Sleep Deprivation Affects Interference Control: A Diffusion Model Analysis

Previous studies suggest that interference control may be unaffected by sleep deprivation based on the unchanged interference effects (reaction time [RT] differences between incongruent and congruent conditions), while ignoring the overall slower RTs after sleep deprivation. In the present study, we interpreted these results from a new angle using a variant of diffusion model, diffusion model for conflict tasks (DMC), and investigated whether and how interference control is affected by sleep deprivation. Mathematical derivations and model simulations showed that unchanged task-irrelevant information processing (i.e., unaffected interference control) may not lead to the observed unchanged interference effects when considering the overall slower RTs after sleep deprivation (due to either decreased drift rate of task-relevant information or increased decision boundary). Therefore, the unchanged interference effects do not necessarily indicate unchanged interference control. We then conducted a Simon task following one night of sleep deprivation or normal sleep, and fitted the DMC to the data. Experimental results showed that the Simon effect was reversed when most of the trials were incongruent, indicating that participants used learned spatially incompatible stimulus-response associations to predict responses. However, the Simon effects in both mean RTs and RT distributions were not significantly modulated by sleep deprivation. Model fits showed that the drift rate of task-relevant information decreased and the time-to-peak of task-irrelevant activation increased after sleep deprivation. These results suggest that central information processing was degraded after sleep loss, and most importantly, task-irrelevant activation increased after sleep deprivation as interference control was impaired.