CN102722255A - Neurofeedback-based motor imagery brain computer interface (BCI) interactive training system and method - Google Patents
Neurofeedback-based motor imagery brain computer interface (BCI) interactive training system and method Download PDFInfo
- Publication number
- CN102722255A CN102722255A CN2012102212301A CN201210221230A CN102722255A CN 102722255 A CN102722255 A CN 102722255A CN 2012102212301 A CN2012102212301 A CN 2012102212301A CN 201210221230 A CN201210221230 A CN 201210221230A CN 102722255 A CN102722255 A CN 102722255A
- Authority
- CN
- China
- Prior art keywords
- training
- eeg signals
- computing machine
- experimenter
- computer interface
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Images
Abstract
The invention discloses a neurofeedback-based motor imagery brain computer interface (BCI) interactive training system and a neurofeedback-based motor imagery brain computer interface (BCI) interactive training method. The interactive training system comprises electrode caps, a plurality of electrodes, signal processing equipment and a computer; and an interactive training method of the interactive training system comprises the following steps of: 1, acquiring an electroencephalogram signal of motor imagery of a subject; 2, amplifying and filtering the electroencephalogram signal; 3, performing human training; and 4, performing system training. By the interactive training system and the interactive training method, the bidirectional training process of the subject and the training system is realized, the training efficiency is improved, the training time is greatly shortened, and the classification accuracy is improved.
Description
Technical field
The present invention relates to a kind of training system and training method of information interaction system, particularly a kind of motion imagination brain-computer interface interactive training system and method based on neural feedback.
Background technology
Brain-computer interface (BCI) provides a kind of new passage that exchanges with the external world, and it is the brand-new information interaction system that does not rely on the normal output channel of brain (peripheral neverous system and musculature).When carrying out the limb motion imagination, the relevant phenomenon that desynchronizes of incident can appear in corresponding motor cortex zone, and (Event Related Desynchronization, ERD), this phenomenon is used to set up the brain-computer interface system of the motion imagination.The basis of this type brain-computer interface need be tried effectively to carry out the modulation of EEG signals, so need set up the method that corresponding training system helps tried to grasp the modulation EEG signals.Training is a two-way process, and one side needs training to be tried to learn how to modulate EEG signals.For the training that is tried, let the experimenter learn how to adjust the brain signal of oneself through feedback system usually.Because different brain electricity conditions can produce different results and then can produce different feedback signals, thus the experimenter according to current feedback signal come adaptive learning how freely the control system feedback reach the purpose of control EEG signals.But this training process needs long-term training such as some months time usually, and is consuming time long, and the discrimination of system is low excessively.
Summary of the invention
The purpose of this invention is to provide a kind of motion imagination brain-computer interface interactive training method, realized the two-way training process of experimenter and training system, improved training effectiveness, shortened the training time greatly, and improved classify accuracy based on neural feedback.
In order to realize above purpose, the present invention realizes through following technical scheme:
A kind of motion imagination brain-computer interface interactive training system based on neural feedback comprises:
Electrode cap;
Several electrodes, described several electrodes are arranged in the described electrode cap, and on the head that electrode cap is worn on the experimenter, these several electrodes can receive the EEG signals of the motion imagination;
Signal handling equipment, the input end of described signal handling equipment links to each other with electrode through circuit, and this signal handling equipment amplifies and the processing EEG signals that electrode collected;
Computing machine, described input end and computer links to each other with the output terminal of signal amplifying apparatus through data line, and the interface of the training software in this computing machine can show that experimenter's motion imagination signal is through the sorted feedback diagram as a result of training software.
The number of described electrode is 13.
