Error-related processes during motor imagery are a driver of motor learning

Motor imagery is known to be effective for the acquisition of motor skills. Given its covert nature and thus an absence of feedback related to task performance, key to improving overt movement performance, how motor imagery drives skill acquisition is unclear. Here we investigated error-related processes in motor imagery and the role errors may play in driving motor learning. Specifically, we examined (a) the change in self-reported accuracy over time when participants trained on a complex motor skill using motor imagery and (b) if greater change in self-reported accuracy across motor imagery training predicted subsequent physical performance of the motor skill. We hypothesized that self-reported accuracy would increase as a function of time and that greater improvements in self-reported accuracy would predict superior physical performance. Results revealed a significant increase in self-reported accuracy over time. Moreover, changes in self-reported accuracy were highly predictive (R² = 0.43) of physical performance, with participants who showed the greatest improvement in self-reported accuracy also demonstrating greater accuracy during physical performance. These findings support the notion that error-related processes are present and likely play a critical role in learning via motor imagery. We discuss possible mechanisms of error processing in motor imagery, including the updating of internal models. We conclude by proposing a "fusion model" of motor imagery, which posits that motor imagery integrates sensory predictions of movement with a cognitive simulation of movement that permits updating of sensory prediction, in turn facilitating performance improvement and ultimately motor learning. (PsycInfo Database Record (c) 2026 APA, all rights reserved).