Cross, E. 1, 2, 3 , Liepelt, R. 1, 4 , Prinz, W. 1 & Hamilton, A. F. d. C. 5
1 Department of Psychology, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany
2 Behavioural Science Institute & Donders Institute for Brain, Cognition & Behaviour, Radboud University Nijmegen, The Netherlands
3 Department of Psychology, Bangor University, Wales, UK
4 Institut für Pscyhologie, Westfälische Wilhelms Universität Münster, Germany
5 School of Psychology, University of Nottingham, Nottingham, UK
In the human brain, top-down modulation has a substantial impact on whether information is processed as ‘social’ or ‘non-social’. Theory of mind studies have used the same stimuli for ‘human’ and ‘computer’ conditions, and found that the identical stimulus activates social brain regions when participants believed the observed action originated in another person, not a computer. Here we tested for modulation of actions perceived as social/non-social, based on instructions and the actor’s form. Before scanning, participants watched a video instructing how avatars could be animated entirely by computer or by capturing motions of a real human actor. During scanning, participants saw computer-generated ‘Poser’ figures (one human-like and the other robot-like in appearance) performing simple goal-directed actions. Both figures performed the same actions with identical kinematics, but participants were told that half the videos were created with motion capture, and the other half with computer animation. Imaging data reveal stronger occipital activations when perceiving the robotic versus the human form, and a broad network of regions including inferior occipital, fusiform, and posterior cingulate cortices when participants believed they were watching human-generated actions. The present data thus demonstrate how top-down (instructions) and bottom-up (agent form) features work together to modulate action perception.