Abstract
Establishing the functional role of brain regions will require a better understanding of the interactions between stimulus information in the input and the subset of this information the brain requires to categorize the stimulus as one thing or another (Schyns, 1998; Pernet et al., 2007). Of the distributed face-preferential regions there are inconsistencies regarding their functional roles. To illustrate, while some studies indicate that the OFA is crucial for processing of identity (Dricot et al., 2008; Kadosh et al., 2011), whereas others suggests that it is involved in processing of generic facial features (Gilaie-Dotan et al., 2010). We addressed the functionality of face-preferential regions in a fast event-related fMRI experiment using different categorizations of faces and objects (male vs. female, or fruit/vegetable decisions on the one hand, identification of exemplars on the other) and parametrically varied the stimuli in terms of their Spatial Frequency (SF) content. We first measured observers’ behavioral performance with each combination of categorization and stimulus, across the different levels of SF content. To ascertain which brain regions code the categorical information, we sampled voxels from the bilateral OFAs and FFAs, built similarity matrices for each conjunction of categorization task and stimulus set (e.g. gender/faces vs. identity/faces) and tracked the regions whose categorical discrimination matched the observers’ behavioral performance. Neural representations generally differentiated with increasing stimulus resolution, but this differentiation followed the specific categorization considered (e.g. gender coding in gender and identity coding in identity task). These findings are in agreement with studies showing stimulus- and/or task-dependent modulation of activation within category-specific regions (e.g. Joseph et al., 2006; Cohen Kadosh et al., 2010) and task-context dependent responses in sub-populations of IT neurons (Pauls & Logothetis, 2002).
Meeting abstract presented at VSS 2013