Classification of complex tasks for Brain-Computer Interface

The aim of this work is to investigate the possibility to use complex mental tasks (imagination of playing tennis and of spatial navigation), translated from the fMRI research, to drive an EEG-based BCI. This could be particularly useful for implementing new BCI paradigms for non-responsive patients. Common Spatial Patterns for the preprocessing and different classifiers were compared on a dataset recorded from 10 subjects. Classification accuracies show that the considered mental tasks could have some potentialities in the development of new paradigms for EEG-based BCIs.