician's shortage, healing robot supplemental training deficiency, there is important economy and social value.

Accompanying drawing explanation

Fig. 1 is the structural representation sketch that the present invention is based on the patients with cerebral apoplexy rehabilitation training system of brain myoelectricity and virtual scene.

Fig. 2 is the brain wave acquisition electrode cap channel position figure that the present invention is based on the patients with cerebral apoplexy rehabilitation training system of brain myoelectricity and virtual scene.

Fig. 3 is the virtual scene display surface chart that the present invention is based on the patients with cerebral apoplexy rehabilitation training system of brain myoelectricity and virtual scene.

Fig. 4 is the rehabilitation training scene display interface figure of different fatigue degree that the present invention is based on the patients with cerebral apoplexy rehabilitation training system of brain myoelectricity and virtual scene.

Fig. 5 is the 4 layers of decomposition result figure of EEG signals wavelet packet that the present invention is based on the patients with cerebral apoplexy rehabilitation training system of brain myoelectricity and virtual scene.

Fig. 6 is the brain fag index F that the present invention is based on the patients with cerebral apoplexy rehabilitation training system of brain myoelectricity and virtual scene _θ/βcurve chart with training time variation.

Drawing reference numeral: 1-electrode for encephalograms cap passage F3, 2-electrode for encephalograms cap passage F4, 3-electrode for encephalograms cap channel C 3, 4-electrode for encephalograms cap channel C 4, 5-brain electricity reference electrode A1, 6-brain electricity reference electrode A2, 7-START button, 8-" rehabilitation training " button, 9-" physical signs " button, 10-" preservation " button, 11-the Close button, 12-eeg data waveform, 13-rehabilitation training scene, 14-fruits basket, 15-fruit figure (for example: Fructus Musae, Fructus Mali pumilae, pears), the virtual palm of 16-, 17-data channel.

The specific embodiment

Below in conjunction with the specific embodiment and accompanying drawing, the present invention is described in further detail:

In structural representation sketch of the present invention as shown in Figure 1, system of the present invention comprises signals collecting part, data preprocessing part, motion index Extraction parts, exercise fatigue index extraction part, virtual scene design part and rehabilitation training experimental section; Wherein,

Described signals collecting part is extracted patient's EEG signals and electromyographic signal;

Described data preprocessing part carries out Filtering Processing to the EEG signals collecting and electromyographic signal;

Described motion index Extraction parts is that the quantitative analysis of discharging during to patient's muscle movement is extracted;

Described exercise fatigue index extraction is partly that patient's EEG signals and electromyographic signal are analyzed respectively to extraction, and then judgement exercise fatigue or brain fag;

Described virtual scene design be partly under computer environment based on Visual C#2010 development and Design, generating virtual scene show output, sets up man-machine interaction feedback mechanism on computers;

Described rehabilitation training experimental section be patient according to rehabilitation training requirement in virtual scene, by arm, bend, stretch with left and right waving and control virtual palm in virtual scene and complete appointed task.

Concrete steps are as follows:

Step 1: signals collecting part

The collection of electromyographic signal adopts the differential input of bikini, two differential input ends that be myoelectricity wherein, and another one be reference, difference input electrode, along muscle fiber direction, is placed on belly of muscle place; The muscle the present invention relates to is biceps brachii m. and triceps brachii, first uses the skin at the tested position of alcohol wipe, to remove skin surface oils and fats and scurf, adhesive electrode.Conducting wire is suitably fixed and reduced the interference that in course of action, conducting wire rocks as far as possible.

With reference to figure 2 brain wave acquisition electrode cap channel position figure.Eeg signal acquisition adopts 8 passage brain myoelectricity synchronous acquisition instrument to gather, and 10～20 electrodes of adopting international standards are placed standard, by electrode cap, electrode are connected with scalp.Because the present invention is the brain fag index that extracts rehabilitation, Representative Region is sensorimotor cortex, therefore selects C3, the C4 in electrode cap and represents that the F3 of front motor region, F4 place gather.Adopt the single-stage method of leading, electrode slice A1, electrode slice A2 lead and are connected respectively to after the ear of left and right mastoid process as with reference to electrode, and ground electrode arrangement overhead hits exactly.

