Classification of Motor Imagery signal using wavelet decomposition: A study for optimum parameter settings

Abstract

Motor Imagery (MI) based Brain Computer Interface (BCI) is an assistive technology, which translates the brain signals into commands to control external devices. A basic MI classification involves different steps of signal processing such as pre-processing, spatial filter, feature extraction and classification. There are numerous combinations of these steps that we can explore to achieve the better result. In this work we have systematically compared different parameter settings for wavelet-based feature extraction in search for optimum performance. Our detailed experimental results illustrate how we can choose appropriate wavelet function, order, number of decomposition levels and finally selection of coefficient at different levels.