Decoding categories shared by the face and body

Humans can easily categorize images of human bodies and faces according to their sex/gender and weight. Objectively speaking this is a difficult task due to category-independent variability in image size, lighting, facial expression and body sex and weight of faces and bodies, and whether the representations would be face- or body-selective. We used fMRI to record the brain activity of subjects viewing faces and bodies that varied in sex, weight, and image size (factor 2). Using multivoxel pattern analyses, we found that the extrastriate body area (EBA), fusiform body area (FBA) and occipital face area (OFA) consistently discriminated bodies of different sexes, including in a cross-classification analysis where training and test data were based on different stimulus sizes. Body weight could be decoded in OFA and FFA, size-invariantly in the latter. When voxels of body-regions were pooled, the sex and weight of bodies could be decoded invariant with respect to image size. No region consistently decoded the sex or weight of faces, nor did face-related decoding work when voxels were pooled across face- or bodyselective regions. We hypothesize that this may be due to the fact that neither weight nor sex appeared very prominently in controlled face stimuli used here (e.g. excluding hair). We conclude that information relating to the body categories sex and weight is found in both body and face responsive brain regions, but that size-invariant information is mostly located in body responsive regions