Step 2: data preprocessing part

Use self adaptation high pass filter and self adaptation 50Hz notch filter wave filter respectively EEG signals, electromyographic signal to be carried out to Filtering Processing, baseline drift and the power frequency removed in signal are disturbed; Re-use the logical FIR wave filter of Butterworth three rank bands EEG signals, electromyographic signal are processed, according to effective frequency range feature of signal, choosing electromyographic signal cut-off frequency is 2Hz～200Hz, and the cut-off frequency of choosing EEG signals is 2Hz～50Hz.

Step 3: motion index Extraction parts

Electromyographic signal motion index is as follows,

$\mathrm{iEMG}=\underset{t}{\overset{t+T}{\∫}}\left|\mathrm{EMG}\left(t\right)\right|\mathrm{dt}$

In formula, iEMG is integration myoelectricity value, the quantity of moving cell and the discharge magnitude of each moving cell during reflection muscle movement; T is for gathering the time of electromyographic signal; T is for analyzing the cycle of the electromyographic signal collecting; The electromyographic signal of the respective muscle motion that EMG (t) collects constantly for t.

Step 4: exercise fatigue index extraction part

Comprise electromyographic signal fatigue index and brain fag index;

(1) electromyographic signal fatigue index as shown in the formula,

$\mathrm{MPF}=\underset{0}{\overset{\∞}{\∫}}f\·P\left(f\right)\mathrm{df}/\underset{0}{\overset{\∞}{\∫}}P\left(f\right)\mathrm{df}$

In formula, MPF is frequency of average power, is the frequency of power spectrum curve position of centre of gravity, and the spectral change of underload motion is had compared with hypersensitivity; F is the frequency of electromyographic signal; P (f) is power spectrum function;

(2) brain fag index, based on wavelet packet decomposition algorithm, adopt Binary Scale conversion, EEG signals f (t) is decomposed into 4 layers (as shown in table 1), obtain EEG signals low frequency sub-band, by wavelet package reconstruction, obtaining slow wave is the frequency band rhythm and pace of moving things at 4～8Hz place, and obtaining fast wave is the frequency band rhythm and pace of moving things at 12～32Hz place, wherein slow wave is θ ripple, fast wave is β ripple, further tries to achieve the energy Ratios of θ ripple and β ripple, and concrete steps are as follows:

1. WAVELET PACKET DECOMPOSITION

$f\left(t\right)=\underset{i=0}{\overset{2^j-1}{\mathrm{\Σ}}}{f}_{j,i}\left(t_i\right)=f_{j,0}\left(t_0\right)+f_{j,1}\left(t_1\right)+...+f_{j,2^j-1}\left(t_{2^j-1}\right)$

In formula, i=0,1,2 ..., 2 ^j-1, f _j,i(t _i) be the reconstruct EEG signals of WAVELET PACKET DECOMPOSITION on j node layer (j, i);

2. wavelet packet decomposition computation formula by Parseval theorem and 1., the energy spectrum that can calculate EEG signals f (t) WAVELET PACKET DECOMPOSITION is:

$E_{j,i}\left(t_i\right)=\∫{\left|f_{j,i}\left(t_i\right)\right|}^2\mathrm{dt}=\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{i,\mathrm{\π}}\right|}^2$

In formula, E _j,i(t _i) be that EEG signals f (t) WAVELET PACKET DECOMPOSITION arrives the frequency band energy on node (j, i); x _{i, π}(i=0,2 ..., 2 ^j-1; π=1,2 ..., n) be reconstruct EEG signals f _j,i(t _i) discrete point amplitude; N is that signal sampling is counted;

