Ill
at the resting potential for unstimulated rods. This potential could change from cell to cell,
and the -40 mV we assumed is likely up to 5 mV more negative than the average rod resting
potential in darkness. This could cause a small error, but the estimates of 7 = Rc∣Rm with
the dual patch recordings agree well with our estimates using the light stimulus.
5.7.3 Significance of rod-rod coupling
It is well known that rods are coupled to one another in several species, but the consequences
of the coupling have not been thoroughly examined quantitatively. We use the coupling data
collected in the first sections to create a first-order 2D model of the rod network (figure 5.6).
We analyze the model response using spectral analysis to show how the spatial frequency
components of visual scenes will be represented in the salamander retina (figure 5.11). This
analysis shows that coupling reduces high frequency components. Features with a wave-
length of ≈ 50 microns would be attenuated to half the amplitude of a whole-field stimulus.
This coupling would reduce the response to sharp contrasts in a visual scene, which are
composed of high frequency changes.
While rod-rod coupling has the disadvantage that it causes attenuation of rapid changes
in contrast in an image, its primary advantage of reducing noise in photoreceptors is a worth-
while tradeoff. One of the sources of intrinsic noise in photoreceptors is the photocurrent,
which drives the photoreceptor response to light. The photocurrent noise has two com-
ponents: continuous noise, which comes from variations in phosphodiesterase (PDE) ac-
tivity, and discrete noise, which comes from spontaneous thermal activation of rhodopsin