Preventing visual feedback of the moving limb during grasping preferentially activates parietofrontal premotor areas.

Accurate control of grasping relies on the brain's ability to estimate the position and configuration of the arm and hand, integrating sensory feedback with motor-based predictions derived from a copy of the movement command. Visual feedback of the moving limb plays an important role in calibrating these internal estimates. When visual feedback is limited, the brain must rely more heavily on proprioception and motor-based predictions. In this study, we use fMRI to test for cortical regions that are preferentially involved in grasp control when visual feedback of the moving limb is selectively prevented. Twenty healthy participants perform reaching-to-touch and reaching-to-grasp actions with and without visual feedback. Critically, target objects are always visible, fixed to illuminate only their surface properties. This contrasts with prior studies that remove vision of both the limb and the target, thereby confounding non-visual control with increased working-memory demands. Our findings reveal a distributed network of posterior parietal and frontal premotor areas that are preferentially active for grasping when visual feedback of the moving limb is unavailable. We interpret these results as reflecting an important role for these areas in maintaining an internal representation of the state of the arm and hand for online grasp control.