No motor learning during action-imagery practice and no intermanual transfer in the two-ball-rotation task

Action-imagery is an internal simulation of sensory movement experiences in the absence of actual execution. Action-imagery practice (AIP), i.e. the repeated use of action-imagery has been shown to improve subsequent motor execution performance in many tasks. However, AIP is less effective than action-execution practice (AEP), particularly in motor tasks that involve few visuospatial components (e.g. spatial stimulus and response mappings). In such a task in AEP, the development of effector-dependent representations is expected, whereas effector-independent motor and visuospatial representations might not evolve. It remains unclear whether such findings hold for AIP as well. An AIP group and an AEP group practiced two-ball-rotations and a control-practice (CP) -group performed a control task. Execution tests were conducted with the practice and transfer hand, using the practiced rotation and an unpracticed rotation direction. Intermanual transfer was observed neither in AEP nor in AIP, indicating that effector-independent representations were not acquired. A rotation-specific decrease in movement time in the practice hand indicated movement-specific effector-dependent learning in AEP, but not in AIP and CP. Because deviations from optimal performance in the two-ball-rotation tasks are not as obvious as in other tasks, participants may not have detected such errors in AIP, hampering motor learning.