Discrimination between cognitive tasks - a comparative study

In this paper we investigate the performance of using, within the EEG classification algorithm, feature extractors such phase synchrony and relative power spectral density indices calculated for seven frequency bands - delta (0.1-4 Hz), theta (4-7 Hz), alpha1 (7-10 Hz), alpha2 (10-13 Hz), beta1 (13-18 Hz), beta2 (18-30 Hz), gamma (30-70 Hz) -, that are believed to mostly reflect functionally different components of cognitive cortical activity. For classification of five mental tasks (baseline, count, letter, math and rotate) a multi layer perceptron (MLP) neural network classifier was used. A standard principal component analysis (PCA) was applied here in order to reduce the high dimensionality of the input data. To quantitatively assess the performance of the extracted features a comparative study between their corresponding results and between them and some other EEG features is also provided.