The Research On Brain-Computer Interface Based On SSVEP And ERN

Brain-computer interface(BCI) is an assistive technology that conveys users’ intentions by decoding various brain activities and translating them into control commands, without the need of verbal instructions and/or physical interactions. It is a comprehensive technology which involves many disciplines such as physiology, neuroscience, signal processing and control theory. With the development of technology. BCI will play an important role in medical rehabilitation, live entertainment and military operations, etc.The processing and recognition on Electroencephalograph(EEG) which is recorded with non-invasive techniques recognition is very complicated and difficult because of its weakness and non-stationary. To solve those problems, this thesis designs a BCI system based on steady-state visual evoked potential(SSVEP) and uses time-frequency analysis method to extract the frequency components, then analyzes the error related negativity(ERN) which is evoked by incorrect command of BCI system, and put forward a new feature extraction algorithm of ERN. The mainly work as follows,Because of the direct relationship between the visual stimulation frequencies and frequency distribution of evoked SSVEP, the thesis designed a multi-frequencies stimulus BCI system based on SSVEP. Firs the preprocessing of EEG such as spatial filtering、time-frequency transform、frequency extraction were used. Then, the pseudo smoothed Wigner-Ville distribution(SPWVD) was used in extract SSVEP frequency and compared it with other commonly used frequency extraction algorithms such CCA and SPDA. The results showed that the SPWVD had advantages in processing non-stationary EEG, although under short time window, the classification accuracy and information transmission rate of SPWVD time-frequency analysis method was lower than other methods, but it was more sensitive under the long time window, the joint time-frequency distribution can more reflect the signal characteristics in a long time window. With the increase of the time windows, the effect of SPWVD time frequency analysis method was more obvious, and the classification accuracy and information transmission rate of the SPWVD were better than other commonly methods.The main limitation of BCI system is that it cannot avoid mistakes in practice. Because the ERN only exists in the error monitoring process and has little relationships with specific paradigms, it can be integrated into the BCI system in other modes, recording the EEG for the second time and recognizing the existence of ERN, so as to correct the mistakes and improve the BCI’s reliability. Because the single trial detection of ERN is difficult and multi-channels data may cause over fitting, the thesis proposed a new approach of combining multiple-channel features from temporal, spectral, and spatial domains through two times of dimensionality reductions based on neural network. It took advantage of information from multiple electrodes and combination of features from different domains rather than single features. Although the ERN signal is inter-individual different, the classification results using receiver operating characteristic curves(ROC) and area under ROC curve(AUC) metrics suggest superior performance with combined features over single features(AUC: 0.7818).The brain-computer interface based on SSVEP and ERN has higher reliability, it will promote the BCI system into practice.