A kind of motion imagination brain-computer interface interactive training method based on neural feedback is characterized in that, comprises following steps:
Step 1: collect the EEG signals that the experimenter moves and imagines through several electrodes in the electrode cap, these EEG signals are transferred to signal handling equipment;
Step 2: signal handling equipment carries out processing and amplifying to these EEG signals, and these EEG signals are carried out filtering;
Step 3: will pass through filtered EEG signals input computing machine; Computing machine carries out feature extraction and classification to filtered EEG signals; And the information of the imagination intensity of the EEG signals that this motion of acquisition is imagined from sorting result; Convert this strength information to image feedback in real time again and show on computers, to let the experimenter in time know the quality of the current motion imagination;
Step 4: computing machine is followed the trail of the EEG signals of the different motion imagination of experimenter, and selects features training according to qualifications and upgrade disaggregated model with different weights through the output parameter of dynamic adjustment sorter.
The present invention compared with prior art has the following advantages:
1, realizes the two-way training process of experimenter and training system, improved training effectiveness, shortened the training time greatly;
2, improved classify accuracy.
Description of drawings
Fig. 1 is a kind of system chart of imagining the brain-computer interface interactive training system based on the motion of neural feedback of the present invention;
Fig. 2 is a kind of systematic training schematic diagram of imagining brain-computer interface interactive training method based on the motion of neural feedback of the present invention.
Embodiment
Below in conjunction with accompanying drawing,, the present invention is done further elaboration through specifying a preferable specific embodiment.
As shown in Figure 1, a kind of motion imagination brain-computer interface interactive training system based on neural feedback comprises: electrode cap 1, several electrodes 2, signal handling equipment 3, computing machine 4; Wherein, several electrodes 2 are arranged in the electrode cap 1, in the present embodiment; Electrode cap 1 is selected 64 electrode caps that lead for use; The number of electrode 2 is 13, and on the head that electrode cap 1 is worn on the experimenter, 13 electrodes 2 can receive the EEG signals of the motion imagination.
The input end of signal handling equipment 3 links to each other with electrode 2 through circuit, these EEG signals that signal handling equipment 3 amplifies and processing electrode 2 is collected.
The input end of computing machine 4 links to each other with the output terminal of signal amplifying apparatus 3 through data line, and the interface of the training software in this computing machine 4 can show that experimenter's motion imagination signal is through the sorted feedback diagram as a result of training software.
Utilize the above-mentioned motion imagination brain-computer interface interactive training system based on neural feedback, the interactive training method that is adopted comprises following steps:
Step 1: collect the EEG signals that the experimenter moves and imagines through 13 electrodes 2 in the electrode cap 1, these EEG signals are transferred to signal handling equipment 3; In the present embodiment, electrode 2 is to adopt 256Hz to carry out data sampling.
Step 2: 3 pairs of these EEG signals of signal handling equipment carry out processing and amplifying, and these EEG signals are carried out filtering; In the present embodiment, adopt 5 ~ 30Hz bandpass filtering.
Step 3: will pass through the EEG signals input computing machine 4 behind the bandpass filtering; 4 pairs of EEG signals of computing machine carry out feature extraction and classification; And from sorting result, obtain the information of the imagination intensity of these EEG signals; Convert this strength information to image feedback in real time again and on computing machine 4, show,, promptly carry out people's training to let the experimenter in time know the quality of the current motion imagination.
In the present embodiment, imagine that with the experimenter left hand motion is example and representes that through green arrow the real-time images displayed in the interface of the training software in the computing machine 4 comprises several tanks; In the present embodiment, two tanks are set, wherein; If the length of each tank is L, a bead is arranged in each tank, the once imagination task of every completion needs 3s; Divide feedback 10 times, every 0.3s feeds back once, promptly corresponding 10 feedbacks of the once imagination task of each tank correspondence.
