Research On Mental Task Classification For Brain Computer Interface Application

Recent developments in computer hardware and signal processing have made it feasible to use human electroencephalograph (EEG) signals to communicate with a computer. Brain computer interface (BCI) is a brand-new interface between the human brain and the computer, which gives their users communication and control channels that do not depend on the brain's normal output channels of peripheral nerves and muscles, but accepts commands encoded in neurophysiology signals. As a wholly new information communication and control technology, BCI can provide the paralyzed, especially those "locked-in" but with intake ideation, with an effective communication and control channels with outside word. BCI technology also has potential applications in other fields such as atuocontrol and defense. Because of the enormous prospects of its applications, BCI technology has provoken interests of international scientists, and BCI research has drawn attention of scientists in the brain-science, rehabilitation engineering, biomedical engineering and human machine automatic control.The physiological phenomenon that different types of mental activity can activate distinct brain cortex area and evoke different EEG rhythmic activities make it possible to classify mental tasks using EEG signal. It realize a BCI based on mental tasks. Mental task-based BCI is based on the recording and classification of circumscribed and transient EEG changes during different types of imagery such as imagination of left-hand, right-hand movement or solving a mathematic question. Through the collection of EEG, we classify the feature vector of data and combine the different mental tasks with different instruction to realize the communication between the brain and outer equipments.In the BCI system, Signal collection, feature extraction and recognition are the crucial works. If those works are incorrectly done, the BCI system will also incorrectly recognize the special instruction in the user's mental activities, and neither can the system send out correct controlling instruction which coincides with the user's intention to the devices. In this thesis, our study is mainly about those works in classification towards different mental tasks based on EEGIn the thesis we adopt Independent Component Analysis to deal with EEG. During the collection process of medical signals, recorded data of scalp EEG are often mixed with artifacts, it is necessary to correct these artifacts. ICA is one kind of new statistical method developing in the recent years, whose basic idea is to decompose observed signals to some independent components through the optimization algorithm on the principle of statistical independence. Applying ICA in the artifact correction and feature extract of EEG has been a hotspot recently. The thesis emphasizes on the introduction of basic ICA principle and typical algorithm, especially the Informax and FastICA algorithm, and we validate the ICA to artifact correction effect on speech signal and EEG.. Experimentation results prove that ICA can successfully impair even totally eliminate all the disturbing signals among the EEG while protect useful signal, which promises great future in the medical signal processing area.In the thesis, Six-order AR model is built to extract facture of mental EEG and use the AR model coefficients directly to represent the data to classify. The proper BP neural network is designed to classify the feature of all mental tasks data, and get high correct rate of classification. In the thesis, we also classify two to five mental tasks, and all can get good result.