Retrograde transport studies have shown that widespread areas of the cerebral cortex project upon the superior colliculus. In order to explore the organization of these extensive projections, the anterograde autoradiographic method has been used to reveal the distribution and pattern of corticotectal projections arising from 25 cortical areas. In the majority of experiments, electrophysiological recording methods were used to characterize the visual representation and cortical area prior to injection of the tracer. Our findings reveal that seventeen of the 25 cortical areas project upon some portion of the superficial layers (stratum zonale, stratum griseum superficiale, and stratum opticum, SO). These cortical regions include areas 17, 18, 19, 20a, 20b, 21a, 21b, posterior suprasylvian area (PS), ventral lateral suprasylvian area (VLS), posteromedial lateral suprasylvian area (PMLS), anteromedial lateral suprasylvian area (AMLS), anterolateral lateral suprasylvian area (ALLS), posterolateral lateral suprasylvian area (PLLS), dorsolateral lateral suprasyvian area (DLS), periauditory cortex, cingulate cortex, and the visual portion of the anterior ectosylvian sulcus. While some of these corticotectal projections target all superficial laminae and sublaminae, others are more discretely organized in their laminar-sublaminar distribution. Only the corticotectal projections arising from areas 17 and 18 are exclusively related to the superficial layers. The remaining 15 pathways innervate both the superficial and intermediate and/or deep layers. The large intermediate gray layer (stratum griseum intermedium; SGI) receives projections from almost every cortical area; only areas 17 and 18 do not project ventral to SO. All corticotectal projections to SGI vary in their sublaminar distribution and in their specific pattern of termination. The majority of these projections are periodic, or patchy, and there are elaborate (double tier, bridges, or streamers) modes of distribution. We have attempted to place these findings into a conceptual framework that emphasizes that the SGI consists of sensory and motor domains, both of which contain a mosaic of connectionally distinct afferent compartments (Illing and Graybiel, '85, Neuroscience 14:455-482; Harting and Van Lieshout, '91, J. Comp. Neurol. 305:543-558). Corticotectal projections to the layers ventral to SGI, (stratum album intermediale, stratum griseum profundum, and stratum album profundum) arise from thirteen cortical areas. While an organizational plan of these deeper projections is not readily apparent, the distribution of several corticotectal inputs reveals some connectional parcellation.

Anterograde and retrograde transport methods have been used to analyze the projection of the superior colliculus upon the dorsal lateral geniculate nucleus in 19 mammalian species. Our retrograde findings reveal that tectogeniculate neurons are relatively small, and lie dorsally within the superficial gray. These small tectogeniculate neurons are spatially related to a dense tier of W-cell retinal input. Our anterograde tracing results show that tectogeniculate axons are visuotopically distributed to small-celled regions of the lateral geniculate in all nineteen species. In the majority of these species, the small-celled, tectally innervated regions of the lateral geniculate lie adjacent to the optic tract and contain W-cell-like neurons. Our findings suggest that neuroanatomical demonstration of the tectogeniculate projection is a relatively simple and straightforward way of revealing regions of the lateral geniculate which contain W-cells. This is true even in species in which the lateral geniculate lacks obvious cellular laminae, and in regions of the lateral geniculate where W-cells are few in number. The present data are especially interesting in light of the cortical projections of tectally innervated, small-celled regions of the lateral geniculate to the patches or puffs within layer III of area 17. Since these regions of small-celled geniculocortical axons are co-extensive with zones ("blobs") rich in cytochrome oxidase, it might be that information carried over the tectogeniculate circuitry plays an important role in the functions of the blob system.

Axonal markers injected into layers 5 and 6 of cortical areas 17, 18, or 19 labeled axons going to the lateral geniculate nucleus (LGN), the lateral part of the lateralis posterior nucleus (LPl), and pulvinar (P). Area 19 sends fine axons (type 1, Guillery [1966] J Comp Neurol 128:21-50) to LGN, LPl, and P, and thicker, type 2 axons to LPl and P. Areas 17 and 18 send type 1 axons to LGN, and a few type 1, but mainly type 2 axons to LPl and P. Type 1 and 2 axons from a single small cortical locus distribute to distinct, generally nonoverlapping parts of LP and P; type 1 axons have a broader distribution than type 2 axons. Type 2 axons, putative drivers of thalamic relay cells (Sherman and Guillery [1998] Proc Natl Acad Sci USA 95:7121-7126; Sherman and Guillery [2001] Exploring the thalamus. San Diego: Academic Press), supply small terminal arbors (100- to 200-microm diameter) in LPl and P, and then continue into the midbrain. Each thalamic type 2 arbor contains two terminal types. One, at the center of the arbor, is complex and multilobulated; the other, with a more peripheral distribution, is simpler and may contribute to adjacent arbors. Type 2 arbors from a single injection are scattered around and along "isocortical columns" in LPl, (i.e., columns that represent cells having connections to a common cortical locus). Evidence is presented that the connections and consequently the functional properties of cells in LP change along these isocortical columns. Type 2 driver afferents from a single cortical locus can, thus, be seen as representing functionally distinct, parallel pathways from cortex to thalamus.

