Changes in Coherent Activity between EEG and Various Frequency Components of Music while Listening to Familiar and Unfamiliar Songs

Music is known to have a pleasurable and even palliative effect. While the effects of music are well known, the causes behind such effects are not. We have previously reported an increase in coherence between electroencephalograph (EEG) and autonomic variables such as heart rate, blood pressure, and respiration when subjects listened to songs. In this study we compared the coherencies between four different features of songs, envelopes of song and songs filtered within frequency bands representing low frequency i.e., bass, medium frequency i.e., mezzo-soprano and high-frequency components with three features of EEG, i.e.The envelopes of EEG and EEG filtered in alpha and beta bands. We report results from 8 subjects (4 males and 4 females) while they listened to 3 different songs. EEG was analyzed from four scalp positions (T3, T4, P3, and P4) during slow and fast tempo songs and a song selected by the subject (referred to as favorite song). Two parameters were derived from coherencies computed between these signals, the frequencies where coherence was above a threshold (0.55), and the average of the coherence values above the threshold as indicators of the coherent bandwidth and the average magnitude of coherence in this bandwidth. We compared the change in these parameters during the slow and fast song from those computed during favorite songs. The reason for comparison between favorite song and slow as well as fast rhythm song was to determine how much of the changes in these are a result of rhythm vs familiarity given that both slow and fast songs were unfamiliar to the subjects. The results showed that overall the coherent bandwidth was lower during the fast and slow song than favorite. These measures were different when bandpass filtered EEGs were used with the most pronounced effect seen in the alpha band with an increase in bandwidth for both fast (for EEG locations T3, T4, and P4) and slow songs. These results provide further insight into how different features of songs affect rhythmic electrical activity in the brain and thus may eventually help in improving music therapy for patients.