| WHO WE ARE |
NeuroVia is an NIH- and NSF-funded research group at the University of Minnesota,
located at the Minneapolis VA Medical Center, whose major interests include scientific
visualization, computational anatomy, PET and fMRI data analysis, and neuroinformatics
with particular emphasis on neuroscience and functional neuroimaging.
Long-term support from the Human Brain Project has fostered interdisciplinary
collaborations with the Department of Mathematics at Florida State University, the
International Consortium for Brain Mapping at UCLA, the Akita Research Institute for Brain
and Blood Vessels (Akita, Japan) and the Department of Mathematical Modeling at the Danish
Technical University (Copenhagen, Denmark).
In addition to basic-science-oriented research programs, Neurovia runs an active clinical
research program in Parkinson's disease at the Minneapolis VA Medical Center. Neurovia
investigators include physicians, nurses, physicists, mathematicians, neuroscientists,
informatics specialists, and computer scientists, as well as doctoral students in
biomedical engineering and computer science.
|
| WHAT WE DO |
Current research efforts are focused on the physiological basis of the BOLD fMRI signal, the
use of resampling techniques for analyzing neuroimaging data, the evaluation of software tools for
the brain imaging community, and the creation of a surface cartography of the brain.
Dr. Seong-Gi Kim and his colleagues at the Center for Magnetic Resonance Research are
studying the relationship between neural activity and the hemodynamic response as a function
of stimulus duration. Dr. Stephen Strother and his colleagues at the Danish Technical Institute
are implementing a software environment with a flexible Java-based interface that includes a
statistical resampling framework and links to heterogeneous software tools for fMRI data processing
written in IDL, MATLAB and C. Dr. David Rottenberg and his colleagues at the University of Minnesota
and Florida State University are testing the hypothesis that 2D surface-based representations of spatial
activation patterns have greater predictive power than their corresponding 3D volumetric
representations by comparing 2D and 3D patterns of cerebral and cerebellar activation in
multi-subject fMRI data sets using overlap, probabilistic, and distance metrics.
|