On the use of spiking neural network for EEG classification

Article type: Research Article

Authors: Goel, Piyush^{a; *} | Liu, Honghai^a | Brown, David^a | Datta, Avijit^b

Affiliations: [a] Institute of Industrial Research, University of Portsmouth, PO1 3QL, UK | [b] Institute of Biomedical and Biomolecular Sciences, University of Portsmouth, UK

Correspondence: [*] Corresponding author. E-mail: [email protected]

Abstract: This paper presents a new classification technique of continuous EEG recordings, based on a network of spiking neurons. Human EEG signals published on the BCI Competition website were used for the study. The signals were pre-processed using Wavelet Transform to remove the noise and to extract the low frequency content. Analysis of the signals was performed on the ensemble EEG and the task of the neural network was to identify the P300 component in the signal. The network employed leaky-integrate-and-fire (LIF) neurons as nodes in a multi-layered structure. The method involved formation of multiple weak classifiers to perform voting. Collective results are used for final classification. Results have shown the method to perform better than a genetic algorithm approach to the same problem.

DOI: 10.3233/KES-2008-12404

Journal: International Journal of Knowledge-based and Intelligent Engineering Systems, vol. 12, no. 4, pp. 295-304, 2008