3. ask for brain fag index

According to table 1, to wavelet packet subband (4,1) reconstruct, obtain 4～8Hz rhythm and pace of moving things, be θ ripple, defining its energy is E _θ, by $E_{j,i}\left(t_i\right)=\∫{\left|f_{j,i}\left(t_i\right)\right|}^2\mathrm{dt}=\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{i,\mathrm{\π}}\right|}^2$ Known:

$E_{\mathrm{\θ}}=E_{\mathrm{4,1}}\left(t_1\right)=\∫{\left|f_{\mathrm{4,1}}\left(t_1\right)\right|}^2\mathrm{dt}=\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{1,\mathrm{\π}}\right|}^2$

In like manner, to wavelet packet subband (4,3), (4,4), (4,5), (4,6), (4,7) reconstruct, obtain 12～32Hz rhythm and pace of moving things, be β ripple, defining its energy is E _β, equally by known:

$E_{\mathrm{\β}}=\underset{i=3}{\overset{7}{\mathrm{\Σ}}}{E}_{4,i}\left(t_i\right)=\underset{i=3}{\overset{7}{\mathrm{\Σ}}}\∫{\left|f_{\mathrm{4,1}}\left(t_1\right)\right|}^2\mathrm{dt}=\underset{i=3}{\overset{7}{\mathrm{\Σ}}}\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{i,\mathrm{\π}}\right|}^2$

Definition brain fag index is F _θ/β,

With reference to figure 6, Fig. 6 is the brain fag index F of certain patients with cerebral apoplexy convalescent period C3 passage _θ/βwith the situation of change of rehabilitation training time, the known increase along with movement time, F _θ/βpresent ascendant trend, the process that this and adult change from normal condition to fatigue state, the slow wave of EEG signals (θ ripple) increases gradually, and it is consistent that fast wave (β ripple) reduces gradually.Therefore, the present invention is by F _θ/βquantitative assessment for brain fag state, realizes the difficulty or ease adjustment of rehabilitation training scene.

Step 5: virtual scene design part

With reference to figure 3, Fig. 3 is virtual scene display surface chart of the present invention, is in conjunction with the needs of patients with cerebral apoplexy late rehabilitation training and physiatrician's suggestion, under computer environment, and the virtual scene based on Visual C#2010 design; Control knob in computer in virtual scene display window comprises START button 1, " rehabilitation training " button 2, " physical signs " button 3, " preservation " button 4, the Close button 5; START button is brain myoelectricity synchronous acquisition button; " rehabilitation training " button is rehabilitation training start button; " physical signs " button is real-time the Show Button of myoelectricity motion index, electromyographic signal fatigue index and brain fag index; " preservation " button is the save button of myoelectricity motion index, electromyographic signal fatigue index and brain fag index; The Close button is rehabilitation training conclusion button; The task platform that the rehabilitation training scene of virtual scene is rehabilitation training, comprising: fruits basket 14, fruit (as: Fructus Musae) 15, virtual palm 16.

With reference to figure 4 (a～f), Fig. 4 (a～f) is for designing different rehabilitation training scenes for rehabilitation different fatigue state, its Scene is reduced to the difficulty of Fig. 4 (f) gradually by Fig. 4 (a), show as along with patient's degree of fatigue changes from small to big, corresponding rehabilitation training scene difficulty is become different by difficulty, be embodied in when the degree of fatigue of patients with cerebral apoplexy increases gradually, patient bends by arm, stretch and a left side, the right side is waved and is controlled the upper of virtual palm, lower and left, move right, complete various fruit is all put into this task process of basket, the kind of fruit and number are reduced to Fig. 4 (f) gradually by Fig. 4 (a), be that rehabilitation training intensity weakens gradually along with the increase of degree of fatigue.Step 6: rehabilitation training experimental section

Patient, according to the requirement of rehabilitation training scene in virtual scene design, completes appointed task by arm bending, stretching, extension and the left and right virtual palm that waves to control in rehabilitation training scene.

