Proper

Abstract

In recent years the developments focused in brain computer interfaces (BCI) have benefited with the technological advances in noninvasive acquisition of encephalographic signals, therefore the researches related of this field have centered in increasing the efficiency of the recognizing of brain activities using different methodologies. This paper has as objective recognize two imaginary movements of a person with physical disability using support vector machines (SVM), for that will use a methodology which will begin with filter the brain signals in the sequential frequencies ranges and in each range will be extract the main features using the independent components analysis (ICA), after will be select the features more relevant among all the frequencies ranges through the analysis of minimum redundancy and maximum relevance (mRMR) for training the SVMs with the selected data. Obtaining as a result that this methodology improved the recognition of an imaginary movement to right in 8% and 9% for the imaginary movement to left in comparison with a general methodology for imaginary movement recognition.