Modelo computacional para classificação de sinais EEG

Abstract

People with motor impairments and Neurological disorders have little control over parts of the body, for that have great difficulty in locomotion. The signs of electroencephalogram (EEG) can be used as a source of information that allows communication between the brain and a intelligent wheelchair (CRI). Several techniques have been proposals to recognize patterns during the imagination of the movement by means of the analysis of EEG. the neural networks artificial are widely used for the classification task for presenting a vast theoretical framework and a good performance on complex problems. This work of The research aimed to develop a model computational for analysis, classification and activation of a CRI through EEG signals from the motor imagination of left, right, both fists and both wrist movements the feet. The model is composed of three modules: signal acquisition, signal processing and CRI activation. EEG signals were acquired through the database eegmmidb – EEG Motor Movement/Imagery Dataset, captured through the system BCI2000, samples of EEG signals from 9 individuals. The results obtained show that the model computation had a performance above 81% in the correct classification of movements imagination. The method based on decision tree (C4.5) presented better performance in relation to vector machine methods of support (SVM) and neural networks (RN). The techniques used are promising and can be used in the classification of patterns of movement image, making its use feasible in a brain-computer interface.