Analysis and classification of electroencephalographic signals (EEG) to identify arm movements

The objective of this work is the analysis and classification of electroencephalographic signals (EEG) to identify arm movements. The system must similarly recognize if the movement is executed or imagined. The EEG signals are recordings of the electrical activity of the brain and are subdivided into frequency bands (Alpha, Beta, Delta, Gamma, Mu, and Theta). The interest waves alpha, beta and mu reflect the cerebral activation due to real or imagined movement. These signals vary in time and frequency hindered their identification and classification. With the help of the literature, statistical analysis, wavelet analysis and classification tests were selected 8 EEG channels to use. The wavelet transformed (WT) was applied to the signal to extract time and frequency characteristics. The approximation coefficients of WT were integrated as vectors to system classification inputs. These vectors are composed of different decomposition levels for some channels depending of the wave (alpha, betha, mu). In the classification we used a multilayer perceptron neural network. Finally we identified four movements (real and imagined) of the arms of a healthy person: right hand back and forth, left hand back and forth. The accuracy obtained was 88.72% for 4 movements and 82.71% for 5 movements with the leg movement as the 5^th class.