Research On Tactile P300 Brian-computer Interface Paradigm

Brain Computer Interface(BCI)is a new type of human-computer interaction channel that does not rely on physical nerve or muscle tissue to convert brain signals into control commands to operate external devices or to communicate.With the deepening of related research,the future of brain-computer interface in medical,rehabilitation,entertainment and military fields has been paid more and more attention.Classic brain-computer interface is divided into induced and spontaneousaccording to whether the Electroencephalogram(EEG)signal calling for external stimuli or not.The character spelling and wheelchair control system are based on the existing induced braincomputer interface.This kind of feedback and stimulus sharing just the common channel is called visual non-independent system.In order to construct a visual independent braincomputer interface,this paper chooses haptics as the stimulus channel,makes the scheme of brain-computer interface system,designs and realizes the hardware circuit,and proves that it can stimulate event-related potentials with time-domain characteristics.The offline training was carried out to analyze the time-domain characteristics such as peak and delay of the peak EEG signal.Then,this feature was combined with the improved stepwise linear discrimination analysis(SWLDA)algorithm to verificate the similarity and separability of signal.The offline analysis show that the characteristics of EEG electrodes and excitation and has similarities and differences due to different subjects,so the system performance was designed with different parameters to obtain a better classifier,the correct rate of the best combination of more than 80% shows the signal separability.During on-line experiment,10 subjects took part in with their appropriate stimulation time,the electrode channel and the corresponding classifier parameters.The average of classification accuracy and information transfer rate at 89.1% and 14.77 bits/min indicate that the characteristic signal stability exists and can be divided,and also prove the feasiblity of the designed tactile brain computer interface system.The work of this paper has important significance for the design and use of the tactile induced brain computer interface.