Characterizing EEG during mental activity using non-linear measures: the more concentration, the higher correlation dimension

We have investigated a method of discriminating EEG changes during emotional and cognitive states in healthy volunteers who responded to three designed states, in which the subjects were relaxing with ease with closed eyes, listening to music and computing a simple subtraction with closed eyes. Especially, we estimated EEG dimensional complexity by Skinner´s point-wise correlation dimension (PD2) method for each mental state. It was found that the subjects, who responded that they had concentrated more during the arithmetic task, show higher PD2 in their non-linear EEG measures, in comparison with the subjects who responded that they had not concentrated during the task. This high PD2 is also significant in statistical analysis. A subject who had the highest score in evaluating the degree of induced emotion during emotional task shows significantly lower PD2 in statistical analysis than other subjects who had low scores. For these results, linear spectral analyses were performed; however, they did not show any significant difference. Only non-linear dynamical analysis shows the significant difference of the mental states