Abstract
Patches of face-selectivity in macaque inferotemporal cortex have been identified with fMRI for over a decade, and have since been investigated with great interest. However, little is known about the mechanisms that bring about intersubject variability in the location and degree of activation in these face patches. Why do they vary in location? Why do some monkeys lack some of these patches altogether, while others demonstrate very strong selectivity? The neural architecture that may underlie an individual macaque's specific and idiosyncratic activation pattern remains unexplored. Since connectivity is the principle aspect of neural architecture that defines what a brain region is capable of computing, we hypothesized that connectivity should be strongly predictive of macaque face patches. We used intrinsic functional connectivity to model and predict the location and activation strength of each monkey's face patches, in two monkeys. The resulting model describes the connectivity fingerprint for face selectivity in the macaque brain, offering potential targets for optogenetic or pharmacological manipulations outside of the traditional set of face patches. Importantly, this approach allows researchers interested in electrophysiological experiments to infer the location of face patches using an individual macaque's connectivity data, which can be acquired with minimal or no training, saving researchers years of experimental preparation.
Meeting abstract presented at VSS 2018