The realization of one’s musical ideas at the keyboard is dependent on the ability to transform sound into movement, a process called audiomotor transformation. Using fMRI, we investigated cerebral activations while classically‐trained improvising and non‐improvising musicians imagined playing along with recordings of familiar and unfamiliar music excerpts. We hypothesized that audiomotor transformation would be associated with the recruitment of dedicated cerebral networks, facilitating aurally‐cued performance. Results indicate that while all classically‐trained musicians engage a left‐hemisphere network involved in motor skill and action recognition, only improvising musicians additionally recruit a right dorsal frontoparietal network dedicated to spatially‐driven motor control. Mobilization of this network, which plays a crucial role in the real‐time transformation of imagined or perceived music into goal‐directed action, may be held responsible not only for the stronger activation of auditory cortex we observed in improvising musicians in response to the aural perception of music, but also for the superior ability to play ‘by ear’ which they demonstrated in a follow‐up study. The results of this study suggest that the practice of improvisation promotes the implicit acquisition of hierarchical music syntax which is then recruited in top‐down manner via the dorsal stream during music performance.
Using fMRI, cerebral activations were studied in 24 classically-trained keyboard performers and 12 musically unskilled control subjects. Two groups of musicians were recruited: improvising (n=12) and score-dependent (non-improvising) musicians (n=12). While listening to both familiar and unfamiliar music, subjects either (covertly) appraised the presented music performance or imagined they were playing the music themselves. We hypothesized that improvising musicians would exhibit enhanced efficiency of audiomotor transformation reflected by stronger ventral premotor activation. Statistical Parametric Mapping revealed that, while virtually 'playing along' with the music, improvising musicians exhibited activation of a right-hemisphere distribution of cerebral areas including posterior-superior parietal and dorsal premotor cortex. Involvement of these right-hemisphere dorsal stream areas suggests that improvising musicians recruited an amodal spatial processing system subserving pitch-to-space transformations to facilitate their virtual motor performance. Score-dependent musicians recruited a primarily left-hemisphere pattern of motor areas together with the posterior part of the right superior temporal sulcus, suggesting a relationship between aural discrimination and symbolic representation. Activations in bilateral auditory cortex were significantly larger for improvising musicians than for score-dependent musicians, suggesting enhanced top-down effects on aural perception. Our results suggest that learning to play a music instrument primarily from notation predisposes musicians toward aural identification and discrimination, while learning by improvisation involves audio-spatial-motor transformations, not only during performance, but also perception.
Mirror neurons in the cerebral cortex have been shown to fire not onlyduring performance but also during visual and auditory observation ofactivity. This phenomenon is commonly called cerebral resonance behavior.This would mean that cortical motor regions would not only beactivated while singing, but also while listening to music. The sameshould hold true for playing a music instrument. Although most individualsare able to sing along when they hear a melody, even highlyskilled instrumentalists, however, are frequently unable to play by ear.They are score-dependent—i.e. they are only able to play a piece of musicwhen they have access to the notes—while musicians who are able to playby ear and improvise are non score-dependent; they are able to playwithout notes. Our hypothesis is that score-dependent instrumentalistswill exhibit less cerebral resonance behavior than non score-dependentmusicians while listening to music. Using fMRI to measure BOLD response,subjects listen to two-part harmony presented with headphones.The following experimental conditions are distinguished: (1) well-knownvs. unknown music (2) motor imagery vs. attentive listening. A voxelbasedanalysis of differences between the condition-related cerebral activationsis performed using Statistical Parametric Mapping.
LINK
