Neural encoding of direction and distance across reference frames in visually guided reaching.

Abstract

Goal-directed actions require transforming sensory information into motor plans defined across multiple parameters and reference frames. Substantial evidence supports the encoding of target direction in gaze- and body-centered coordinates within parietal and premotor regions. However, how the brain encodes the equally critical parameter of target distance remains less understood. Here, using Bayesian pattern component modeling of fMRI data during a delayed reach-to-target task, we dissociated the neural encoding of both target direction and the relative distances between target, gaze, and hand at early and late stages of motor planning. This approach revealed independent representations of direction and distance along the human dorsomedial reach pathway. During early planning, most premotor and superior parietal areas encoded a target's distance in single or multiple reference frames and encoded its direction. In contrast, distance encoding was magnified in gaze- and body-centric reference frames during late planning. These results emphasize a flexible and efficient human central nervous system that achieves goals by remapping sensory information related to multiple parameters, such as distance and direction, in the same brain areas.Significance statement Motor plans specify various parameters, e.g., target direction and distance, each of which can be defined in multiple reference frames relative to gaze, limb, or head. Combining fMRI, a delayed reach-to-target task, and Bayesian pattern component modeling, we present evidence for independent goal-relevant representations of direction and distance in multiple reference frames across early and late planning along the dorsomedial reach pathway. Initially, areas encoding distance also encode direction, but later in planning, distance encoding in multiple reference frames was magnified. These results emphasize central nervous system flexibility in transforming movement parameters in multiple reference frames crucial for successful goal-directed actions and have important implications for brain-computer interface technology advances with sensory integration.