Random feedforward projections account for orientation selectivity of the thalamic input in rodent visual cortex

G. MATO; D. HANSEL; C. VAN VREESWIJK

Copenhagen

Conferencia; FENS 2016; 2018

FENS

Neurons in mouse primary visual cortex (V1) exhibit orientation selectivity (OS). However, in contrast to cat and monkey, mouse V1 does not exhibit orientation maps. It has been recently shown that another difference with the later species is that lateral geniculate nucleus (LGN) neurons display significant OS (Zhao et al 2013, Scholl et al, 2013, Sun et al 2015). Whether OS in rodent visual cortex is primarily due to feedforward (FF) connectivity or LGN selectivity is not presently known. To address this issue we consider a computational model comprising two networks of neurons, one representing the LGN and the other Layer 4, connected with completely random feedforward connections. The stimuli are drifting gratings. We show that even in the absence of OS in LGN, the F1 component of the thalamic inputs to Layer 4 has a typical average orientation selectivity index, OSIF1 ~ 0.2. We also find that OSIF1 depends very weakly on the OS in LGN. In contrast, for the selectivity of the F0 component (OSIF0), OS in LGN is essential. However, for the degree of LGN selectivity observed experimentally, our model gives an average OSIF0 ~ 0.05. These values for OSIF0 , OSIF1 are comparable with recent experimental data (Lien et al., 2013). We therefore conclude that in rodent, the OS of the thalamic input can be accounted for by a purely random organization of the LGN projections.