Heart rate deceleration following movement error has been observed previously and often interpreted as an autonomic response to adverse behavioral outcomes. The present study sought to dissociate whether HR changes after movement perturbations primarily index negative movement outcomes or instead reflect violations of expectation about movement outcomes. The authors used a reaching paradigm to test whether expectation or error valence better explains HR modulation during sensorimotor behavior.
The investigation employed a robotic manipulandum to impose visuomotor perturbations during reaching movements. Human participants of both sexes performed the reaching tasks while physiological measures, including heart rate, were recorded. The experimental design manipulated contextual information about upcoming outcomes and varied task constraints to separate the influences of outcome valence, expectation, and different types of error.
The study reproduced established visuomotor perturbations that induce movement errors. In a core condition, perturbations produced deviations that caused observable task errors. In separate conditions, the experimenters provided contextual cues that predicted upcoming movement outcomes, allowing participants to form expectations prior to movement. Additionally, target width was manipulated in some conditions to permit successful task completion despite the presence of visuomotor rotations, thereby enabling dissociation between task success/failure and sensory mismatch.
As in prior reports, HR deceleration occurred following visuomotor perturbations that induced movement errors. Crucially, when contextual cues reliably predicted upcoming movement outcomes, HR deceleration shifted earlier and emerged during movement preparation rather than following the occurrence of an error. Across conditions with varying levels of expectation about different movement outcomes, HR deceleration was associated with outcomes that were unexpected relative to the cue-based predictions, rather than being tied to the negative valence of outcomes. These observations indicate that the temporal profile and presence of HR deceleration depend on the participant's expectation state.
To further probe underlying drivers of HR deceleration, the authors altered target width so that participants could still accomplish the task despite visuomotor rotations. Under these circumstances, sensory feedback and internal predictions could still be mismatched (a sensory prediction error), but the task outcome might remain successful (no or reduced task error). The results indicated that HR deceleration was more strongly associated with task error—failures to achieve the task goal—than with sensory prediction error alone. In other words, HR modulation aligned more closely with whether the movement led to an unexpected task-level outcome than with discrepancies between predicted and actual sensory feedback.
Taken together, the findings support an interpretation that HR responses during reaching reflect violations of movement expectation. HR deceleration appears sensitive to whether an outcome is anticipated, and its timing can shift to the preparation phase when contextual cues create clear expectations. The stronger link to task error than to sensory prediction error suggests that HR deceleration indexes higher-level expectation about task success rather than low-level sensorimotor mismatch alone. The authors propose that autonomic responses, as measured by HR, may therefore serve as physiological markers of expectation-related processing during sensorimotor tasks.
These results provide a new perspective on autonomic correlates of motor behavior by emphasizing the role of expectation violation. If HR deceleration reliably signals unexpected movement outcomes, it could be used as a noninvasive physiological index of expectation-related processing in studies of motor learning, decision-making under uncertainty, or clinical conditions with altered predictive processing. The dissociation between task error and sensory prediction error reported here may help refine models that link autonomic measures to different components of sensorimotor control.
The authors declared funding from multiple grants from the Japan Society for the Promotion of Science and support from the Japan Science and Technology Agency. The authors declared no competing interests.
The source article provides the overall design, primary manipulations, and core conclusions described above. Detailed methodological parameters (for example, sample size, exact cueing procedures, timing windows for HR analysis, statistical values, and effect sizes) were not reported in the summary provided here and therefore are not included. Readers should consult the full preprint for those experimental and analytic details if required.