Clustering the Sources of EEG Activity during Motor Imagery by Attractor Neural Network with Increasing Activity (ANNIA)
Electrical brain activity in subjects controlling Brain-Computer Interface (BCI) based on motor imagery is studied. A used data set contains 7440 observations corresponding to distributions of electrical potential at the head surface obtained by Independent Component Analysis of 155 48-channel EEG recordings over 16 subjects. The distributions are interpreted as produced by the current dipolar sources inside the head. To reveal the sources of electrical brain activity the most typical for motor imagery, the corresponding ICA components were clustered by Attractor Neural Network with Increasing Activity (ANNIA). ANNIA was already successfully applied to clustering textual documents and genome data [8,11]. Among the expected clusters of components (blinks and mu-rhythm ERD) the ones reflecting the frontal and occipital cortex activity were also extracted. Although the cluster analysis can not substitute careful data examination and interpretation however it is a useful pre-processing step which can clearly aid in revealing data regularities which are impossible to tract by sequentially browsing through the data.
KeywordsBrain computer interface motor imagery independent component analysis attractor neural network with increasing activity
Unable to display preview. Download preview PDF.
- 2.Bobrov, P., Frolov, A., Cantor, C., Fedulova, I., Bakhnyan, M., Zhavoronkov, A.: Brain-Computer Interface Based on Generation of Visual Images. PLoS One 6(6), e20674 (2011), doi:10.1371/journal.pone.0020674Google Scholar
- 4.Faber, J., Novak, M.: Thalamo-cortical reverberation in the brain produces alpa and deta rhythms as iterative convergence of fuzzy cognition in a stochastic ebvironment. Neural Network World 21(2), 169–192 (2011)Google Scholar
- 5.Frolov, A.A., Sirota, A.M., Husek, D., Muraviev, I.P., Polyakov, P.A.: Binary factorization in hopfield-like neural networks: single-step approximation and computer simulations. Neural Network World 14(2), 139–152 (2004)Google Scholar
- 7.Frolov, A.A., Polyakov, P.Y., Husek, D., Rezankova, H.: Neural Network Based Boolean Factor Analysis of Parliament Voting. In: Proceedings in Computational Statistics, Heidelberg, pp. 861–868 (2007)Google Scholar
- 9.Frolov, A.A., Husek, D., Bobrov, P.: Comparison of four classification methods for brain-computer interface. Neural Network World 21(2), 101–115 (2011)Google Scholar
- 10.Frolov, A., Husek, D., Bobrov, P., Korshakov, A., Chernikova, L., Konovalov, R., Mokienko, O.: Sources of EEG activity most relevant to performance of brain-computer interface based on motor imagery. Neural Network World 22(1), 21–37 (2012)Google Scholar