If the imagination is correct, the fluid column in the tank then can be compared the current liquid-column height segment distance that rises, thereby bead is upwards promoted a segment distance; Otherwise bead then can move down corresponding distance, and the distance that bead moves is represented with H; Can calculate H according to formula, i.e. H=P * 0.3s * L/3s.In the process of feedback,, suppose three differing heights if the fluid column in each tank arrives different height; Be respectively the 1/2L of tank, when 3/4L and L, then the fluid column color in the tank also can present various colors; In the present embodiment, once imagine task termination after, also can be according to the height of fluid column; Corresponding mark occurs, encourage accordingly to give the experimenter.Wherein, P is the classification output probability, this classification output probability be the current EEG signals of experimenter according to the sorted output probability of SVMs (SVM) disaggregated model, what P can be along with characteristic and disaggregated model is different and different, specifically sees also step 4.
Step 4: the EEG signals of the motion imagination that 4 couples of experimenters of computing machine are different are followed the trail of, and select features training according to qualifications and upgrade disaggregated model with different weights through the output parameter of dynamic adjustment sorter, promptly carry out systematic training.
As shown in Figure 2; In the present embodiment, training system adopts is a kind of training of the disaggregated model based on feature selecting, in the process of train classification models; The characteristic of each disaggregated model training comes from the up-to-date training set (runs) that the experimenter accomplishes; Number≤5 of training set (runs), each training set (run) comprises 20 training (trials), every completion is once moved and is imagined that being completion once trains (trial).
Select the principle of trial to be among each run: to select the correct trials of all classification earlier, and be arranged in order the correct trials of all classification from high to low according to classification output probability P.Wherein, If among the trials that selects; The ratio r that imagines the number of the correct trials that representes left hand and right hand action respectively satisfies 3/5≤r≤1 or 1≤r≤5/3; All trials that selected have been considered to reach two types of levels than balance so, and the trials that the classification of then having sorted is correct forms a new run; Otherwise; Think that then the trials that has selected is considered to not reach two types of levels than balance; Therefore, the trials that classification among this run is correct carries out descending sort from high to low according to classification correct probability P, and the trials of all classification errors among this run; Then will carry out ascending order from low to high and arrange, form a new run according to classification output probability P.
The rest may be inferred, and other new run also can form in the same way, and the runs of all new formation (runs≤5) selects according to different weights then; The concrete definition as follows: when the quantity of new runs is not more than 5, select the ratio of trials to define as follows among each new runs: W (n)=X (n)/5, n={1; 2; 3,4,5}.In formula, establish the number of n for the new runs of formation, set: the value of the corresponding X of run (n) that forms with the last time is 5; Be W (n)=1; The up-to-date run that accomplishes from the experimenter is to a last run, and the value of X (n) subtracts 1 successively, and then the value of W (n) successively decreases 1/5 successively; So, selecting the number of trials among each new run is 20W (n); Therefore, when the number of runs greater than 5 the time, from up-to-date run to preceding several run the classification output probability of trial fix, promptly setting up-to-date run is newrun5, is forward newrun4 successively; Newrun3, newrun2, newrun1, then its classification output probability is respectively 1,4/5 successively; 3/5,2/5,1/5, promptly from each run, select the number of trial respectively to be 20; 16,12,8,4.Newrun5 in view of the above, newrun4, newrun3, newrun2, newrun1 forms a new model newmodel.If the number of trial is less than the number of selecting trial among each run of correspondence that designs in top ratio among the runs of each new formation; Then be as the criterion, otherwise from new runs, carry out the selection of trial successively by trial number among top each run that designs with the actual trial number among the new runs.The principle that the electrical activity of brain tried improves along with the importance of carrying out in model training of experiment is successively followed in the selection of this ratio.
In sum; Brain-computer interface interactive training system and method are imagined in a kind of motion based on neural feedback of the present invention, have realized the two-way training process of experimenter and training system, have improved training effectiveness; Shorten the training time greatly, and improved classify accuracy.
Although content of the present invention has been done detailed introduction through above-mentioned preferred embodiment, will be appreciated that above-mentioned description should not be considered to limitation of the present invention.After those skilled in the art have read foregoing, for multiple modification of the present invention with to substitute all will be conspicuous.Therefore, protection scope of the present invention should be limited appended claim.
