In the early steps of visual perception, the retina encodes visual scenes in the form of action potentials in ganglion cells, which convey information through the optic nerve. Like every neural circuit, the retina faces the problem of encoding a wide range of stimuli using neurons that have a limited dynamic range of firing rates. Many retinal ganglion cells encode the scene efficiently by adapting to predictable spatio-temporal patterns over several seconds, thus becoming more sensitive to novel patterns. For example, in response to a pattern of flickering vertical bars, many cells decrease their sensitivity to vertically oriented stimuli, while decreasing their sensitivity to horizontal stimuli. How the circuit performs this “pattern adaptation” is not known. Inhibitory neurotransmission is required for pattern adaptation, indicating that inhibitory amacrine cells play an important role. One possibility for how amacrine cells could produce pattern adaptation is that individual amacrine cells are selective for particular patterns and change their activity when strongly activated by that pattern. A second possibility is that amacrine cells can selectively adjust their synapses, strengthening those synapses that adapt the circuit to a particular pattern while weakening other synapses from the same cell. This synapse-specific plasticity would give the circuit a much greater flexibility in its adaptive properties. It has been proposed that this synaptic adaptation could be achieved by an “anti-Hebbian” plasticity mechanism, whereby if an amacrine cell�s activity is correlated with that of a postsynaptic ganglion cell, the strength of the inhibition increases. Using the isolated retina of the tiger salamander, I am directly testing these two possibilities by intracellularly recording from individual amacrine cells while simultaneously recording from many ganglion cells using an extracellular multielectrode array. By injecting current into the amacrine cell while projecting visual stimuli onto the retina, I directly measure how amacrine to ganglion cell transmission changes during pattern adaptation. Initial results suggest that amacrine cells only slightly change their sensitivity in response to visual patterns, but they consistently change their transmission to a subset of postsynaptic ganglion cells depending on the pattern. For some amacrine and ganglion cell pairs, patterns that increase the correlation between the two cells also increase the inhibitory transmission, suggesting an anti-Hebbian plasticity mechanism.