Research¶
Computational methods for modeling neural representational spaces¶
We are developing computational methods for aligning neural representational spaces across subjects at a fine spatial scale. Using these methods, we can now use multivariate pattern (MVP) analyses to build a model on a group of subjects and use that model to classify responses in a new subject. We are using these methods to build a common high-dimensional model of the representational space in ventral temporal (VT) cortex for complex visual stimuli.
Representation of faces and person knowledge¶
The distributed neural system for face perception, recognition of familiar faces, and the role of face perception in social communication are longstanding research themes in the lab. Current projects include investigations of the automaticity of face recognition and activation of person knowledge – to what extent can these processes be accomplished with minimal attention or without conscious awareness.
Representation of animate entities¶
The ventral object vision pathway in the human brain appears to have a lateral-to-medial topography in ventral temporal cortex that reflects a distinction between the representation of animate and inanimate stimuli. A similar distinction is found in lateral temporal cortex (superior temporal sulcus to middle temporal gyrus). MVP analysis of suggests that the representations of animate entities in ventral temporal cortex embodies semantic knowledge of animal species.