Emotion in real and imagined music: Same or different?

This paper reports two studies on the emotional effect of imagined music. The first employed a professional pian-ist who listened to a recording of his own (already re-corded) performance of a piece while at the same time registering the emotion expressed on a two-dimensional emotion space (2DES). The process was repeated with the same pianist imagining the same performance of the piece. While the shape of the emotional response pro-files in the two (arousal and valence) dimensions were similar across real and imagined conditions, the imagined condition always produced a temporally stretched re-sponse. Cognitive load was proposed as an explanation of this time warping. Interference due to cognitive load was removed in a sec-ond exploration by asking the participant to make post-performance, rather than continuous, responses to imag-ined and real performances. Forty-five undergraduate music students were asked to make emotional ratings (in-cluding valence and arousal used in the 2DES) in re-sponse to highly familiar pieces which were either played to the participants over loudspeakers, or imagined in the absence of sound. Repeated measures ANOVA revealed no significant difference in emotional responses between imagined and real conditions. These explorations suggest that emotional response to music can be reliably produced in the absence of music, although continuous response collection produces some interference, and in any case the music must be highly familiar. The study has important implications in emo-tion in music research, and in mood induction and regula-tion procedures