Selective cortical neuronal damage after middle cerebral artery occlusion in rats

We studied histopathologic changes in cerebral cortex of 20 rats after middle cerebral artery occlusion by using the Fink-Heimer suppressive silver impregnation method and conventional stains. At 6 hours after occlusion, Fink-Heimer-stained sections revealed abundant coarsely granular, intensely argyrophilic neurons in the ischemic cortex. These distinctive argyrophilic neurons could be clearly differentiated from neurons that suffered postmortem changes; argyrophilic neurons were present in all layers of the lateral parietal cortex but in only the superficial cortical layers II and III in the parasagittal area of the frontoparietal cortex and the temporo-occipital area. At 24 hours after occlusion as the ischemic region progressed to pannecrosis, argyrophilic neurons were still evident in peri-infarct regions, with more prominent neuritic silver deposits but no changes in number or spatial distribution. Over 2-7 days, the argyrophilic neurons gradually disappeared while many fine silver-impregnated degenerating terminals appeared in the peri-infarct regions. At 3-6 weeks after occlusion, no more argyrophilic neurons were seen in the cortex although degenerating axons were still present in the deep white matter. Our results indicate selective neuronal damage in the superficial cortical layers and massive axonal degeneration in the cerebrum surrounding infarcts. The neuronal damage does not appear to progress beyond 6 hours after middle cerebral artery occlusion. The Fink-Heimer method has many advantages over existing conventional stains for documenting selective neuronal damage in focal cerebral ischemia.     

Neuronal network disturbance after focal ischemia in rats

We studied functional disturbances following left middle cerebral artery occlusion in rats. Neuronal function was evaluated by [14C]2-deoxyglucose autoradiography 1 day after occlusion. We analyzed the mechanisms of change in glucose utilization outside the infarct using Fink-Heimer silver impregnation, axonal transport of wheat germ agglutinin-conjugated-horseradish peroxidase, and succinate dehydrogenase histochemistry. One day after occlusion, glucose utilization was remarkably reduced in the areas surrounding the infarct. There were many silver grains indicating degeneration of the synaptic terminals in the cortical areas surrounding the infarct and the ipsilateral cingulate cortex. Moreover, in the left thalamus where the left middle cerebral artery supplied no blood, glucose utilization significantly decreased compared with sham-operated rats. In the left thalamus, massive silver staining of degenerated synaptic terminals and decreases in succinate dehydrogenase activity were observed 4 and 5 days after occlusion. The absence of succinate dehydrogenase staining may reflect early changes in retrograde degeneration of thalamic neurons after ischemic injury of the thalamocortical pathway. Terminal degeneration even affected areas remote from the infarct: there were silver grains in the contralateral hemisphere transcallosally connected to the infarct and in the ipsilateral substantia nigra. Axonal transport study showed disruption of the corticospinal tract by subcortical ischemia; the transcallosal pathways in the cortex surrounding the infarct were preserved. The relation between neural function and the neuronal network in the area surrounding the focal cerebral infarct is discussed with regard to ischemic penumbra and diaschisis.     

Early and specific expression of monocyte chemoattractant protein-1 in the thalamus induced by cortical injury

