Dynamic interaction between “Go” and “Stop” signals in the saccadic eye movement system: New evidence against the functional independence of the underlying neural mechanisms

Abstract

We investigated human oculomotor behaviour in a Go–NoGo saccadic task in which the saccadic response to a peripheral visual target was to be inhibited in a minority of trials (NoGo trials). Different from classical experimental paradigms on the inhibitory control of intended actions, in our task the inhibitory cue was identical to the saccadic target (used in Go trials) in timing, location and shape—the only difference being its colour. By analysing the latency and the metrics of saccades erroneously executed after a NoGo instruction (NoGo-escapes), we observed a characteristic pattern of performance: first, we observed a decrease in the amplitude of NoGo-escapes with increasing latency; second, we revealed a consistent population of long-latency small saccades opposite in direction to the NoGo cue; finally, we found a strong side-specific inhibitory effect in terms of saccadic reaction times, on trials immediately following a NoGo trial. In addition, we manipulated the readiness to initiate a saccade towards the visual target, by introducing a probability bias in the random sequence of target locations. We found that the capacity to inhibit the impending saccade was improved for the most likely target location, i.e. the condition corresponding to the increased readiness for movement execution. Overall, our results challenge the notion of a central inhibitory mechanism independent from movement preparation. More precisely, they indicate that the two mechanisms (action preparation and action inhibition) interact dynamically, possibly sharing spatially-specific mechanisms, and are similarly affected by particular contextual manipulations.