Transmission of Mixed Convergent Signals at the Mouse Retinogeniculate Synapse

There are two broad modes of information transfer in the brain: the labeled line model, where neurons relay inputs they receive, and the mixed tuning model, where neurons transform and integrate different inputs. In the visual pathway, information transfer between retinal ganglion cells (RGCs) and the dorsal lateral geniculate nucleus (dLGN) neurons is primarily viewed as a labeled line. However, recent work in mice has demonstrated that different RGC types, encoding distinct visual features, can converge onto a dLGN neuron, raising the fundamental question of whether the dLGN transforms visual information. Using optogenetics we activated distinct RGC populations and assessed spiking output of dLGN neurons by in vivo recordings. We found that visual response properties of dLGN neurons driven by a specific RGC population largely matched properties of the activated RGCs. Furthermore, in vitro dual-opsin experiments demonstrate that strong functional convergence from distinct RGC types rarely occurs. Thus, retinogeniculate information transfer in mice largely adheres to a labeled line model.