Together, these circuit properties endow the retina with complex signal processing capabilities, which have only partially been elucidated and whose characteristics check details remain a central topic of current research in neuroscience. The spike patterns of ganglion cells do not simply represent the level of incident light at a certain spot within the visual field, but rather can encode more complex features of the visual stimulus. Several recent examples have shown that the specific computations underlying the detection and representation of these features can be

understood based on how the respective ganglion cells pool visual inputs over space and time. These findings have called renewed attention to the critical role of nonlinearities in retinal signal integration (Gollisch and Meister, 2010, da Silveira and Roska, 2011 and Schwartz and Rieke, 2011). Although it has long been known that nonlinear integration exists in the retina and that ganglion cells can distinctly

differ in whether they act linearly or nonlinearly (Enroth-Cugell and Robson, 1966), there are only few examples Talazoparib molecular weight of quantitative assessments of the relevant nonlinearities. This calls for new efforts and approaches to take nonlinear signal integration explicitly into account in both experimental and modeling studies. Here, we discuss some emergent ideas regarding the computational roles, the functional forms, and the experimental assessment of nonlinearities in the receptive fields of retinal ganglion cells. Ganglion cells receive their excitatory input from bipolar cells, which in turn are driven by photoreceptors.

This structure leads to a high degree of signal convergence onto single ganglion cells (Hartline, 1940b and Barlow, 1953), leading to the pooling of signals from more than a hundred bipolar cells by some ganglion cells (Freed and Sterling, 1988). Bipolar cells of the same type are organized in fairly regular spatial patterns (Lin and Masland, 2005 and Wässle et al., 2009), and their dendritic first trees – and correspondingly their receptive fields – are typically much smaller than that of the postsynaptic ganglion cell. Bipolar cells, in turn, collect inputs in a similar fashion from typically several photoreceptors (Freed et al., 1987 and Tsukamoto et al., 2001). This stage therefore provides another important site of stimulus integration. Both sites of spatial signal integration – from photoreceptors to bipolar cells and from bipolar cells to ganglion cells – are modulated by inhibitory interactions, mediated by horizontal cells and amacrine cells, respectively. These add lateral interactions over space and thereby directly influence spatial integration. But they also act locally by modulating or antagonizing the feed-forward excitation of individual bipolar cells and thereby influence which local signals are integrated by ganglion cells.