For many years it has been known that retrograde degeneration of thalamic neurons occurs following damage to the cerebral cortex, however, the molecular mechanisms which control this process are unknown. Recent studies have demonstrated microglial activation in thalamic nuclei well before the onset of retrograde neuronal cell death. Activated monocytes and microglia synthesize factors detrimental to neuronal survival as well as phagocytose damaged and dying neurons. Our previous studies demonstrated that monocyte chemoattractant protein-1 (MCP-1), a β chemokine which attracts cells of monocytic origin to sites of injury, is rapidly expressed in the brain following visual cortical lesions. The present study examined the expression of MCP-1 messenger RNA and protein in the thalamus following a visual cortical lesion. Aspiration lesions of visual cortex were made in adult mice. At specific times after lesion, brains were harvested and dissected into specific regions. MCP-1 message as detected using northern analysis was absent in uninjured brain, but was elevated in the ipsilateral thalamus as rapidly as 1 h following the lesion. In situ hybridization localized MCP-1 message to subpial glial cells of the lateral geniculate nucleus (LGN) of the ipsilateral thalamus after injury. ELISA showed that MCP-1 protein levels were significantly elevated in the ipsilateral thalamus at 6 h, peaked at 12 h, and remained above baseline levels for at least 1 week post lesion. In addition, anti-GFAP staining demonstrated activated astrocytes localized to the ipsilateral LGN at 24 and 72 h after injury. The early expression and regional localization of MCP-1 mRNA and protein strongly suggest that MCP-1 is a critical molecule in the regulation of thalamic retrograde neuronal degeneration.     

Ischemic axonal injury and its recovery after focal cerebral ischemia

After focal cerebral infarction by occluding the middle cerebral artery (MCA) of the rat, the neuronal death occurred in the ipsilateral thalamic neurons, because axons of the thalamic neurons were injured by infarction and retrograde degeneration occurred in the thalamic neurons. However, cortical neurons adjacent to the infarction survived despite their axons injured by ischemia. We employed immunohistochemical staining for 200 kilodalton (kD) neurofilament (NF), in order to study those responses of cortical and thalamic neurons against axonal injury caused by focal cerebral infarction. In the sham operated rats the immunoreactivity to the anti-200 kD NF antibody was only detected in the axon but not in the cell bodies and dendrites. At 3 days after MCA occlusion, axonal swelling proximal to the site of ischemic injury was found in the caudoputamen and internal capsule of the ipsilateral side. At 7 days after occlusion, cell bodies and dendrites of the neurons in the ipsilateral cortex and thalamus were strongly stained with anti-NF antibodies. At 2 weeks after occlusion these responses disappeared in the cortex, but lasted in the thalamus. These phenomena are caused by stasis of the slow axonal transport, because the NF is transported by slow axonal transport. In the cortical neurons impairment of slow axonal transport recovered in the early phase after injury, but in the thalamic neurons the impairment prolonged up to 3 weeks after occlusion. The early recovery of axonal transport from ischemia seemed to be essential for survival of neurons after ischemic axonal injury.     

Functional changes in thalamic relay neurons after focal cerebral infarct: a study of unit recordings from VPL neurons after MCA occlusion in rats.

We evaluated neuronal and histological changes of thalamic neurons 1, 4, 7, and 14 days after middle cerebral artery (MCA) occlusion in rats. After the somatosensory evoked potentials (SEPs) were measured from the cerebral cortex, the thalamic relay neuronal activities were recorded with a glass microelectrode following repetitive electrical stimulation of the contralateral forepaw at frequencies ranging from 1 to 50 Hz. In approximately 95% of the occluded rats, the ipsilateral somatosensory cortex and/or the subcortical somatosensory pathway developed infarct, resulting in SEP loss. We evaluated unit data from rats with abolished SEPs. The average firing rate of the nucleus ventralis posterolateralis (VPL) neurons in response to 25 stimulations at 30 Hz was significantly reduced to 0.1 spike/stimulus 1 day after MCA occlusion. In sham-operated rats, the same stimulation produced 0.7 spike/stimulus. The firing rate recovered to 0.4 spike/stimulus at 30-Hz stimulation 4 and 7 days after occlusion. This was followed by resuppression (0.1 spike/stimulus) 14 days after occlusion. Histological study revealed some abnormal neurons in the ipsilateral thalamus 7 days after occlusion. We were unable to find normal-shaped neurons in the VPL 14 days after occlusion. The present study demonstrates that cortical infarct produces functional and morphologic changes that gradually and progressively affect the ipsilateral thalamus, although incomplete transient recovery of somatosensory transmission may occur.     

