Purpose of the study: The current study investigates the auditory perception of improvised and imitated piano melodies. Improvisation in music is a creative process during which musicians can choose relatively freely the specific notes they play and when to play them. Such an action can be described to be intention-based. On the other hand, imitating a phrase of music can be characterized as stimulus-based action – the notes of someone else’s playing guide the action during imitative music playing. Recent brain imaging studies have shown differences in neural activity during the action of improvising and imitating melodies. We test whether there are differences in brain activations when jazz musicians listen to improvised and imitated piano melodies, i.e., the perception of the end products of different action modes. Such differences may be expected based on claims that action and perception recruit overlapping neural networks.

Methods: Recorded piano melodies that had been played over novel backing tracks of three contrasting jazz styles (swing, bossa nova, blues ballad) were used as stimuli. These melodies were either spontaneous improvisations or practiced imitations of the improvisations. Using event-related fMRI and a sparse sampling technique (interleaved silent steady state imaging, where longitudinal magnetization is maintained during silent periods), 10-second excerpts from these melodies were presented with backing during the silent periods. Following each stimulus presentation, 7 functional whole brain scans were recorded and participants (22 jazz musicians, who had on average 12.8 years of piano playing experience) indicated by key press whether the heard melody was improvised or imitated. Data were analyzed with SPM5 according to a 2 x 2 factorial design that took into account the objective classification of stimuli (real improvisations/real imitations) and subjective classifications based on participants’ responses (judged improvised/judged imitated). After scanning, each participant’s musical and jazz experience was assessed by a questionnaire.

Results: Behavioral data show that participants could distinguish between improvised and imitated melodies with a correct response rate (55%; range 44-65%) and a d-prime (.25; range -.30 to .83) that were significantly better than chance (p<.001). These performance rates were correlated with various measures of musical and jazz experience (e.g., number of hours, participants spent playing jazz together with other musicians). Listening to piano melodies revealed activation of a widespread network comprising auditory cortex, premotor area, supplementary motor area (SMA), and cerebellum (conjunction analysis of all listening conditions). Listening to real improvisations compared with listening to real imitations (objective classification) showed stronger activation in the amygdala region for improvisations. This result is in accord with the suggested role of amygdala in detection of stimuli that are behaviorally relevant. For the contrast based on the subjective classification, listening to melodies that were judged to be improvised compared with those that were judged to be imitated revealed stronger activations in a network including pre-SMA, frontal operculum and insula. This activation pattern could reflect a stronger motor simulation process during listening to melodies that are were judged to be improvised, which would be also consistent with studies showing stronger activation in pre-SMA during the act of improvising music.

Conclusions: Taken together, these findings suggest that differences in brain activation arise at two levels when listening to improvisations and imitations: one largely independent of listeners’ judgments and the other more strongly linked with listeners’ judgments.