Anterograde and retrograde transport methods have been used to explore the interconnections between the thalamic reticular nucleus (TRN) and the dorsal lateral geniculate nucleus of Galago crassicaudatus. We first defined the region of the TRN, which is connected to the lateral geniculate nucleus, by examining the distribution of geniculo-TRN axons, cortico-TRN axons arising from area 17, and the location of TRN-geniculate neurons. Following an intraocular injection of 3H-proline/3 H-leucine, trans-synaptically transported protein is present bilaterally within the lateral portion of the caudal TRN. This same caudal and lateral region is also targeted by cortico-TRN axons and contains neurons which project upon the lateral geniculate nucleus. Light microscopic anterograde transport methods were used to analyze the distribution of TRN-geniculate axons. Our data reveal that all layers and interlaminar zones of the dorsal lateral geniculate nucleus contain TRN axons. Electron microscopic-autoradiographic data support and extend our light microscopic findings by revealing labeled TRN terminals within all geniculate layers. These TRN profiles are the same size throughout the geniculate and exhibit morphological characteristics similar to F1 terminals described by others. That is, they possess predominantly pleomorphic vesicles, a dark cytoplasmic matrix, dark mitochondria, and symmetrical synaptic contacts. Two additional features of TRN terminals have been observed in some profiles. These include dense-core vesicles and a dense, punctate cytoplasmic matrix, which is sometimes associated with the postsynaptic specialization. In addition to their morphology and size, the postsynaptic targets of TRN terminals are similar within the three sets (parvi-, magno-, and koniocellular) of geniculate layers. TRN profiles terminate upon dendrites of all sizes and somata. These findings suggest that the TRN modulates the retino-geniculocortical pathway and that this modulation is occurring in all three streams.

We have used retrograde and anterograde transport methods to analyze the nigrotectal projection in the cat. This projection arises from both pars reticulata (SNr) and pars lateralis (SNl) and distributes to all cellular laminae of the superior colliculus. This extensive nigrotectal innervation is not a simple, single circuit. Rather it appears to consist of several parallel channels, with each taking origin from a particular zone of the substantia nigra and terminating within specific collicular laminae and/or sublaminae. For instance, only neurons within the SNl project to the stratum griseum superficiale; such neurons also project diffusely to all other tectal laminae. Cells in the most lateral portion of the SNr project to a horizontal, patchy tier in the interface region between the stratum opticum and the stratum griseum intermediate (SGI). Finally, more medially placed neurons within the SNr project to a horizontal patchy tier within the middle of the SGI and to a wedge-shaped locus in the stratum griseum profundum. Our findings provide an anatomical substrate for electrophysiological data (Karabelas and Moschovakis: J. Comp. Neurol. 239: 309-329, '85) showing a widespread distribution of nigrorecipient tectal neurons in the cat.

We have utilized two different anterograde transport methods (Phaseolus vulgaris leucoagglutinin [PHA-L] immunocytochemistry and autoradiography) in the same experiment to compare the sublaminar location and arrangement of tectopetal axons arising from the substantia nigra pars reticulata, the spinal trigeminal nucleus, and the pedunculopontine tegmental nucleus. Our findings reveal that the nigrotectal projection terminates in a patchy fashion within three horizontally oriented sublaminae of the stratum griseum superficiale (SGI), the dorsal, middle and ventral. The middle tier of nigrotectal axons exhibits an exquisite, puzzle-like, complementary spatial relationship with trigeminotectal axons. In contrast, axons arising from the pedunculopontine tegmental nucleus overlap with patches of nigrotectal axons within the middle tier. Thus the middle tier of the SGI consists of domains of overlapping nigral and pedunculopontine tegmental inputs which interdigitate with domains rich in somatosensory inputs.

The visual-recipient sector of the cat striatum receives corticostriate input from over 15 higher visual and oculomotor-related areas of the cortex and appears homologous with the physiologically characterized region of mixed visual and oculomotor inputs within the primate caudate nucleus. This area in the cat involves the dorsolateral caudate and a strip of the caudal putamen. In a first series of experiments, the former was injected with a retrograde tracer in several cats. Thalamostriate cells were found in extensive regions, including the intralaminar nuclei, certain motor-related nuclei, and, most notably, across much of the extrageniculate visual thalamus. In another set of experiments, anterograde tracers were also injected into the superior colliculus (SC), and labeled tectothalamic fibers were observed in all thalamic sites projecting to the visual-recipient striatum. These findings highlight for the first time the need for the SC to be considered in models of thalamostriate and visual/oculomotor-striatal function(s). Moreover, the data bring to light the fact that basal-ganglia outflow reaching the SC via striatonigro-nigrotectal circuitry is well positioned to modulate ascending tecto-thalamic-thalamostriatal signals destined for the visual-recipient striatum.

Anterograde and retrograde tracing methods have been used to analyze the origin and distribution of parabigeminogeniculate axons in the gray squirrel, the gopher, the rat, the opossum, the cat, the greater bushbaby, the squirrel monkey and the macaque monkey. Our findings reveal that parabigeminogeniculate axons most heavily innervate regions of the lateral geniculate that are also targeted by axons arising from the superior colliculus (tectogeniculate). These geniculate layers and zones of parabigeminal and tectal overlap contain small cells, and in several species are associated with the small W cell retino-geniculocortical pathway. In addition to the dense input to small-celled layers and zones, parabigeminal axons in several species also innervate regions of the lateral geniculate nucleus that are relatively free of tectogeniculate axons and that are associated with the medium (X) and large (Y) cell streams. Finally, our data reveal that the laterality of parabigeminogeniculate pathways varies across mammals, being primarily crossed in the gray squirrel, the gopher, the rat, and the opossum, bilateral in the cat, and primarily ipsilateral in the three primates.