Abstract
Numerous fMRI studies have examined the representation of faces and headless bodies in the face-selective and body-selective brain areas. However, in real life our visual system is exposed to the face and body attached together and it is therefore of interest to understand how face and body-selective areas represent the whole person. The few studies that examined this question tested whether the response to the whole person equals or deviates from the mean of the response to the face and body. However, a deviation from the mean may reflect several possible responses: The whole person may be equal the max response of the preferred stimulus (i.e., the response to the face alone in face areas or body alone in body areas); a weighted mean of the response to the face and body in which the weights differ from 0 and 1 (non-max responses) or the sum of the response to the face and body. Here we used two methods that allowed us to directly examine each of these different models: a single-voxel based approach that models the response to a person as the sum of responses to the face and the body with an interaction term; a multi-voxel based approach that models the response to the person using a linear combination of the responses to the face and the body (Reddy et al. 2009). Both methods revealed very similar findings. The fMRI response to the whole person was consistent with a weighted mean response, with most responses close to the max response, indicating a large influence of the preferred stimulus on the response to the whole person. These findings suggest relatively little influence of the response of the non-preferred stimulus on the preferred stimulus even when both are parts of the same stimulus.
Meeting abstract presented at VSS 2018