Abstract
Our interactions with the environment require the smooth execution of object-directed actions. These actions seem to be represented not only in terms of specific muscular patterns, but also in terms of the achievement of a more general goal, e.g. grasping an object irrespective of the means (e.g. the adopted effector). Neuroimaging studies support the idea that actions might be represented at different levels of abstraction, i.e. generalizing across different motor features. Recent studies using MVPA of fMRI data showed action representations tied to the hand performing the action (i.e. at an effector-dependent level), but also generalizing across left and right hand (i.e. at an effector-independent level). Here we used MVPA of fMRI data to examine action representations that generalize across effector, hand orientation, or both. Participants were instructed to perform different non-visually guided object-directed actions. The experimental conditions were embedded in a 2x2x2 factorial design with the following factors: action type (precision grip, whole hand grip), hand orientation (0°, 90°) and effector (left hand, right hand). To investigate the encoding of action representations at different levels of abstraction, we used a whole-brain cross-condition MVPA searchlight approach. Specifically, we classified the type of action within and across hand orientation and effector. Our results showed widespread encoding of specific effector-dependent actions (within effector and hand orientation) in the fronto-parietal prehension network. Several regions also showed encoding for more abstract effector-specific actions, generalizing across hand orientation. Finally, a more circumscribed set of cortical areas, comprising posterior parietal, premotor and lateral occipitotemporal regions, hosted effector-independent and orientation-independent action representations. These data widen our understanding on how actions are represented in the brain by demonstrating that, even during movement execution, actions are encoded following a hierarchical organization spanning different levels of abstraction.
Meeting abstract presented at VSS 2016