Concrete grammar is:

First gather patient's EEG signals and electromyographic signal and carry out signal processing, the size of muscular strength and movement velocity trend in the time of reflecting muscular movement to a certain extent due to electromyographic signal motion index iEMG, adopt iEMG to be intended to for identifying patient's motion as characteristic vector, after asking for iEMG, support vector machines-1 training before sending into, bending of identification rehabilitation arm, stretch and a left side, the action that wave on the right side, and according to recognition result, drive the virtual palm control rehabilitation training scene the inside, make it drag corresponding article (being all kinds of fruit in virtual scene display interface) and arrive assigned address (fruits basket), and then complete the appointment rehabilitation training project that rehabilitation scene requires.Meanwhile, consider electromyographic signal fatigue index MPF and brain fag index F _θ/βcan reflect respectively patient moving muscle fatigue degree and motion brain district degree of fatigue, by the corresponding MPF of rehabilitation and F _θ/βmaximum respectively on average divide 6 grades, the fatigue state that different brackets is corresponding different, higher grade degree of fatigue is larger, on the contrary degree of fatigue is less, wherein corresponding diagram 4 (a～f) is distinguished in class 6～1, i.e. the rehabilitation training scene of the corresponding different difficulty of different fatigue grade.Every a cycle of training, obtain electromyographic signal fatigue index MPF and brain fag index F _θ/βand the support vector SVM-2 training before sending into, identify the fatigue state of this cycle of training, and according to the tired grade W (W ∈ 1～6) of identification, trigger rehabilitation training scene difficulty level control, choose corresponding rehabilitation training scene M (M ∈ a～f), thereby the complexity of adaptive adjusting rehabilitation training, and then realize best rehabilitation training effect.

Above-described embodiment is described the preferred embodiment of the present invention; not scope of the present invention is limited; design under the prerequisite of spirit not departing from the present invention; various distortion and improvement that those of ordinary skills make technical scheme of the present invention, all should fall in the definite protection domain of the claims in the present invention book.

Claims (8)

1. the patients with cerebral apoplexy rehabilitation training system based on brain myoelectricity and virtual scene, is characterized in that: described system comprises signals collecting part, data preprocessing part, motion index Extraction parts, exercise fatigue index extraction part, virtual scene design part and rehabilitation training experimental section; Wherein,

Described signals collecting part is extracted patient's EEG signals and electromyographic signal;

Described data preprocessing part carries out Filtering Processing to the EEG signals collecting and electromyographic signal;

Described motion index Extraction parts is that the quantitative analysis of discharging during to patient's muscle movement is extracted;

Described exercise fatigue index extraction is partly that patient's EEG signals and electromyographic signal are analyzed respectively to extraction, and then judgement exercise fatigue or brain fag;

Described virtual scene design be partly under computer environment based on Visual C#2010 development and Design, generating virtual scene show output, sets up man-machine interaction feedback mechanism on computers;

Described rehabilitation training experimental section be patient according to rehabilitation training requirement in virtual scene, by arm, bend, stretch with left and right waving and control virtual palm in virtual scene and complete appointed task.

2. the patients with cerebral apoplexy rehabilitation training system based on brain myoelectricity and virtual scene according to claim 1, it is characterized in that: in described signals collecting part, the collection of electromyographic signal adopts the differential input of bikini, two differential input ends that are myoelectricity wherein, another one is with reference to ground, difference input electrode, along muscle fiber direction, is placed on belly of muscle place; Eeg signal acquisition adopts 8 passage brain myoelectricity synchronous acquisition instrument to gather, 10～20 electrodes of adopting international standards are placed standards, by electrode cap, electrode is connected with scalp, adopt the single-stage method of leading, reference electrode leads and is connected respectively to mastoid process place after the ear of left and right, and ground electrode arrangement overhead hits exactly.

