| |
Abstract:
Auditory cortical processing is often modeled as a
hierarchical, feed-forward analysis of incoming signals by neurons
that act as feature detectors. Previous computational models have
not accounted for the plasticity of cortical neurons or the effects
of neuromodulation on neuronal receptive fields. Recent
neurophysiological studies have shown that stimulation of the basal
forebrain during presentation of acoustic stimuli leads to massive
reorganization of the auditory cortex. This reorganization is
stimulus-dependent; auditory feature maps appear to be warped to
enhance the representation of spectral and temporal properties of
the presented stimulus. The aim of this study was to develop a
simple connectionist model of auditory cortical processing that can
be used to investigate the role of basal forebrain modulation in
plasticity. A number of mechanisms whereby neuromodulation could
affect plasticity have been proposed including (a) the
amplification of behaviorally relevant inputs, (b) the inhibition
of less relevant inputs, (c) the modulation of intracortical
signal-to-noise ratios (e.g., through lateral inhibition), and (d)
the facilitation of synaptic plasticity. A self-organizing feature
map served as a cortical model, and the effects of neuromodulation
were modeled using parameters intrinsic to this connectionist
architecture. Simulations examined how parameter modulation
affected the structure of representational maps, and whether
reorganization in the model was comparable to the reorganization
observed in the auditory cortex.
|