Claims (3)
1. the motion imagination brain-computer interface interactive training system based on neural feedback is characterized in that, comprises:
Electrode cap (1);
Several electrodes (2), described several electrodes (2) are arranged in the described electrode cap (1), and on the head that electrode cap (1) is worn on the experimenter, these several electrodes (2) can receive the EEG signals of the motion imagination;
Signal handling equipment (3), the input end of described signal handling equipment (3) links to each other with electrode (2) through circuit, the EEG signals that this signal handling equipment (3) amplifies and processing electrode (2) is collected;
Computing machine (4); The input end of described computing machine (4) links to each other with the output terminal of signal amplifying apparatus (3) through data line, and the interface of the training software in this computing machine (4) can show that experimenter's motion imagination signal is through the sorted feedback diagram as a result of training software.
2. the motion imagination brain-computer interface interactive training system based on neural feedback as claimed in claim 1 is characterized in that the number of described electrode (2) is 13.
3. the motion imagination brain-computer interface interactive training method based on neural feedback is characterized in that, comprises following steps:
Step 1: collect the EEG signals that the experimenter moves and imagines through several electrodes (2) in the electrode cap (1), these EEG signals are transferred to signal handling equipment (3);
Step 2: signal handling equipment (3) carries out processing and amplifying to these EEG signals, and these EEG signals are carried out filtering;
Step 3: will pass through filtered EEG signals input computing machine (4); Computing machine (4) carries out feature extraction and classification to filtered EEG signals; And the information of the imagination intensity of the EEG signals that this motion of acquisition is imagined from sorting result; Convert this strength information to image feedback in real time again and on computing machine (4), show, to let the experimenter in time know the quality of the current motion imagination;
Step 4: computing machine (4) is followed the trail of the EEG signals of the different motion imagination of experimenter, and selects features training according to qualifications and upgrade disaggregated model with different weights through the output parameter of dynamic adjustment sorter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210221230.1A CN102722255B (en) | 2012-06-29 | 2012-06-29 | Neurofeedback-based motor imagery brain computer interface (BCI) interactive training system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210221230.1A CN102722255B (en) | 2012-06-29 | 2012-06-29 | Neurofeedback-based motor imagery brain computer interface (BCI) interactive training system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102722255A true CN102722255A (en) | 2012-10-10 |
| CN102722255B CN102722255B (en) | 2015-07-15 |
Family
ID=46948050
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210221230.1A Expired - Fee Related CN102722255B (en) | 2012-06-29 | 2012-06-29 | Neurofeedback-based motor imagery brain computer interface (BCI) interactive training system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102722255B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103054573A (en) * | 2012-12-31 | 2013-04-24 | 北京师范大学 | Multi-user neural feedback training method and multi-user neural feedback training system |
| CN105962935A (en) * | 2016-06-14 | 2016-09-28 | 中国医学科学院生物医学工程研究所 | Brain electrical nerve feedback training system and method for improving motor learning function |
| CN106371590A (en) * | 2016-08-29 | 2017-02-01 | 华南理工大学 | High-performance motor imagery online brain-computer interface system based on OpenVIBE |
| CN109907941A (en) * | 2019-04-02 | 2019-06-21 | 西安交通大学 | A kind of wrist rehabilitation control device based on focus level |
| WO2022017202A1 (en) * | 2020-07-24 | 2022-01-27 | 天津大学 | Method and apparatus for dynamic spatial filtering and amplification of electroencephalogram, electronic device, and storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923392A (en) * | 2010-09-02 | 2010-12-22 | 上海交通大学 | Asynchronous brain-computer interactive control method for EEG signal |
-
2012
