Computing stress-related emotional state via frontal cortex asymmetry to be applied in passive-ssBCI

The identification of stress-related emotional states allows the use of a passive single-switch BCI (passive-ssBCI) together with a main multi-class BCI. Thus, the passive-ssBCI could work in the background of the main BCI, allowing it to recognize stress-related emotions and to switch to a “stress mode”, which allows a better adaptation to the user current mental state, leading the system to better success rates. Frontal cortex asymmetry gave evidence that greater right frontal activity seems to be more highly related to negative emotional states. In this work, index of asymmetry of alpha band was computed by comparing the power of contra-lateral frontal electrodes. Discrete Wavelet transformation was used to decompose the EEG signals in frequency bands. The power spectral density was then calculated using modified Hamming periodogram. Signals from a public database for the analysis of human affective states, which was labeled with volunteers self-assessment scores, were employed. In order to label negative stress-related emotional state, rules based on valence, arousal and dominance dimensional emotions were defined. The results were promissory, because computed index of asymmetry in the alpha band indicated activity into the right frontal hemisphere was higher than the left one for seven out of nine volunteers during negative emotional states. This index could be employed as switch to a “stress mode” in a passive-ssBCI.