Efferent connections of the orbitofrontal cortex in the marmoset(Saguinus oedipus)

Unilateral partial ablations were made in the orbitofrontal cortex of 4 adult marmosets(Saguinus oedipus) and fiber degeneration was traced using the Nauta-Gygax and Fink-Heimer selective silver impregnation techniques. Corticocortical projections were found to the ipsilateral convexity and medial aspect of the frontal lobe and to the homologous orbitofrontal areas of the contralateral hemisphere. Fiber degeneration was followed through the uncinate fascicle to the temporal and insular cortices, and caudally into the rostrolateral entorhinal cortex. Other fibers joined the cingulum bundle and terminated throughout the cingulate cortex.Subcortical projections were observed to the lateral and basal amygdaloid nuclei, caudate head, ventrolateral putamen and ventral claustrum. The lateral preoptic and hypothalamic areas received a small number of fibers, as did the intralaminar and reticular thalamic nuclei. The dorsomedial nucleus of the thalamus was recipient of a large group of fibers which followed the ventral internal capsule and joined the inferior thalamic peduncle to terminate there. Preterminal debris appeared heaviest in the dorsomedial thalamic nucleus, pars magnocellularis (MDmc) in more caudal orbital lesions. A subthalamic projection to field H of Forel was observed. A small number of fibers terminated in the lateral midbrain tegmentum, but no appreciable fiber degeneration was observed more caudally than the midbrain. These results are compared in some areas to findings in the rhesus monkey. The possibility of a topical organization in the orbital cortical and thalamic projections is discussed.     

Projections of thalamic and cortical gustatory areas in the rat

The method of Nauta was used to stain axons which degenerated after lesions of the thalamic projection nucleus and cortical zone subserving gustation in the rat. Lesions of the medial subnucleus of the thalamic ventral nuclear complex resulted principally in degeneration which passed rostrolaterally through the thalamus, internal capsule, globus pallidus, and striatum to terminate in insular and inferior parietal cortex around the region of the middle cerebral artery. Lesions of this cortical projection zone resulted in degeneration which passed caudomedially through the same striatal and pallidal areas traversed by the thalamocortical fibers. One component of these fibers continued into the thalamus to terminate just dorsal to the medial subnucleus. A second component turned caudally into the cerebral peduncle to terminate in the substantia nigra. There was evidence of terminal degeneration in the striatum but not in the pallidum after both thalamic and cortical lesions. There was no evidence of degeneration in any structures primarily associated with olfactory or alimentary functions.     

A population of nicotinic receptors is associated with thalamocortical afferents in the adult rat: Laminal and areal analysis

In the adult rat brain, a prominent population of nicotinic cholinoceptors binds ³H-nicotine with nanomolar affinity. These receptors are abundant in most thalamic nuclei and in neocortical layers 3/4, which receive a major thalamic input. To test whether cortical nicotinic receptors are associated with thalamocortical afferents, unilateral excitotoxic (N-methyl-D-aspartate) lesions were made in one of four thalamic nuclear groups (anterior, ventral, medial geniculate, or dorsal lateral geniculate) or in temporal cortex. After 1 or 4 weeks of survival, cortical ³H-nicotine binding was quantified via autoradiography. Thalamic lesions resulted in a partial loss of ³H-nicotine binding in ipsilateral cerebral cortex. In each thalamic lesion group, the greatest decrease (35–45%) occurred within the cortical layers and area (i.e., cingulate, parietal, temporal, or occipital cortex) receiving the densest thalamocortical innervation. Binding of ³H-nicotine was also reduced within the thalamus local to the lesion, particularly at the longer survival time. Saturation analysis, performed in frontoparietal cortical tissue homogenates following ventral thalamic lesions, revealed a significant (34%) reduction in receptor density but not affinity. Direct excitotoxic lesions of the neocortex (temporal cortex) tended to preserve ³H-nicotine binding in layers 3/4, despite local neuronal loss. These results, taken with other published findings, suggest that some nicotinic cholinoceptors in adult rat cerebral cortex are located on thalamocortical terminals. This organizing principle appears to apply not only to sensory and motor relay projections but also to association nuclei that project to allocortical areas. These receptors may provide a local mechanism for nicotinic cholinergic modulation of thalamocortical input. J. Comp. Neurol. 380:175–190, 1997. © 1997 Wiley-Liss, Inc.     

