Classification of Motor Imagery Tasks by means of Time-Frequency-Spatial Analysis for Brain-Computer Interface Applications

We have developed new algorithms for classification of motor imagery tasks for brain-computer interface applications by analyzing single trial scalp EEG signals in the time-, frequency-, and space-domains. These new algorithms have been evaluated using a publically available dataset. The results are promising, suggesting that the newly developed algorithms may provide useful alternative for noninvasive brain-computer interface applications