Methamphetamine-dependent adults showed differences in adjusting after stop errors even though their average stop-signal reaction time did not significantly differ from controls. A single summary measure of inhibition can miss how people adapt to unexpected events.
Harlé and colleagues studied 62 recently abstinent methamphetamine-dependent participants recruited from an inpatient program and 34 healthy controls. Participants performed a stop-signal task during functional MRI. A computational model estimated their changing expectation that a stop signal would occur and how surprising an outcome would be. Methamphetamine was not administered experimentally.
The average stop-error rate did not significantly differ between groups (p = 0.13). However, the interaction with stop-signal delay was significant: the dependence group was more likely to fail on difficult trials. Stop-signal reaction time was also similar between groups (p = 0.61). The researchers nevertheless found altered post-stop slowing, including reduced additional slowing after an error compared with a successful stop.
The dependence group showed reduced neural tracking of a model-derived surprise signal in selected cortical and subcortical regions. This offers a possible account of weaker adjustment when expectations are violated. The surprise measure was calculated from the task history; it was not a direct readout of awareness, intent or a blood methamphetamine concentration.
The groups differed in education and other substance-use histories. Task instructions also encouraged limiting excessive slowing, so reduced slowing could partly reflect how instructions were followed. The model did not separately resolve every possible contribution of uncertainty and environmental volatility. These limitations prevent a general claim that people with methamphetamine dependence cannot learn from mistakes. The supported conclusion is task-specific: adaptive responses to difficult and surprising stop events differed in this sample despite several nonsignificant overall comparisons.
See also why executive-function tests answer different questions.