fMRI Investigation of Representation of Axis Structure

A core question in visual neuroscience is: what makes two shapes similar? And how, exactly, does the ventral visual cortex - our neural substrate for object recognition - encode the similarities we perceive? Several theories of object recognition suggest that our ability to perceive structure (i.e., the relationships between object parts) is at the root of our ability to generalize across views of an object and across exemplars of a category type.

Using fMRI multi-voxel pattern analysis, I am currently testing whether human lateral occipital cortex can distinguish between highly similar objects that differ only in the relative positions of their component parts.

(MRI subjects click here)

Cross-Cultural Behavioral Study

Do we all see the world the same way? I went to Africa in the summer of 2006 as part of an ongoing project in our lab to determine whether the brain's sense of geometry is innate or learned. We studied a tribe called the Himba.

The work I did was published in Visual Cognition in Jan. 2010; another related study from our lab appeared in Psychological Science in 2010, as well. Click the image to the left for figures from the paper, videos of the Himba, and some other related links.

The whole Himba shape recognition project was recently featured in a pop-level write-up on USC's web page. Also: A Himba mother and child will be featured in an upcoming documentary (due out in 2010, unrelated to our work).

Multiple-Unit Electrophysiology

How are shape-selective areas in the primate brain organized? During the summer of 2008, in collaboration with Takayuki Sato and Manabu Tanifuji at the Riken Brain Science Institute, I helped begin a (still ongoing) project to investigate how information is organized at different scales in macaque inferotemporal cortex (an area critical for shape processing).

We found electrophysiological evidence for "functional domains," several millimeters across, which had consistent preferences for particular classes of images. Out of a set of about 100 stimuli, nearby cells responded in a similar fashion to the same class of image (faces, body parts, tools).