Effective Computational Techniques for Generating Electroencephalogram Data

Abstract

The complexity of the electroencephalogram makes it a significant challenge for physicians and engineers to extract useful information from, process, and classify the electroencephalogram signals. Moreover, the difficulty in conducting clinical experimentation limits the collection of a sufficient number of electroencephalogram data samples for further processing using advanced computational techniques such as deep learning. This complexity and difficulty together with the inflexibility and the subtle linearity of the traditional signal processing techniques motivate us to find innovative techniques to address the problem of insufficient electroencephalogram data. In this paper, a number of computational and statistical techniques to generate electroencephalogram data from a previously done experiment on 30 healthy participants experiencing painful stimuli are applied. We believe this application will benefit the research in the field of biomedical signal processing.