Modeling Music Emotion Judgment of Cochlear Implant Listeners

Abstract

Different studies have outlined the limitations of the cochlear implant, CI, to transmit pitch, harmony, mode and chord progressions. Despite these difficulties, CI users can discriminate happy songs from sad songs (Vannson et al., 2015, Trends in Hearing). Given this lack of tonal perception, it would have been expected than CI listeners base their emotional judgment only the tempo. However, statistical analysis revealed only a weak effect of this musical parameter. To investigate, how CI listeners can extract emotional information, 12 normal- hearing, NH, listeners were asked to rate modified versions of the stimuli from the study of Vannson et al. In this latter study, the stimuli were short piano pieces composed to induce a specific emotion (Vieillard et al., 2008, Cognition and Emotion), while in current study; the same pieces were performed on two congas. Consequently, NH listeners had access to only purely temporal information to form their emotional judgment. Emotional ratings of the congas pieces from NH listeners were very similar to the rating of the piano pieces from CI listeners. As the tempo only explains a small part of the variability, other temporal aspects such as the rhythmic pattern, syncopation and pulsation were used to model the music emotion judgment. Although CI users have difficulties perceiving modes, they can rely on subtle temporal cues to extract the emotional context of music.