Classification of Encephalographic Signals using Artificial Neural Networks

Roberto Sepulveda, Oscar Montiel, Gerardo Diaz, Daniel Gutierrez, Oscar Castillo


For the signal classification of eye blinking andmuscular pain in the right arm caused by an externalagent, two models of artificial neural network architecturesare proposed, specifically, the perceptron multilayerand an adaptive neurofuzzy inference system. Bothmodels use supervised learning. The ocular and electroencephalographictime-series of 15 people in the rangeof 23 to 25 years of age are used to generate a database which was divided into two sets: a training set anda test set. Experimental results in the time and frequencydomain of 50 tests applied to each model show that bothneural network architecture proposals for classificationproduce successful results.


EEG, BCI, brain-computer interface, blink, artificial neural network, FFT

Full Text: PDF (Spanish)