Abstract
Simple physical actions, such as hand gripping, impose measurable mental costs, impairing attention, memory, and decision-making. However, the mechanisms underlying this action-cognition trade-off remain elusive. A resource-sharing account posits that action and cognition draw resources from a common pool; thus, engaging the muscular system may reduce one's ability to actively retain information in mind, resulting in a working memory (WM) retention cost. In contrast, a control-cost account suggests that physical exertion primarily increases demands on control-related processes such as distractor inhibition without reducing overall WM retention. We tested these different accounts across two experiments, both of which consistently showed that concurrent physical load impaired visual WM performance at the behavioral level, especially in the presence of task-irrelevant distractors. In Experiment 1, EEG recordings revealed that stronger concurrent handgrip force did not reduce the contralateral delay activity (CDA), a neural marker of WM retention. Instead, higher physical load increased CDA amplitude when more distractors were present, consistent with increased retention of task-irrelevant information during concurrent physical exertion. In Experiment 2, fMRI revealed that this interaction was preferentially expressed within a frontoparietal network, encompassing the bilateral inferior frontal and posterior parietal cortices, rather than sensory and motor cortices associated with visual input and physical action. Together, these findings indicate that the cognitive cost of physical exertion arises not from an overall reduction in WM retention, but from increased demands on control-related processes that regulate which information gains access to memory, leading to greater inclusion of task-irrelevant content under elevated physical load.