Brain Computer Interface. Comparison of Neural Networks Classifiers.

Abstract

Brain Computer Interface is an emerging technology that allows new output paths to communicate the user’s intentions without use of normal output ways, such as muscles or nerves (Wolpaw, J. R.; et al., 2002).In order to obtain its objective BCI devices shall make use of classifier which translate the inputs provided by user’s brain signal to commands for external devices. The primary uses of this technology will benefit persons with some kind blocking disease as for example: ALS, brainstem stroke, severe cerebral palsy (Donchin et al., 2000).This report describes three different classifiers based on three different types of neural networks: Radial Basis Functions RBF, Probabilistic Neural Networks PNN, and Multi-Layer Perceptions MLP. The report compares the results produced by them in order to obtain conclusions to apply to an on-line BCI device; it also describes the experimental procedure followed in the experiments. As result of the tests carried out on five healthy volunteers an estimation of the success rate for each type of classifier, the type and architecture of the classifier, and filtering windows are established.