Object Representations
How are objects represented in the brain? How do these representations change as information progresses along the visual processing pathway?

All the images of the same object embedded in an n-dimensional space make up that object's manifold. Each view or image of the object is represented as a point along the manifold in such a way that relationships between points are maintained across all lower-dimensional mappings. For example, within the manifold for a mug, a snapshot of the mug at 10º will occupy a point along the manifold that is closer to the point representing the mug rotated at 15º than the point at which the mug is rotated at 45º. Inspired by DiCarlo & Cox (2007), we are interested in exploring whether manifolds and similar methods can be applied to brain patterns as a way to uncover how objects might be represented as visual information moves along the brain's processing stream.

Poster presented at the annual meeting of SfN in November 2016 (PDF)



Distributional Learning
Without any feedback or explicit instruction, how do the distributions for specific features affect how people learn about a group of novel objects?

Project description coming soon.