The Organization of Global Motion and Transparency
Oliver Braddick and Ning Qian, in Motion Vision - Computational, Neural,
and Ecological Constraints, (J. M. Zanker and Z. Jochen, eds.),
Springer-Verlag, 2001, 86-112.
Click here to download the full paper
(PDF file, 0.19MB)
Abstract
The visual system has the task of computing global motions associated
with objects and surfaces. This task strongly involves extrastriate
brain areas, particularly V5/MT. Motion transparency provides a
particularly challenge for understanding how global motions are
computed and represented in the brain. Psychophysical experiments
show that, for a single region, multiple motions can be quantitatively
represented. However, at the most local scale, motion signals have a
suppressive interaction so that only a single motion can be
represented. Neurophysiological experiments show that this
suppression is a property of MT, not of V1, reflecting a subunit
structure within MT receptive fields and showing that transparency
perception is related to MT rather than V1 activity. A full
understanding of transparency perception and other global motion
phenomena will require us to understand how perceived motions are
related to the distribution of activity across a population of
directionally selective neurons, and how the brain implements the
representation of motion assigned to extended objects rather than to
specific retinotopic locations.
Back to Qian Lab Home Page