Transformations of cognitive maps for sensorimotor control

Abstract Adaptive embodied behavior involves transforming structured knowledge about the relationship between environment and action into motor signals, but how these transformations are coordinated across brain networks remain unknown. Participants learned associations between visual cues and isometric exertions that varied in force and duration, forming a two-dimensional cognitive map of a force–time space. During behavior, this force-time space was expressed in several cortical regions using grid-like coding schemes, indicating sensorimotor cognitive maps. Importantly, while mnemonic regions such as the entorhinal cortex maintained an unwarped, task-relevant representation, the primary motor cortex encoded a force-time space distorted by perceived effort during motor execution. Dynamic causal modeling showed inhibitory motor-to-mnemonic coupling that predicted the transformation of effort-weighted motor signals into sensorimotor maps. Furthermore, individual differences in learning and navigating the force-time space independently shaped mnemonic map geometry and perceived effort. These findings demonstrate that sensorimotor cognitive maps emerge from dynamic interactions between motor and mnemonic systems and are shaped by individual differences during the learning and execution of movement.