Mouse control by Independent Component Analysis and Differential Evolutions

摘要

During the past 10 years, the research pace of brain computer interfaces (BCIs) has quickened greatly because of their potential application value. The goal of a BCI is to provide its users a communication and control channel that do not depend on brain's traditional output pathways of peripheral nerves and muscles. Its potential applications include restoring functions to those with motor disabilities, alarming paroxysmal diseases (e.g., epileptic seizure prediction), manipulating human's control in inhospitable or even dangerous environments, etc. Inherently, research on BCIs is an interdisciplinary field involving neuroscience, psychology, engineering, mathematics, clinical rehabilitation, and computer science. Feature subset selection is the process of identifying and removing as much irrelevant and redundant information as possible. This reduces the dimensionality of the data and may allow learning algorithms to operate faster and more effectively. In some cases, accuracy on future classification can be improved; in others, the result is a more compact, easily interpreted representation of the target concept. In new BCI systems for increase accuracy, usually used increased number of electrodes. In this case the increased number of electrodes causes a non-linear increase in computational complexity (i.e. decrease transfer rate). In this paper, we attempt to enhance the single trial EEG patterns and Redundancy Reduction using the components obtained by independent component analysis (ICA) for reduction of artifacts and Differential evolution (DE) for Feature subset selection.