Skip to main content


Functional magnetic resonance (fMRI) imaging

Education Achievements

Functional magnetic resonance (fMRI) imaging

Many IGERT students' research projects involve functional magnetic resonance (fMRI) imaging. In some cases, students who are primarily imagers take on a project in modeling or statistical analysis of fMRI data. In other cases, students who are primarily computational researchers use their IGERT training to conduct fMRI experiments. Daniel Leeds is one of the latter. Working in the lab of Dr. Michael Tarr at Carnegie Mellon University, Leeds studies visual object recognition in the brain. An important question is whether activity in visual areas correlates with theoretical models of visual encoding? Leeds found that a model called SIFT (Scale-Invariant Feature Transform) explains encoding of objects in mid-level cortical visual regions better than biologically-inspired models such as the Gabor filter bank and HMAX. But in earlier visual areas, activity better matched the more simplistic Gabor approach or the more holistic "shock graphs", confirming that the brain uses multiple codes.