Emergence and refinement of clustered horizontal connections in cat striate cortex.

Pyramidal cells in layer 2/3 of adult cat striate cortex have long, intrinsic horizontal axon collaterals within both layer 2/3 and layer 5. These collaterals form periodic "clusters" of finer axon branches that link columns of similar orientation selectivity. We have investigated the sequence of events and possible mechanisms underlying the development of these clustered intrinsic horizontal connections using a combination of neuronal tracers and intracellular staining. Small injections of fluorescent latex microspheres made during the first postnatal week (at P4-6), when examined in tangential sections, produced an even, unclustered distribution of retrogradely labeled cells up to 2 mm from the injection site. At P8, retrograde labeling extended over a larger area and clustering was discernible, primarily among the most distant labeled cells. At both P6 and P8, labeling was similar in layers 2/3 and 5, indicating that the transition from clustered to unclustered connections occurred simultaneously for cells in superficial and deep laminae. By the end of the second postnatal week (P12-15), retrogradely labeled cells were far more clustered both within and beyond the extent of P6 label; the density of labeled cells was high throughout the labeled region, but much higher within clusters. The periodicity of these nascent clusters was similar to that in the adult. Despite obvious clustering, the pattern of retrograde label observed following injections at 2-3 weeks (P12-21) differed markedly from the adult, in that the regions between clusters contained many labeled cells. Over the next 3 weeks, the connections were refined, so that by the sixth postnatal week (P36-38), regions between clusters contained very few retrogradely labeled cells and the overall pattern of retrograde label was indistinguishable from that in adults. Despite differences in postmigratory ages of neurons from the superficial and deep laminae, clustering of retrogradely labeled cells from these 2 populations was similar at all ages. Experiments in which 2-3 weeks elapsed between the time microsphere injections were made and animals were killed demonstrated that neither the initial formation of crude clusters nor their refinement was due to cell death. Instead, cluster refinement resulted from specific process elimination. When a red microsphere injection at P15 was followed by a green microsphere injection at exactly the same location on P29, the earlier injection resulted in crude clustering, as expected. Virtually all of the cells double-labeled by the later injection were within the densest clusters of label from the early injection.(ABSTRACT TRUNCATED AT 400 WORDS)