Switching between cognitive control states? No, thank you.

Cognitive control enables adaptive behavior by flexibly adjusting information processing to meet changing environmental demands. While extensive research has examined the costs of switching between different tasks, a fundamental question remains unanswered: Does switching between different control states (relaxed and focused) within a single task carry intrinsic costs? We hypothesized that internally switching between control states while performing a single task incurs a subjective effort cost and therefore people should avoid such demands. Additionally, we examined the tradeoff between this cost and an established demand source—interference. We conducted a series of experiments combining the Stroop task with the Demand Selection Task where participants choose between different options. Participants consistently avoided options with high internal state switching (Experiments 1a, N = 64; 1b, N = 81) or frequent interference (Experiment 2, N = 86). Regarding the potential tradeoff, Experiment 3 (N = 121) revealed weak evidence that participants prioritized avoiding internal state switching over interference when both demands were manipulated. However, due to constraints imposed by the structure of the Stroop task, simultaneous manipulation of switching and interference was only possible with relatively weak manipulations for each. To address this limitation, we developed a novel parametric flanker task (Experiment 4, N = 50) enabling strong simultaneous manipulation of both demands in Experiment 5 (N = 130). With stronger manipulations, both switching and interference costs influenced selections, revealing trade-offs between these competing demands. Our findings refine models of control regulation, highlighting that adjusting control—not just exerting it—carries intrinsic costs.