Differential properties of cells in the feline primary visual cortex providing the corticofugal feedback to the lateral geniculate nucleus and visual claustrum

We have examined the responses of 141 layer VI cells in the feline visual codex. Within this group we compared the responses of a subpopulation of cells checked for connectivity by electrical stimulation in the dLGN and the visual claustrum. The antidromically identified corticogeniculate projecting cells had relatively short receptive fields, as judged from length response curves, measured quantitatively, and were located at the 'short' end of the receptive field length spectrum seen in the general population. Of the 17 corticogeniculate projecting cells, 71% were S type cells, which were typically monocular and directionally selective, with relatively long latencies following electrical stimulation. The remaining 29% were C type cells, also directionally selective, but with a wider spread of ocular dominance preferences and shorter latencies following electrical stimulation. S and C type subpopulations did not differ in their receptive field lengths. The mean receptive field length for this subpopulation was 2.2° ± 0.27, the shortest field being 1° and the longest 5°. The five layer VI cells activated by electrical stimulation from electrodes within the dorsocaudal (visual) claustrum all had much longer receptive field lengths than the corticogeniculate population, often 10° or longer and were monocular and directionally selective S type cells. These data indicate that the information carried in the corticogeniculate stream (and that from layer VI directly to layer IV carried by axon collaterals) is relatively tightly focused in spatial terms whilst the less spatially focused, long receptive field output from layer VI projects to the claustrum.