The effect of temporal accuracy constraints on movement-related potentials.

Both self-paced movements (internally generated) and movements paced by a fixed interval cue (externally cued) are preceded by a slow-rising movement-related potential (MRP) of similar timing, magnitude and topography. When the timing of the external cue is variable (temporally unpredictable), this MRP is absent. These findings have been interpreted to suggest that MRPs reflect neural activity mediating the preparation of temporally predictive movements, irrespective of whether the movement is internally generated or externally cued. However, the apparent similarity between MRPs preceding self-paced and predictably cued movements might be explained by the absence of control for the timing of movement onset, that is, MRPs preceding regularly-paced cues may simply reflect activity associated with self-paced movements initiated at times that more or less coincide with the timing of the imperative cue. The objective of this study was to reexamine the comparison of MRPs preceding temporally predictive (self-paced and predictably cued) versus reactive movements. To circumvent the potential confound of movement onset timing, constraints were placed on the temporal accuracy of movements cued by a regularly-paced tone. This design permitted post-hoc classification of trials into predictive or reactive movements to the tone. Three movement initiation conditions were tested: (1) self-paced (SP), (2) in response to an irregularly-paced cue (IC), and (3) in response to a regularly-paced cue (RC). In the latter condition, subjects were trained to initiate movement to within less than one simple reaction time in at least 50% of trials, and MRPs were compared between movements that were initiated "too early" (predictive), "too late" (reactive), or were temporally accurate (predictive). Cerebral potentials were recorded from 87 scalp surface electrodes. Consistent with previous studies, an early slow-rising MRP with maximum negativity over the midline frontal cortex was present only when the timing of movement onset could be predicted in advance (SP and RC conditions). Moreover, MRPs for movements that were temporally accurate or were initiated "too early" were significantly larger than the MRPs that preceded SP movements (P < 0.018). In contrast, movements initiated in reaction to the cue (IC condition), even when the timing of the cue could be predicted in advance (movement initiated "too late" in the RC condition), were associated with a significant attenuation of premovement activity (P < 0.005). Differences between conditions (RC > SP > IC) were significantly greater over the midline frontal cortex than the contralateral or ipsilateral sensorimotor cortex (P < 0.038). These findings show that the imposition of accuracy constraints on the timing of movement onset markedly enhances preparatory activity in the cortical or subcortical networks that mediate the predictive initiation of movement.