The 3H-thymidine method of birth-dating was used to determine when the cells belonging to each of the principal cellular layers of the cat's primary visual cortex are generated. In order to detect systematic differences in the position of radioactively labeled cells following 3H-thymidine administration at different prenatal ages, a geometric method was devised to represent the distribution of labeled cells in the form of depth histograms. Results show that visual cortical neurogenesis occurs largely during the second half of gestation between embryonic day 31 (E31) and E57. Cells of layer 6 are generated early, between E31 and E38, whereas cells destined for successively more superficial layers are generated at progressively later times. Layer 4 cells, the principal targets of geniculocortical afferents, are generated between E37 and E44. In addition, a special population of cells embedded in the white matter below layer 6 was found to be produced throughout the week-long period immediately prior to the onset of layer 6 neurogenesis. Overall, this radial pattern of cortical neurogenesis closely resembles the inside-first, outside-last, spatiotemporal sequence of development described for the monkey's primary visual cortex (Rakic, '74).
In addition to finding this pronounced gradient in the radial dimension, we were also able to detect a less pronounced gradient along the tangential dimension: neurons destined for any given layer in the anterior part of the cortex (inferior visual, field representation) are generated slightly in advance of neurons destined for more posterior regions (superior visual field). However even our more quantitative histogram analysis failed to reveal a mediolateral (central to peripheral visual field) gradient within area 17.
In the cat, layers 6, 5, and 4 each take about a week to be generated, although their total cell numbers and packing densities differ in the adult. About 2 weeks are required to produce the cells of layers 2 and 3 combined. Furthermore, we found that neurons belonging to different layers and different morphological classes can be generated simultaneously. This suggests that the identity of a cortical neuron is not solely a function of the time of neurogenesis.