Decreases of cortical and thalamic glucose metabolism produced by parietal cortex stimulation in the rat

Parietal cortex stimulation elicited focal decreases as well as increases of brain glucose metabolism in ipsilateral cortex, ipsilateral thalamus, and contralateral cortex of rats in a pattern resembling 'surround inhibition'. It is proposed that parietal stimulation activated inhibitory circuits which decreased cortical and thalamic glucose metabolism. This decrease of cerebral glucose metabolism is important for interpreting brain glucose metabolic studies particularly when metabolic changes do not correlate with changes of neuronal activity.     

Corticofugal axonal degeneration in rats after middle cerebral artery occlusion

We used the Fink-Heimer method to study degenerating corticofugal axons after unilateral middle cerebral artery occlusion in 14 adult male Long-Evans hooded rats. Axonal degeneration in the pyramidal tracts was prominent at 1-3 weeks, manifesting in well-defined silver-impregnated axonal bundles coursing from the internal capsule to the pyramids and crossing completely to the contralateral spinal cord. In half of eight rats examined at 1-3 weeks, the cortical infarct included the forelimb region of the sensorimotor cortex, and degenerating corticospinal axons could be traced to the lower cervical segments; in rats with involvement of the hindlimb cortical area as well, axonal degeneration extended to the lumbosacral segments. Terminal degeneration products were present in the forebrain, midbrain, and brainstem within 2 days after arterial occlusion; the number of degenerating terminals peaked at 7 days and decreased gradually thereafter up to 6 weeks. Dense terminal degeneration was observed in the trigeminal nuclear complex of all seven rats studied at 2 and 7 days. In these seven rats, five had small cortical infarcts, and silver-impregnated terminals were observed in the lateral reticular formation; in two rats with large cortical lesions, terminal degeneration was prominent in the medial reticular formation as well. We conclude that infarcts produced by middle cerebral artery occlusion cause axonal degeneration in the brainstem and spinal cord. The Fink-Heimer method may be useful for evaluating the rat middle cerebral artery occlusion model.     

Adult rat motor cortex connections to thalamus following neonatal and juvenile frontal cortical lesions: WGA-HRP and amino acid studies

Frontal cortex was removed in 1- and 30-day-old rats. When both groups reached 90 days of age, the forelimb motor/sensory cortex in the unlesioned hemisphere was injected with wheat germ agglutinin-horseradish peroxidase (WGA-HRP) or tritiated leucine. Thalamic neurons were retrogradely labeled only ipsilateral to the WGA-HRP injection site in both neonatally and juvenile-lesioned subjects. Ventrolateral (VL), ventromedial (VM), centromedial (CM), centrolateral (CL), parafascicular (PF), posteromedial (POm), and posterior (PO) thalamic nuclei were labeled. This and the demonstration of only ipsilateral thalamocortical connections at birth helped explain the marked thalamic atrophy which developed ipsilateral to neonatal frontal cortex lesions. Death of thalamic neurons after neonatal removal of their normal cortical target could be due to their failure to sprout into the opposite cortex because that cortex was already innervated by the opposite thalamus at birth. Leucine motor/sensory cortex injections in both neonatally and juvenile-lesioned subjects labeled the ipsilateral VL, VM, CM, CL, PF, POm, and PO thalamic nuclei; contralateral CM, CL, and PF thalamic nuclei; ipsilateral medial, ventral, and lateral pontine nuclei; and parts of the contralateral pontine nuclei. The ipsilateral connections were always more robust than the contralateral connections. The contralateral corticothalamic and corticopontine projections, however, were much more numerous and widespread in neonatally compared to juvenile-lesioned subjects. The greater sparing of some motor functions said to occur in neonatal compared to adult motor cortex-lesioned subjects could be due to the plasticity of corticothalamic, corticopontine, and other corticofugal pathways, but not to the plasticity of thalamocortical pathways.     