3. the patients with cerebral apoplexy rehabilitation training system based on brain myoelectricity and virtual scene according to claim 1, it is characterized in that: described data preprocessing part, use self adaptation high pass filter and self adaptation 50Hz notch filter wave filter respectively EEG signals, electromyographic signal to be carried out to Filtering Processing, baseline drift and the power frequency removed in signal are disturbed; Re-use the logical FIR wave filter of Butterworth three rank bands EEG signals, electromyographic signal are processed, according to effective frequency range feature of signal, the cut-off frequency of choosing electromyographic signal is 2Hz～200Hz, and the cut-off frequency of choosing EEG signals is 2Hz～50Hz.

4. the patients with cerebral apoplexy rehabilitation training system based on brain myoelectricity and virtual scene according to claim 1, is characterized in that: described motion index Extraction parts, and electromyographic signal motion index is as follows,

$\mathrm{iEMG}=\underset{t}{\overset{t+T}{\∫}}\left|\mathrm{EMG}\left(t\right)\right|\mathrm{dt}$

In formula, iEMG is integration myoelectricity value, the quantity of moving cell and the discharge magnitude of each moving cell during reflection muscle movement; T is for gathering the time of electromyographic signal; T is for analyzing the cycle of the electromyographic signal collecting; The electromyographic signal of the respective muscle motion that EMG (t) collects constantly for t.

5. the patients with cerebral apoplexy rehabilitation training system based on brain myoelectricity and virtual scene according to claim 1, is characterized in that: described exercise fatigue index extraction partly comprises electromyographic signal fatigue index and brain fag index;

(1) electromyographic signal fatigue index as shown in the formula,

$\mathrm{MPF}=\underset{0}{\overset{\∞}{\∫}}f\·P\left(f\right)\mathrm{df}/\underset{0}{\overset{\∞}{\∫}}P\left(f\right)\mathrm{df}$

In formula, MPF is frequency of average power, is the frequency of power spectrum curve position of centre of gravity, and the spectral change of underload motion is had compared with hypersensitivity; F is the frequency of electromyographic signal; P (f) is power spectrum function;

(2) brain fag index, based on wavelet packet decomposition algorithm, adopt Binary Scale conversion, EEG signals f (t) is decomposed into 4 layers, obtain EEG signals low frequency sub-band, by wavelet package reconstruction, obtaining slow wave is the frequency band rhythm and pace of moving things at 4～8Hz place, and obtaining fast wave is the frequency band rhythm and pace of moving things at 12～32Hz place, wherein slow wave is θ ripple, fast wave is β ripple, further tries to achieve the energy Ratios of θ ripple and β ripple, and concrete steps are as follows:

1. WAVELET PACKET DECOMPOSITION

$f\left(t\right)=\underset{i=0}{\overset{2^j-1}{\mathrm{\Σ}}}{f}_{j,i}\left(t_i\right)=f_{j,0}\left(t_0\right)+f_{j,1}\left(t_1\right)+...+f_{j,2^j-1}\left(t_{2^j-1}\right)$

In formula, i=0,1,2 ..., 2 ^j-1, f _j,i(t _i) be the reconstruct EEG signals of WAVELET PACKET DECOMPOSITION on j node layer (j, i);

2. wavelet packet decomposition computation formula by Parseval theorem and 1., the energy spectrum that can calculate EEG signals f (t) WAVELET PACKET DECOMPOSITION is:

$E_{j,i}\left(t_i\right)=\∫{\left|f_{j,i}\left(t_i\right)\right|}^2\mathrm{dt}=\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{i,\mathrm{\π}}\right|}^2$

In formula, E _j,i(t _i) be that EEG signals f (t) WAVELET PACKET DECOMPOSITION arrives the frequency band energy on node (j, i); x _{i, π}(i=0,2 ..., 2 ^j-1; π=1,2 ..., n) be reconstruct EEG signals f _j,i(t _i) discrete point amplitude; N is that signal sampling is counted;

