Zelal Eltas¹, Murat B. Tunca^1,2, Burcu A. Urgen¹; ¹Bilkent University, ²Justus Liebig University Giessen

Perceiving the movement direction of observed actions is essential for interpreting intentions and guiding social interactions. Direction selectivity has been widely studied in motion perception literature using simple stimuli like dots, gratings, or point-light displays (PLDs). However, little is known about how the brain encodes direction during the natural, repetitive actions seen in daily life. To investigate this, a set of novel video stimuli was created using four classes of action: (i) object manipulation (stretching, shaking), (ii) object manipulation using a tool (wiping, painting), (iii) self-directed actions (scratching, rubbing), and (iv) self-directed actions using a tool (brushing, drying). Each action was recorded in three spatial directions (left-right, up-down, and front-back) with two versions and two actors, resulting in a total of 96 naturalistic action videos. 25 participants took part in an event-related fMRI experiment while they observed these action videos and made a judgment about movement direction. The brain activity was analyzed using multivariate pattern analysis (MVPA) and multiple regression representational similarity analysis (RSA). MVPA results suggested that action direction could be classified above chance levels in occipital, parietal, and motor cortices. Specifically, the highest decoding accuracy was observed in occipital, primary motor, and somatosensory regions. Importantly, RSA results demonstrated that direction information is still represented in early visual, occipito-temporal, parietal, and motor regions after controlling for low-level visual and motor features. These findings indicate that the movement direction of actions is represented at multiple levels of the action observation network (AON), extending from early sensory regions to higher-order parietal and frontal areas. More importantly, this study suggests that direction selectivity is a core feature of the action observation network, linking the processing of basic motion with the understanding of everyday actions.

Acknowledgements: This project is funded by Turkish Health Institutes Presidency (TUSEB) A2 Project Supporting Program (Project No: 37734).