Purchase this article with an account.
Mona Rosenke, Nicolas Davidenko, Kalanit Grill-Spector, Kevin Weiner; Differential responses across body- and face-selective cortex predict visual categorization behavior. Journal of Vision 2018;18(10):1091. https://doi.org/10.1167/18.10.1091.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Humans categorize objects remarkably fast. Prior research shows that neural responses within a single region are correlated with categorical judgments of visual stimuli, which likely contributes to this efficient behavior. However, it is presently unknown if not just one, but instead, several functional regions that are adjacent in cortex may work together to perform categorical judgments. Here, we leveraged the fact that regions selective for faces and bodies are adjacent in human ventral temporal cortex (VTC) to test if and how responses from both types of regions contribute to categorical judgments. To do so, we generated a novel set of parameterized silhouette stimuli that spanned a continuous morph space between faces and hands, while controlling for low-level image properties. We defined stimuli at 5 morph levels, ranging from fully face-like (level-1) to fully hand-like (level-5), and behaviorally calibrated intermediate (level-3) stimuli to appear equally face-like and hand-like in a large group of participants (N>60). Using these stimuli, we conducted two types of experiments in 14 independent participants: (i) an fMRI block-design experiment during which we measured mean responses in face- and body-selective regions in VTC and (ii) a behavioral categorization experiment. This two-pronged approach allowed us to examine the relationship between neural responses and behavioral categorization. We report three main findings. First, adjacent face- and body-selective regions in VTC illustrate functionally distinct neural tuning to face-hand morphs. Second, neural tuning is correlated with behavioral categorization judgments (R2 range: .35 - .54). Third, a linear regression model reveals that the combination of neural responses of face- (b=.31) and body-selective (b=-.22) regions best predicts human judgements (cross-validated R2: .70±16). Together, these findings support a new idea in which the differential response between regions selective for different domains more accurately explains human categorical judgements than neural responses from one domain-selective region alone.
Meeting abstract presented at VSS 2018
This PDF is available to Subscribers Only