Overview
Nonlinear Systems Analysis with Application to Localizing VEP Sources
Recent advances in functional imaging of brain activity using PET and fMRI
have created much excitement among neuroscientists. Researchers are now able to
extend their study of brain processing to humans. Unfortunately, the new methodologies
do not have adequate temporal resolution. Our recent research has produced a
breakthrough in applying nonlinear analysis to the visual evoked potential (VEP)
and magnetoencephalogram (MEG) enabling the underlying neural generators to be
isolated with temporal resolution 1000 times better than is possible with either
fMRI or PET and with comparable spatial resolution. The new approach has the potential
of isolating the signals from each of the separate retinotopically organized visual
areas. Information on the recent advances can be found at Localization in Time and Space of Evoked EEG and MEG Sources.
For details on our earlier research see the Nonlinear Systems Analysis and Visual Evoked Potentials publications, as well as associated "Research Commentaries".
Developing a Model of Spatial Vision
The visual system consists of an enormous number of spatial filters with different
positions, sizes, orientations and bandwidths. These filters are arranged in sequential
stages with nonlinear interactions among and within the stages. In collaboration with
Dennis Levi's group we do psychophysical experiments to characterize these stages of
visual processing and their interactions. The role of top-down feedback is especially
interesting because we intend to carry out parallel experiments with VEP/MEG/fMRI source
localization. We also study the effects of visual noise and the effects of visual learning
on human performance. One result of this research is that we now hold the Guinness record
for position acuity. This modeling research has applications to image compression. For
further details see the Spatial Vision and Applied Vision publications together with
the associated "Research Commentaries".
