The output from the retina comprises the activity of 10-15 distinct classes of retinal ganglion cells, each optimized for specific spatial and temporal properties. Information transfer over the limited bandwidth available to retinal neurons is optimized by mechanisms that remove redundant information. One hypothesis for such a mechanism is predictive coding, which collects luminance signals from the surrounding regions and subtracts them from the center response, thereby removing correlations, and enhancing the signal-to-noise ratio and transmission of information. We hypothesize that predictive coding can also be implemented by the subtraction of more complex statistics such as contrast, color, motion, or orientation. Surround antagonism, generated first in the outer retina by horizontal cell feed-back onto photoreceptors, propagates to bipolar cells and therefore must be inherent in all ganglion and amacrine cells, but there it is mixed with a surround generated from the inner retina. Feedback from amacrine cells onto bipolar cell terminals and feed-forward from amacrine cells onto a ganglion cell's dendrites endow its surround with non-linear properties. Both outer and inner retinal surrounds are fundamental for the function of vision, but their relative roles are unknown. We pro- pose to test several hypotheses about amacrine and ganglion cell surrounds. We will record from live amacrine and ganglion cells, measure the spatio-temporal extent of their inner and outer retinal surrounds, and using blockers of neurotransmitters GABA and glycine, distinguish between the linear and nonlinear surround properties. We will test the hypothesis that feedback onto a bipolar cell's terminals produces surround inhibition common to more than one postsynaptic ganglion cell type. Second, we will determine which complex receptive field properties unique to a specific ganglion cell type are mediated by feed-forward inhibition. Third, we will study receptive field properties of specific amacrine cell types to determine whether they can convey the nonlinear properties observed in ganglion cells. The experiments will focus on 2 well-characterized concentric ganglion cells, the brisk-transient (BT) and brisk-sustained (BS) cells, and on 2 well characterized complex ganglion cells, the On-Off direction-selective cells (DSGC), and local-edge-detectors (LED), as well as several types of narrow- and wide-field amacrine cell. This work will collect information about retinal structure and function vital to a better understanding of information processing in the visual system and the brain. It will help to understand better how the eye functions, which will help clinical researchers determine what has gone wrong in many types of eye disease and bioengineers in developing prosthetic retinal devices that more closely match the function of the living retina.
Blindness affects millions of Americans and constitutes a significant cost to public and private health sectors. Development of prosthetic devices that can replace the function of the retina is an avenue of treatment that is being actively pursued. This project will elucidate the properties of neural signals in normal retina, which will allow development of prosthetic devices that can better mimic normal retinal function.
Showing the most recent 10 out of 24 publications
