Timothée Maniquet¹, Hans Op de Beeck¹; ¹KU Leuven, ²Leuven Brain Institute

One of the hallmarks of the human visual system is the functional specialisation of areas in the occipito-temporal cortex (OTC). Category-selective patches of cortex can be found where neurons preferentially respond to a given semantic category. Examples include the fusiform face area (FFA), extrastriate body area (EBA), and parahippocampal place areas (PPA), with selective responses to faces, bodies, and places, respectively. Across clusters of selectivity, super-clusters can be observed, in particular for the distinction between animate and inanimate categories. This animacy representation seems to be continuous, which has been interpreted as preference for categories closer or further away on an intuitive taxonomy scale, or alternatively as a reflection of how bodies and faces are coded as being more or less human-like. Here, we test the latter hypothesis through the combination of behavioural and neuroimaging methods. We recorded participants behaviour as they classify images either as animate or inanimate, or as faces, bodies and scenes, and tried predicting reaction times and gradual movement trajectories from univariate activity from FFA, EBA and PPA in response to the same images.  Our results show a nuanced pattern whereby FFA activity is strongly predictive of reaction times and mouse positions, more so than EBA and PPA. While EBA and PPA can both predict increases in reaction times for non body and non scene stimuli, respectively, only FFA activity can accurately predict decreases in reaction times for stimuli rated as more face-like. Additionally, FFA surpasses others in predicting patterns of the movement trajectories during animacy classification. Overall, we find partial support for the face & body account of animacy selectivity in the cortex. We interpret our results in light of theories of OTC organisation.