Neural quantification of timbre and emotions from Indian Classical Music: A multifractal exploration

Music is believed to be one of the strongest mediums to affect both psychological and neural responses in humans. In case of Indian Classical Music (ICM), raga rendition has the power of evoking emotions of certain colors due to its characteristic note structures, tempo and method of unfoldment. The aim of this work is to study and understand the neural origin of different timbres (Sitar and Sarod), emotions (happy and sad) and audience category (musician and non-musician), using certain Alaap (without any accompaniment) portions from renditions of several maestros as input signals. Using robust non linear parameters like multifractal width (MW) and asymmetry, we have tried to characterize and compare the EEG responses originating from the temporal, occipital and frontal lobes of brain. From MW and asymmetry tools, significant classification has been obtained for timbre, emotions as well as audience classification. To compute the statistical significance of the results, tools like one-way ANOVA and Mahalanobis Distance (M.D.) were used to compare the two classes under comparison (emotion/timbre/audience category). While ANOVA revealed significant timbre classification for happy clips and significant emotion classification for Sitar clips, M.D. showed prominently higher values for sad clips of both timbres in occipital lobe, significant separation in different lobes were obtained characterizing musicians and non-musicians. The paper presents new and interesting results regarding the neural processing and characterization of timbral and emotional parameters in musicians and non-musicians using Indian Classical Music as the input stimuli. •Brain response study using EEG with instrumental Indian Classical Music clips.•Studying the neural manifestations of happy and sad clips, played in Sarod and Sitar among musicians and non-musicians.•Non-linear analysis of alpha brain rhythms using multifractal width and asymmetry as classificatory parameters.•Neural response characterization with statistical tools like ANOVA and Mahalanobis Distance.