Dissociating Task Selection and Response Selection in Dual-Task Contexts: Evidence from a Novel Trial-by-Trial Analysis of Temporal Overlap between Tasks

This study examined the effect of temporal overlap in dual-task processing on task switch costs. Participants performed a psychological refractory period (PRP) experiment with a varying stimulus onset asynchrony (SOA), which is the time interval between the onsets of the stimuli for Task 1 (T1) and Task 2 (T2). Trials included task repetitions, where T2 was identical to T1, and task switches, where T2 differed from T1. T2 performance was worse with shorter SOAs than with longer SOAs and in switch trials than in repetition trials, indicating a PRP effect and switch costs. Notably, switch costs were not modulated by the SOA. However, SOA cannot precisely determine whether T1 and T2 are performed with or without temporal overlap in a given trial. To distinguish between these trials, we assessed the time interval between the T1 response and the onset of the T2 stimulus, known as the response-stimulus interval (RSI). The RSI acts as a proxy to temporally localize when T1 response selection is completed and the response-selection bottleneck is released, This novel trial-by-trial approach revealed that switch costs did not differ between trials with and without temporal overlap in task processing. Moreover, RSI was found to predicted T2 performance more accurately than SOA. Using RSI as a predictor of RT2 provides persuasive evidence that task selection and response selection rely on independent cognitive processes. Alternatively, both processes use shared central processing limitations, but temporal and/or strategic factors prevent these processes from overlapping in time and thereby interfering with each other.