The estimated latency of the unobservable stop response, the so-called stop-signal reaction time (SSRT), has been the established measure of performance in the stop-signal task. While it is currently debated whether SSRT is a suitable marker of inhibition performance, other markers such as the reliability of triggering the stop process (“stop trigger failures”) are coming into focus. In the present study, we elucidated the mechanisms associated with trigger failures using a model-based neuroscience approach by means of functional magnetic resonance imaging for the first time. To this end, we used a large, open-access fMRI data set to investigate the relationship between the probability of trigger failures and fMRI signal change in a stop-signal task in healthy adults (n = 113). Stop trigger failures were associated with less activity in the substantia nigra during unsuccessful stopping and with less activity in the subthalamic nucleus (STN) region during successful inhibition. Although stop trigger failures strongly correlated with SSRT, we found only little evidence for a correlation between SSRT and stopping-related fMRI signal. Thus, in particular, the reliability of the stop process and not its estimated latency depends on the recruitment of key nodes within the prefrontal-subthalamic hyperdirect pathway. More specifically, stop trigger failures may be linked to inadequate substantia nigra innervation at the neural network level. As current evidence suggests that the hyperdirect pathway is engaged by the processing of salient stimuli, deficiencies in assessing the relevance of the stop signal may represent a phenotype associated with a propensity to trigger the stop process unreliably.