Abstract
Purpose:
To characterize contrast-dependent gain changes in the crab eye as a function of mean illumination level and investigate their physiological origins.
Methods:
A visual stimulation system was built to drive single or multiple ommatidial receptors of the horseshoe crab eye, consisting of an optical coupler that focused light from a computer monitor onto a 150um or larger optical fiber. The monitor output was modulated according to random binary sequences having different mean luminance and luminance variance. Single optic nerve fiber responses to the white noise stimulus were then recorded from live male horseshoe crabs and saved to computer. The spike train records were analyzed in terms of a linear-nonlinear model. In other experiments, electroretinograms (ERGs) were recorded from the eye stimulated by a linearized LED modulated by the same random binary sequences. ERG records were also evaluated by white noise analysis.
Results:
Optic nerve recordings from the horseshoe crab eye demonstrate existence of variance adaptation. The sensitivity of the adaptive process to stimulus contrast was the same across mean light levels ranging from 0.6 to 60 cd/m2. Single ommatidial and whole eye illumination gave similar results, indicating that this process does reside in lateral inhibitory synapses of the retinal network. ERG recordings did not show evidence of variance adaptation, implying that contrast-dependent gain changes originate in post-receptoral mechanisms.
Conclusions:
Variance adaptation does not depend on mean light level over the photopic to mesopic range. Its origin appears to lie in eccentric cells, which encode photoreceptor signals in optic nerve spike trains. Possible mechanisms include spike generation, self-inhibition, or a heretofore unknown process.