- 2012-06-29 CN CN201210221230.1A patent/CN102722255B/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923392A (en) * | 2010-09-02 | 2010-12-22 | 上海交通大学 | Asynchronous brain-computer interactive control method for EEG signal |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103054573A (en) * | 2012-12-31 | 2013-04-24 | 北京师范大学 | Multi-user neural feedback training method and multi-user neural feedback training system |
| CN103054573B (en) * | 2012-12-31 | 2015-11-18 | 北京师范大学 | Many people neural feedback training method and many people neural feedback training system |
| CN105962935A (en) * | 2016-06-14 | 2016-09-28 | 中国医学科学院生物医学工程研究所 | Brain electrical nerve feedback training system and method for improving motor learning function |
| CN105962935B (en) * | 2016-06-14 | 2019-01-08 | 中国医学科学院生物医学工程研究所 | The brain electric nerve feedback training system and its method improved for motor learning function |
| CN106371590A (en) * | 2016-08-29 | 2017-02-01 | 华南理工大学 | High-performance motor imagery online brain-computer interface system based on OpenVIBE |
| CN106371590B (en) * | 2016-08-29 | 2019-06-18 | 华南理工大学 | The online brain machine interface system of high-performance Mental imagery based on OpenVIBE |
| CN109907941A (en) * | 2019-04-02 | 2019-06-21 | 西安交通大学 | A kind of wrist rehabilitation control device based on focus level |
| WO2022017202A1 (en) * | 2020-07-24 | 2022-01-27 | 天津大学 | Method and apparatus for dynamic spatial filtering and amplification of electroencephalogram, electronic device, and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102722255B (en) | 2015-07-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103876734B (en) | A kind of EEG signals feature selection approach based on decision tree | |
| CN102722255B (en) | Neurofeedback-based motor imagery brain computer interface (BCI) interactive training system and method | |
| CN101515199B (en) | Character input device based on eye tracking and P300 electrical potential of the brain electricity | |
| CN101201696B (en) | Chinese input BCI system based on P300 brain electric potential | |
| CN107822623A (en) | A kind of driver fatigue and Expression and Action method based on multi-source physiologic information | |
| KR20190035368A (en) | Method and Apparatus for Recognizing Emotion Based on Image Converted from Brain Signal | |
| CN103699226A (en) | Tri-modal serial brain-computer interface method based on multi-information fusion | |
| CN101515200B (en) | Target selecting method based on transient visual evoked electroencephalogram | |
| CN108420429A (en) | A kind of brain electricity epilepsy automatic identifying method based on the fusion of various visual angles depth characteristic | |
| CN102940490A (en) | Method for extracting motor imagery electroencephalogram signal feature based on non-linear dynamics | |
| CN107577343A (en) | It is a kind of based on the notice of haptic device and electroencephalogramsignal signal analyzing training and evaluating apparatus | |
| CN104548347A (en) | Pure idea nerve muscle electrical stimulation control and nerve function evaluation system | |
| CN107656612B (en) | Large instruction set brain-computer interface method based on P300-SSVEP | |
| CN103631941A (en) | Electroencephalogram-based target image retrieval system | |
| CN104473636A (en) | Brain fatigue network analysis method based on partial orientation coherence | |
| CN102542283A (en) | Optimal electrode assembly automatic selecting method of brain-machine interface | |
| CN106778444A (en) | A kind of expression recognition method based on multi views convolutional neural networks | |
| CN105962935A (en) | Brain electrical nerve feedback training system and method for improving motor learning function | |
| CN103294192A (en) | LED lamp switch control device and control method thereof based on motor imagery | |
| CN107390869A (en) | Efficient brain control Chinese character input method based on movement vision Evoked ptential | |
| CN106491129A (en) | A kind of Human bodys' response system and method based on EOG | |
| CN110584597A (en) | Multi-channel electroencephalogram signal monitoring method based on time-space convolutional neural network and application | |
| CN107799180A (en) | Mental Capital training for promotion method and system based on cognitive-behavioral therapy | |
| CN109034015A (en) | The demodulating system and demodulating algorithm of FSK-SSVEP | |
| CN108319367B (en) | Brain-computer interface method based on motion initiation evoked potential |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20150715 Termination date: 20180629 |