Mental task classifications using prefrontal cortex electroencephalograph signals.

Main Methods:

• EEG signals were recorded from three prefrontal electrode positions (Fp1, Fpz, Fp2) during mental tasks (mental arithmetic, ringtone, finger tapping, words composition), baseline, and eyes closed conditions.
• Feature extraction was performed using the Hilbert Huang Transform (HHT) energy method.
• Classification was achieved using an artificial neural network (ANN) optimized with a genetic algorithm (GA).

Main Results:

• The dominant alpha wave during the eyes closed state was clearly detected in the prefrontal cortex.
• The system achieved an average classification accuracy of 73% for distinguishing mental tasks from the baseline task across five subjects.
• An average accuracy of 72% was obtained for classifying pairs of mental task combinations.

Conclusions:

• The prefrontal cortex non-hair area is a viable and convenient alternative for EEG-based BCI applications.
• The proposed HHT energy feature extraction and GA-optimized ANN classification method effectively discriminates between mental tasks.
• This approach significantly enhances the practicality of BCI systems by eliminating the need for scalp gel and reducing setup time.