, 2009; Kay et al., 2011). Here, we provide functional evidence in support of layer-specific DS-RGC input by directly imaging presynaptic DS Ca2+ signals in the most superficial retinorecipient layers (Figure 6). This is consistent with the recent finding that Ca2+ signals are tuned to tail-to-head (CR) motion in a superficial sublayer of SFGS (Nikolaou et al., 2012), using presynaptic Ca2+ indicators of the SyGCaMP family (Dreosti et al., 2009). Given the tight regulation of laminar specificity

by molecular recognition mechanisms (Huberman et al., 2010; Sanes and Zipursky, 2010), it seems plausible that the genetic expression profile determines both the dendritic MK 2206 wiring pattern in the retinal inner plexiform layer (IPL), which determines the PD (Briggman et al., 2011), and the precise tectal stratum the axon terminals preferentially innervate. Postsynaptic tectal cell types arborize in different layers in the SFGS, which correlates with their molecular profile (Robles et al., 2011; this paper). In such a model of lamina-specific functional specialization, basic DS is not the result of intratectal computation within the local circuitry. Instead, it is the result of spatial separation of different features of the visual scene already analyzed in the retina (Gollisch and Meister, 2010) and conveyed to the tectum selleck chemicals by different signaling channels into different

strata. This model also provides a simple explanation why tectal cells show matching PDs when either the contra- or ipsilateral eye is stimulated in artificially induced binocular tectal circuits (Ramdya and Engert, 2008): if DS-RGCs innervate different tectal sublaminae depending on a molecular recognition mechanism, they are likely to do so independent of which eye they are located in. A tectal neuron will then arborize and receive

input from the tectal lamina(e) it is specified to connect to and therefore receive consistent DS signals from both eyes. The two DS cell classes identified here were often whatever inhibited by stimuli moving in nonpreferred directions. What may be the source of these DS inhibitory inputs? GABAergic SINs branch horizontally in the dorsal neuropil (Del Bene et al., 2010), where they could contact the distal dendrites of type 1 and/or type 2 neurons. Another attractive possibility is that type 1 and type 2 cells inhibit each other reciprocally. This is because (1) their spike output is tuned in opposite directions, (2) they exhibit a GABAergic phenotype, and (3) their lower dendritic/axonal compartments branch in a similar layer at the SGC/SFGS border, where they could form synaptic contacts between each other. In this model of reciprocal inhibition, homotypic inhibitory connections within the class of type 1 and type 2 cells would occur less frequently because inhibitory currents were relatively small during preferred-direction stimuli.