Interpersonal predictive coding in piano duets: the role of multimodal micro-cues in joint coordination

Piano duets provide an ecologically meaningful model of high-precision joint action. Performers must coordinate note onsets, tempo changes, articulation, and expressive dynamics despite delays in auditory transmission and sensorimotor processing. This temporal constraint suggests that auditory feedback alone cannot fully explain fine-grained interpersonal synchrony. This narrative review examines how pianists may combine prospective prediction with retrospective error correction during joint performance. Drawing on behavioral, computational, motion-capture, and neuroimaging evidence, we propose that performers use learned expectations about a partner's timing and expressive style together with current auditory, visual, and kinematic information to estimate upcoming musical events. Breathing and body sway may provide phrase-level information about tempo and structural boundaries, although much of this evidence comes from non-piano ensembles. Preparatory hand movements and cueing gestures provide more immediate information about impending note onsets, while auditory feedback remains essential for evaluating expectations and correcting subsequent timing. Computational error-correction and coupled-oscillator models offer complementary descriptions of immediate onset adjustment, longer-term tempo adaptation, and reciprocal temporal influence. Current evidence nevertheless relies heavily on simplified tapping tasks, cross-ensemble extrapolation, and correlational measures of neural coupling. Future research should directly manipulate cue reliability in ecologically valid piano-duet tasks and combine behavioral timing, motion capture, eye tracking, and hyperscanning to test how prediction and feedback jointly support role-dependent interpersonal coordination.