The detection and classification of the mental state elicited by humor from EEG patterns

In this paper we investigate the use of EEG to detect the affective state of humor. The EEG of five subjects was recorded while they recalled humorous videos. Extracted frequency features were compared to a control state in which users where asked to remain in a neutral mental state. An ANOVA test performed on the two groups: neutral and humor recall found a statistically significant difference in the frequency range 28-32 Hz for a number of channels including T7 and P7. Both of which presented the greatest statistically significant results with p values of 0.009 and 0.0 respectively Furthermore, we demonstrate that these mental states can be classified using Principal Component Analysis followed by a 3 features Linear Discriminant Analysis resulting in a leave one out classification accuracy of 95%.