Representing action in terms of what not to do: Evidence for inhibitory coding during multiple action control

Performing two actions at the same time (vs. in isolation) usually results in performance costs. However, recent studies have reported that it is also possible to observe dual-action benefits, a finding that challenges standard theories of multiple action control. This issue is typically resolved by assuming that under certain circumstances, performing only one of two possible actions-cognitively represented in terms of what to do-necessitates the costly stopping of the initially activated but unwarranted second action. Here, we test this hypothesis against an alternative inhibitory coding account which rests on the assumption that actions might be cognitively represented in terms of what not to do. Across four experiments, participants responded to a single stimulus with either single or dual responses in the manual and vocal domains, while the relative frequency of response types was systematically manipulated. The results revealed robust dual-action benefits in manual response times and error rates, and the pattern across experiments clearly supported the novel inhibitory coding framework. Crucially, this implies that even though the motor actions required for single and dual responses are physically the same, they are represented very differently. Specifically, dual responses can be represented holistically (noncompositionally). Overall, these findings demonstrate an astonishing flexibility in the mental representation of behavior demands. (PsycInfo Database Record (c) 2021 APA, all rights reserved).