A New Scheme of EEG Signals Processing in Brain-Computer Interface Systems

Author

Esmaeili, Maryam ; Jabalameli, Mohamad H. ; Moghadam, Zeinab

Author_Institution

Amirkabir Univ. of Technol., Tehran

fYear

2007

fDate

2-4 Nov. 2007

Firstpage

522

Lastpage

522

Abstract

In this paper, dynamic synapse neural network (DSNN) has been applied to perform EEG signal recognition task. The wavelet packet transform is applied to the EEG signal in order to decompose it into frequency sub-bands, before being introduced to the neural network. In this study we have applied a genetic algorithm (GA) learning method with different fitness functions to optimize the neural network. The advantage of the GA method is that it facilitates finding of a semi-optimal parameter set in the search space domain. The network has been testes for EEG signals tat are provided from BCI Competition 2003 and the results show the power of DSNN in processing of noisy nature signals as EEG signals.

Keywords

electroencephalography; genetic algorithms; learning (artificial intelligence); medical signal processing; neural nets; source separation; wavelet transforms; EEG signal decomposition; EEG signal processing; EEG signal recognition; brain-computer interface system; dynamic synapse neural network; fitness function; genetic algorithm learning method; wavelet packet transform; Biological neural networks; Brain computer interfaces; Electroencephalography; Frequency; Genetic algorithms; Learning systems; Optimization methods; Signal processing; Wavelet packets; Wavelet transforms;

fLanguage

English

Publisher

ieee

Conference_Titel

Granular Computing, 2007. GRC 2007. IEEE International Conference on

Conference_Location

Fremont, CA

Print_ISBN

978-0-7695-3032-1

Type

conf

DOI

10.1109/GrC.2007.149

Filename

4403154