Applications Of Nonlinear Dynamics In The Brain-computer Interface

Recently, Brain Computer Interface becomes a very popular resaerch problem, it is an interface between brain and machine and not a communicating system depending on conventional brain output channel which means peripheral nerve and muscle tela. It can be used as an newly assistant communicating technique which helps people who suffer from severe disability, so it attracts more and more attention in medical, engineering field especially healing engineering field. It is also a new way of electroencephalo-graph(EEG) application. Human brain is a very complex non-linear, non-stationary system, so the EEG and electrocorticogram(ECoG) which are derived from human brain are also non-linear and non-stationary. Signal processing analyzing method based on linear and stationary assumptions, although has obtained important results, can not digging into the problem. Complexity analyzing method based on Non-linear Dynamics and time-frequency analyzing means of Hilbert-Huang Transform(HHT) extended a new approach to EEG or EcoGMotor imagery analysis is a branch of BCI, whose analysis difficulty is in that EEG or ECoG will have great difference in EEG or ECoG for different person or even one person at different time, so different volunteer or the same volunteer at different time, to verify the imaginary movement, meaningful frequency and time are also different. According to this, it is a completely new attempt using Lattice Complexity and HHT to analyze motor imagery based on ECoG, whose aim is to classify the imaginary movement of tongue or left small finger, in this paper. We won the third place in the Third International Brain Computer Interface Competition, which also proved the effect of the methods. Surface Laplacian Filter is a effective EEG preprocessing method, we use this method to preprocess ECoG is also an new innovation.