Distribution of synapses in the lateral geniculate nucleus of the cat: differences between laminae A and A1 and between relay cells and interneurons.

Laminae A and A1 of the lateral geniculate nucleus in the cat are generally considered to be a structurally and functionally matched pair of inputs from two eyes, although there are subtle light microscopic and physiological differences. The present study aims to display ultrastructural differences between these two laminae based on electron microscopic observances on the connectivity patterns of their afferents onto two main cell types: relay cells, and interneurons present in this nucleus. In a design of population measurement from randomized sample areas in laminae A and A1 from six brains, all synaptic contacts made by three terminal types of the geniculate nucleus were identified, and a number of relative distribution properties were analyzed. When the A-laminae were considered as a homogeneous structure, the distribution of the three terminal types on geniculate cells was similar to previously reported results, confirming the validity of the sampling strategies used; RLP (retinal) terminals provided one-fifth of all synapses, whereas RD (from cortex and brainstem) and F (inhibitory) types constituted one-half and one-third, respectively. The relay cells alone received a similar composition of afferents. However, interneurons alone received approximately equal amounts of synapses from the three sources. Similar analyses comparing the distributions in lamina A and A1 revealed that RD and F terminals, but not RLP terminals, innervate these two laminae differently; more RD and fewer F terminals were found in lamina A1. This difference was also present in the distribution of terminals on relay cells alone, but not on interneurons. These results suggest that (1) retinal terminals form a significantly larger fraction of the input to interneurons than to relay cells; correspondingly, cortex and brainstem provide a smaller fraction of all inputs to interneurons than to relay cells; and (2) laminae A and A1 are not strictly equivalent projection sites of the two retinae. The results are discussed in relation to the Y-cell subpopulation in lamina A1 that is involved in corticotectal, as well as corticogeniculate circuits, as opposed to Y-cells of lamina A that are involved in only the latter.