3. ask for brain fag index

Table 1

According to table 1, to wavelet packet subband (4,1) reconstruct, obtain 4～8Hz rhythm and pace of moving things, be θ ripple, defining its energy is E _θ, by $E_{j,i}\left(t_i\right)=\∫{\left|f_{j,i}\left(t_i\right)\right|}^2\mathrm{dt}=\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{i,\mathrm{\π}}\right|}^2$ Known:

$E_{\mathrm{\θ}}=E_{\mathrm{4,1}}\left(t_1\right)=\∫{\left|f_{\mathrm{4,1}}\left(t_1\right)\right|}^2\mathrm{dt}=\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{1,\mathrm{\π}}\right|}^2$

In like manner, to wavelet packet subband (4,3), (4,4), (4,5), (4,6), (4,7) reconstruct, obtain 12～32Hz rhythm and pace of moving things, be β ripple, defining its energy is E _β, equally by known:

$E_{\mathrm{\β}}=\underset{i=3}{\overset{7}{\mathrm{\Σ}}}{E}_{4,i}\left(t_i\right)=\underset{i=3}{\overset{7}{\mathrm{\Σ}}}\∫{\left|f_{\mathrm{4,1}}\left(t_1\right)\right|}^2\mathrm{dt}=\underset{i=3}{\overset{7}{\mathrm{\Σ}}}\underset{\mathrm{\π}=1}{\overset{n}{\mathrm{\Σ}}}{\left|x_{i,\mathrm{\π}}\right|}^2$

Definition brain fag index is F _θ/β,

6. the patients with cerebral apoplexy rehabilitation training system based on brain myoelectricity and virtual scene according to claim 1, is characterized in that: described virtual scene design part, under computer environment, based on Visual C#2010 design virtual scene; In virtual scene display window in computer, be provided with control knob, described control knob comprises START button, " rehabilitation training " button, " physical signs " button, " preservation " button, the Close button; START button is brain myoelectricity synchronous acquisition button; " rehabilitation training " button is rehabilitation training start button; " physical signs " button is real-time the Show Button of myoelectricity motion index, electromyographic signal fatigue index and brain fag index; " preservation " button is the save button of myoelectricity motion index, electromyographic signal fatigue index and brain fag index; The Close button is rehabilitation training conclusion button; The task platform that the rehabilitation training scene of virtual scene is rehabilitation training.

7. the patients with cerebral apoplexy rehabilitation training system based on brain myoelectricity and virtual scene according to claim 1, it is characterized in that: described rehabilitation training experimental section, patient, according to the requirement of rehabilitation training scene in virtual scene design, completes appointed task by arm bending, stretching, extension and the left and right virtual palm that waves to control in rehabilitation training scene.

8. the patients with cerebral apoplexy recovery training method based on brain myoelectricity and virtual scene, it is characterized in that: first described method gathers patient's EEG signals and electromyographic signal and carry out signal processing, adopt electromyographic signal motion index to be intended to for identifying patient's motion as characteristic vector, in support vector machines-1 training before electromyographic signal motion index is sent into, the bending of identification rehabilitation arm, stretch and a left side, the action that wave on the right side, according to recognition result, drive the virtual palm of rehabilitation training scene the inside, make it drag corresponding article and arrive assigned address, complete rehabilitation training project, in during rehabilitation training, the maximum of patient's electromyographic signal fatigue index and brain fag index is on average divided respectively to a plurality of grades, the fatigue state that different brackets is corresponding different, higher grade degree of fatigue is larger, otherwise less, every a cycle of training, obtain electromyographic signal fatigue index and brain fag index and send into support vector machines-2 that train before, identify the fatigue state of this cycle of training, according to the tired grade of identification, trigger rehabilitation training scene difficulty level control, self adaptation regulates the complexity of rehabilitation training.