The countermanding (or stop signal) task requires subjects try to withhold a planned movement upon the infrequent presentation of a stop signal. We have previously proposed a within-trial measure of movement cancellation based on neck muscle recruitment during the cancellation of eye-head gaze shifts. Here, we examined such activity after either a bright or dim stop signal, a manipulation known to prolong the stop signal reaction time (SSRT). Regardless of stop signal intensity, subjects generated an appreciable number of head-only errors during successfully cancelled gaze shifts (compensatory eye-in-head motion ensured gaze stability), wherein subtle head motion toward a peripheral target was ultimately stopped by a braking pulse of antagonist neck muscle activity. Both the SSRT and timing of antagonist muscle recruitment relative to the stop signal increased for dim stop signals and decreased for longer stop signal delays. Moreover, we observed substantial variation in the distribution of antagonist muscle recruitment latencies across our sample. The magnitude and variance of the SSRTs and antagonist muscle recruitment latencies correlated positively across subjects, as did the within-subject changes across bright and dim stop signals. Finally, we fitted our behavioral data with a race model architecture that incorporated a lower threshold for initiating head movements. This model allowed us to estimate the efferent delay between the completion of a central stop process and the recruitment of antagonist neck muscles; the estimated efferent delay remained consistent within subjects across stop signal intensity. Overall, these results are consistent with the hypothesis that neck muscle recruitment during a specific subset of cancelled trials provides a peripheral expression of oculomotor cancellation on a single trial. In the discussion, we briefly speculate on the potential value of this measure for research in basic or clinical domains and consider current issues that limit more widespread use.