A combination of neuroanatomical and electrophysiological techniques was used to study the effects of peripheral axotomy and regeneration of primary afferents on their central projections in the spinal cord. Individual regenerated afferent fibers were impaled with HRP-filled electrodes in the dorsal columns of alpha-chloralose-anesthetized cats and activated by current pulses delivered via the intracellular electrode. The resulting cord dorsum potentials (CDPs) were recorded at four rostrocaudal locations and HRP was iontophoretically injected into the fiber. Central distributions of boutons and CDPs were compared with peripheral receptor type to determine the accuracy of peripheral regeneration and the effects of central-peripheral mismatches. Reconstruction of the central projections of 13 individual afferents for which the adequate stimulus and CDPs had been recorded revealed many abnormalities. For example, unlike controls, four group I and II afferents with central projections typical of proprioceptors (concentrated in laminae V, VI, and VII) innervating either cutaneous or noncutaneous targets evoked measurable CDPs. Three other group II or A beta afferents innervating low-threshold mechanoreceptors with central terminations confined to the dorsal horn exhibited extensive collateralization in laminae I and II in addition to large numbers of terminals in laminae III-IV. These fibers activated central networks whose adaptation behavior was identical to those evoked by high-threshold mechanoreceptive afferents in controls. These results suggest that primary afferents and their central connections are capable of significant modifications following axotomy and regeneration. In addition, the anatomical studies indicate some reorganization in the laminar distribution of boutons as well as in bouton size.

Horseradish peroxidase injection of identified low threshold cutaneous mechanoreceptor (LTCM) primary afferent axons was used to assess the somatotopic organization of hindlimb projections to laminae III and IV of cat dorsal horn. Multiple injections in the same animals were used to assess bilateral symmetry and precision. Thirty-one axons were injected, with more than 1 axon injected in each of 8 animals (25 axons). Somatotopic relations between their receptive field (RF) centers and the centers of their dorsal horn projections were similar to the somatotopic relations between dorsal horn cell RF centers and cell locations. Very few reversals of mediolateral somatotopic gradients (proximodistal RF location as a function of mediolateral projection center) were observed. Two afferents with nearly identical RFs in 1 animal had nearly identical projections. These observations held for many different combinations of receptor types. A simple mathematical model was used to demonstrate that assembly of dorsal horn cell RFs via passive sampling of the presynaptic neuropil by dorsal horn cell dendrites cannot account for the sizes of dorsal horn cell LTCM RFs. Hypothesized mechanisms for assembly of dorsal horn cell RFs must take into account the functional selectivity of connections required to produce RFs smaller than those predicted by the passive assembly model.

1. To test the hypothesis that subtotal deafferentation of dorsal horn cells can stimulate plastic changes in their receptive fields (RFs), diffuse deafferentation of the cat hindlimb dorsal horn was produced by transection of L7 or L6 and L7 dorsal roots. The following single-unit cutaneous low-threshold mechanoreceptor RF properties were compared between operated and control dorsal horns: 1) distance of RF center from tips of toes, 2) RF length-width ratio; and 3) RF area. 2. In both L7 and L6-L7 rhizotomized animals there was an increased incidence of silent electrode tracks in the most deafferented portion of the hindlimb map (the foot and toe representation). In the rhizotomized L6-L7 animals, there was also an increased incidence of symmetrically placed tracks in deafferented and control dorsal horns, in which cell RFs had no mirror-symmetrical components. In addition, cells in the lateral half of the L6 and L7 dorsal horns exhibited a proximal shift in the location of their RFs. In the rhizotomized L7 animals there was a distal shift of RFs in the L5 segment at long survival times. RFs had lower length-width ratios in L5 and L6 at short survival times and in L6 at long survival times. 3. In intact preparations, dorsal horn cells normally respond to inputs via single or small numbers of low-threshold cutaneous mechanoreceptors. Because these rhizotomies do not remove all inputs from any given area of skin, the deafferentations would produce only patchy loss of input from individual receptors. Therefore observed changes cannot be accounted for entirely by loss of afferent input, suggesting that some reorganization of dorsal horn cell RFs occurred. We conclude that the threshold stimulus for plastic change is less than total deafferentation of dorsal horn cells. At least some of the mechanisms underlying these changes may be active in normal animals in the maintenance of the somatotopic map or in conditioning.

Single dorsal roots of spinal nerves that contribute to the cat lumbosacral plexus (L3-S2) were cut to evoke degeneration of centrally projecting axons. Serial sections throughout lumbosacral cord levels were impregnated by the Fink-Heimer method (20) to permit charting of the distribution patterns of segmental dorsal root afferent fibers. Afferent fibers that enter a single dorsal root have an extensive distribution to multiple cord segments; their longitudinal extent varies with entry level and with laminar targets. Afferent projections to the ventral horn reach motor nuclei only in their entry segment and the adjacent segments just above and below their entry. Those afferent fibers projecting to intermediate gray (laminae VI and VII) have the most extensive spinal distribution of any types; they may, from a single dorsal root, reach as many as 13 or 14 cord segments. Dorsal horn projections of single roots are also longitudinally expansive. Small-diameter afferent fibers course rostrally and caudally in Lissauer's tract (LT) for up to 9-10 segments. They appear to terminate in at least laminae I and II in and near their entry segment; their endings are difficult to demonstrate at greater distances where they are probably less dense. Larger caliber axons entering the dorsal horn generate a somatotopically organized projection, especially to laminae III and IV. Collaterals of these fibers appear to course longitudinally within the gray matter and they distribute to as many as six to seven segments.