Modeling Early Auditory Processing Using GENESIS

HARRY R. ERWIN herwin at osf1.gmu.edu
Mon Jul 7 12:42:43 EST 1997


As a class project, I've been working on a GENESIS model of the neural
pathway from the inner hair cell to the bushy cells of the cochlear
nucleus. My advisor points out that the data I'd need for bat is mostly
lacking, and I'm looking for pointers from the auditory community: 

1. I'm using a vesicle release model at the inner hair cell that treats it
as a poisson process, with rates of 100 vesicles/second (no signal) and
700 (pressure wave present). This model is suggested by Mountain and
Hubbard's papers in Auditory Computation and by the know rates of
spontaneous spiking in spiral ganglion neurons. Are there any bat data on
this? The resulting conductance values at the type I spiral ganglion cell
synapse seem to be very noisy.

2. The type I spiral ganglion cell is modeled as a synaptic structure at
the inner hair cell, ten dendritic compartments, and a soma, where the
spikes are generated. The synapse is ignited by the inner hair cell
vesicle release process, and the resulting depolarization is passively
transmitted down the dendrite to the soma. Is this valid? The reduction in
the strength of the depolarization depends on the dendritic diameter, and
specific membrane and axial resistance and seems to be related to the
sensitivity of the cell. In any case, the cell starts spiking periodically
(500-1000 Hz) very soon after the arrival of the wave of depolarization,
bridging over and erasing any internal structure in the return. If
elements of the internal structure of the return are to be used in
downstream processing, they would have to be selected out by inhibiting
the early ignition of the synapse. 

Again, this model is calibrated by using information related to
spontaneous spiking rates in type I spiral ganglion neurons in cat and
rat. Has anyone studied the phenomenon in bat?

3. The End Bulb of Held/Bushy Cell complex is modeled as a soma, a short
dendritic tree, and a collection of excitatory axo-somatic synapses from
the type I spiral ganglion cell. These synapses fire over a period of
about 4 microseconds after the arrival of the action potential. (Valid in
bat?) This complex seems to be an edge detector. I get an initial spike
from the bushy cell at the beginning of a spike train from the type I
spiral ganglion cell, and then the follow-on spikes in that train seem to
prevent the bushy cell soma from resetting/respiking for a while.  Has
anyone looked at this in bat?

--
Harry Erwin, Internet: herwin at gmu.edu, 
Web Page: http://osf1.gmu.edu/~herwin 
PhD student in computational neuroscience (how bats echolocate)
Lecturer for CS 211 (data structures and advanced C++)
Senior Software Analyst supporting the FAA



More information about the Audiolog mailing list