A Laboratory for Neural Computation Publication
In NIPS vol. 10 , M.I. Jordan, M.J. Kearns,
S.A. Solla, (Eds.), Cambridge, MA: MIT Press, 1998,
pp. 208-214.
Toward a Single-Cell Account of Binocular Disparity Tuning:
An Energy Model May be Hiding in Your Dendrites
Bartlett W. Mel ,
Biomedical Engineering Department,
USC
Daniel L. Ruderman
, Salk Institute
Kevin A. Archie
, Neuroscience Program,
USC
ABSTRACT
Hubel and Wiesel (1962) proposed that complex cells in visual
cortex are driven by a pool of simple cells with the same
preferred orientation but different spatial phases. However, a
wide variety of experimental results over the past two decades
have challenged the pure hierarchical model, primarily by
demonstrating that many complex cells receive monosynaptic
input from unoriented LGN cells, or do not depend on simple
cell input. We recently showed using a detailed biophysical
model that nonlinear interactions among synaptic inputs to an
excitable dendritic tree could provide the nonlinear subunit
computations that underlie complex cell responses (Mel et
al. 1997). Our present work extends this result to the case of
complex cell binocular disparity tuning, by demonstrating in an
isolated model pyramidal cell (1) disparity tuning at a
resolution much finer than the the overall dimensions of the
cell's receptive field, and (2) systematically shifted optimal
disparity values for rivalrous pairs of light and dark
bars---both in good agreement with published reports (Ohzawa et
al., 1997). Our results reemphasize the potential importance
of intradendritic computation for binocular visual processing
in particular, and for cortical neurophysiology in general.
Click here for preprint.