Mediodorsal nucleus: Areal, laminar, and tangential distribution of afferents and efferents in the frontal lobe of rhesus monkeys

Abstract

The terminal distribution of thalamic afferents in primate prefrontal cortex has never been examined in any detail. In the present study, WGA-HRP was injected into major subdivisions of the mediodorsal nucleus (MD) in the rhesus monkey in order to determine (1) The areal distribution of MD projections, (2) the layer (s) in which MD afferents terminate, (3) the tangential pattern of the MD axonal terminals, (4) the cells of origin of the reciprocal corticothalamic pathway, and (5) the degree of reciprocity between the corticothalamic and thalamocortical pathways in the different regions of the prefrontal cortex. As expected on the basis of retrograde degeneration and transport studies, injections centered in the magnocellular (MDmc) subnucleus of MD labeled cells and terminals in the ventral and medial prefrontal cortex. Injections involving ventral MDmc labeled the more lateral of these areas (Walker's areas 11 and 12); injections of the dorsal MDmc labeled the ventromedial regions (areas 13 and 14). In contrast, injections involving mainly the lateral, parvicellular (MDpc) moiety labeled cells and terminals in dorsolateral and dorsomedial areas (Walker's 46, 9, and 8B). Area 8A was labeled most prominently when injections included the multiform portion of MD (MDmf) and area 10 had connections with anterior portions of MD. A dorsal-ventral topography for MDpc exists with dorsal MDpc labeling dorsal and dorsomedial prefrontal areas and ventral MDpc labeling dorsolateral prefrontal cortex. Our findings with respect to MD are consistent with a nucleus-to-field organization of its thalamocortical projection system. Outside of the traditional boundaries of prefrontal cortex, lateral MD projections extended to the supplementary motor area (SMA) and the dorsal part of the anterior cingulate (AC) whereas the medial MD projection targeted the ventromedial cingulate cortex and spared SMA. In addition, a few labeled cells and sparse terminals were found in the inferior parietal lobule, the superior temporal sulcus, and the anterior part of the insula after injections that involved the medial part of MD. Labeled terminals were invariably confined to layer IV and adjacent deep layer III. No terminal label was ever observed in layers I, II, superficial III, V, or VI in any part of the cerebral cortex following injections confined to any part of MD. The projection from MD to the PFC formed a disjunctive pattern of bands ranging from 0.3 to 1.3 mm, which were especially evident when injections were relatively small; when the injection involved a larger thalamic volume, the pattern of terminal labeling was somewhat more diffuse and tangentially extensive throughout layers III and IV. These data may indicate an interdigitation of the terminal fields originating from adjacent clusters of thalamic cells (Goldman-Rakic and Porrino: J. Comp. Neurol. 242:535–560, '85). Analysis of retrograde transport in the PFC, SMA and AC revealed unexpected regional differences in the laminar origin of corticothalamic cells. In all prefrontal areas, labeled corticothalamic neurons were found mainly in the superficial part of layer VI with a few labeled cells situated also in the superficial part of layer V. In the supplementary motor area the distribution of cells was essentially bilaminar with the number of labeled cells in superficial V much more substantial than in PFC. Finally, in the anterior cingulate cortex, the labeled cells were confined to and stratified in layer VI with about equal density in its superficial and deep strata. Regional differences were also apparent in relationships of retrogradely labeled cells and anterogradely labeled terminal fields. In prefrontal areas, fluctuations in the density of labeled cells paralleled changes in terminal field density. The anterior cingulate and supplementary motor cortex, however, contained high concentrations of labeled neurons but only sparse terminal label, indicating that the mediodorsal nucleus may have a different anatomical relationship with its “essential” target than with its “accessory” or “secondary” targets.