ABSTRACT
We tested the hypothesis that dorsal horn laminae III-IV cell receptive fields (RFs) are initially established in three steps: cutaneous axons penetrate the dorsal horn near their rostrocaudal (RC) levels of entry into the spinal cord. Their terminal branches distribute mediolaterally (ML) according to their relative distoproximal RF locations on the leg, and form nonselective synapses with nearby dorsal horn cell dendrites, establishing the initial dorsal horn cell RFs. Rootlet axon RFs in adult cats were used to approximate the RC entry levels of hindlimb skin input. Cord dorsum recordings of monosynaptic field potentials evoked by electrical skin stimulation provided the RC distributions of synaptic input. These were in close agreement. Simulated projections of all 22,000 hindlimb axons were similar to projections predicted from EPSP distributions, and with the observed projections of dorsal roots, cutaneous nerves, and individual axons. The simulated terminals were connected nonselectively to nearby dendrites of 135,000 simulated lamina III-IV cells whose dendritic surface area distributions were based on intracellularly stained cells. There was an overall similarity among pre- and postsynaptic embryonic and adult somatotopies, with a progressive transformation of RF angular location as a function of RC, ML dorsal horn location from an initial embryonic presynaptic concentric pattern to an adult postsynaptic radial one. The initial embryonic dorsal horn cell RF assembly hypothesis was supported by the simulations, as was the additional hypothesis that further refinement of connections would be necessary to establish sufficient selectivity to account for observed adult RFs and somatotopy.
