Title: Nonlinear
Neural Field Models for Spatio-temporal Receptive
Fields
Thomas Wennekers
Max-Planck-Institute for Mathematics in the Sciences
Abstract:
Neural Field Models are a
powerful tool in modeling spatio-temporal phenomena
in extended cortical tissue. Localized solutions in such models can be
interpreted as describing the dynamic behavior of cortical receptive fields (RFs). We present a recently developed approximation method
that enables an analytical treatment of localized solutions in nonlinear field
models of arbitrary dimension. This enables in especially the study of
interactions between several feature dimensions (space, orientation, spatial
frequency, etc.). The method reduces the full field dynamics to a small set of
dynamic equations for response amplitudes and tuning widths only. It provides
intuitive interpretations of qualitative spatio-temporal
response properties and establishes a close link to the dynamics of previously
studied small neural systems (e.g., the Wilson-Cowan oscillator). Furthermore, it relates the amplitude
dynamics of localized solutions/dynamic RFs to the
functional connectivity between different cell classes and layers.