| |
Abstract:
The notion of functional specialization is exceptionally
useful in studying the psychophysics and neural basis of perceptual
capacities. Here a model for a possible specialization in audition
is proposed. However, unlike research in which perceptual
attributes (e.g. pitch, timbre) provide the desiderata for
specialized processing systems, the notion of a computational
specialization in the time domain is explored. Psychophysics and
neurophysiology point to the idea that time is processed in the
central nervous system as discrete, quantized chunks that form the
basis for coordinated information processing. In the visual system,
e.g., 40Hz oscillations (corresponding to temporal quanta with a
period of ~25ms) are regularly implicated. Suppose the temporal
quanta associated with left versus right auditory cortices differ.
Specifically, neurons in left auditory cortex integrate over
temporal windows of 25ms (40Hz), right hemisphere neurons over
temporal windows of ~250ms (4Hz). This asymmetric filtering in time
model (AFT) assumes that left and right primary auditory cortices
have access to the same input representations. However, nonprimary
areas differentially process input and construct representations
that highlight completely different aspects of the signal. The
proposal is similar to Ivry and Robertson (1998), who suggest that
perceptual hemispheric asymmetry follows from the differential
amplification of stimulus frequencies by the hemispheres,
conditioned by an attentional mechanism. AFT, however, does not
rely on attention; differing representations arise from the
different time windows over which information is extracted. AFT
predicts that suprasegmental phonological phenomena, particularly
syllable-level and prosodic phenomena, are preferentially
associated with right auditory cortical mechanisms.
|