Thalamotelencephalic projections in the turtle (Pseudemys scripta)

The telencephalic projections of the turtle thalamus were studied using the Fink-Heimer ('67) technique for staining degenerated axons and their terminals. Large thalamic lesions produced terminal degeneration in the basal telencephalic nuclei, the core of the dorsal ventricular ridge and the outer half of layer I in general cortex. A variety of control lesions confirmed that these projections arise in the thalamus. Circumscribed thalamic lesions revealed first, that there is some degree of spatial organization in the turtle's thalamocortical projection system and second, that at least one sensory relay nucleus, the dorsal lateral geniculate, projects to general cortex. Detailed comparisons of the turtle's thalamotelencephalic projections with those present in two primitive mammalian species, the hedgehog and the opossum, provided a basis for identifying probable homologies in the forebrains of reptiles and mammals.     

Projections of thalamic gustatory and lingual areas in the monkey, Macaca fascicularis

The efferent projections of the parvicellular division of the ventroposteromedial nucleus of the thalamus (VMPpc; thalamic taste area) were traced to cortex in Macaca fascicularis by using tritiated amino acid autoradiography. Labeled fascicles could be traced from VPMpc to two discrete regions of cortex. The primary efferent projection was located on ipsilateral insular-opercular cortex adjacent to the superior limiting sulcus and extended as far rostrally as the posterior lateral orbitofrontal cortex. An additional projection was located within primary somatosensory (SI) cortex subjacent to the anterior subcentral sulcus. Following autoradiographic injections in VPM, the trigeminal somatosensory relay, a dense terminal plexus was labeled on SI cortex of both pre- and postcentral gyri, but not within insular-opercular cortex. The autoradiographic data were verified by injecting each cortical projection area with horseradish peroxidase (HRP) and observing the pattern of retrogradely labeled somata within the thalamus. Injections in the precentral gyrus near the anterior subcentral sulcus retrogradely labeled neurons within VPMpc, whereas injections further caudally near the floor of the central sulcus labeled neurons within VPM. Injections of HRP within opercular, insular, or posterior lateral orbitofrontal cortex retrogradely labeled neurons within VPMpc.     

Recombinant basic fibroblast growth factor spares thalamic neurons from retrograde degeneration after ablation of the somatosensory cortex in rats

Retrograde degeneration of thalamic neurons after cortical ablation has long been recognized. Neuronal loss following axotomy eliminates the possibility of regeneration and might prevent the recovery from axonal injury in patients with brain trauma. We investigated whether CS23, a stable recombinant variant of human basic fibroblast growth factor (bFGF), could protect neurons from retrograde degeneration. Four weeks after ablation of the somatosensory cortex in young female rats, there was extensive neuronal degeneration and loss in the lateral ventro-posterior nucleus (VPL) of the ipsilateral thalamus. When Gelfoam soaked in bFGF(CS23) (1 μg/0.l ml) was applied topically at the time of surgery, this neuronal degeneration in the VPL was markedly reduced and macroscopic atrophy of the lateral and medial ventroposterior nucleus (VPL + VPM) was significantly reduced. In contrast, application of bFGF at three days after surgery failed to prevent retrograde degeneration. These resuts indicate that bFGF can prevent thalamic atrophy after ablation of the somatosensory cortex and that administration of bFGF is only effective in the very early period after brain injury.     

Thalamocortical and corticothalamic connections of multiple functional domains in posterior parietal cortex of galagos

In prosimian galagos, the posterior parietal cortex (PPC) is subdivided into a number of functional domains where long-train intracortical microstimulation evoked different types of complex movements. Here, we placed anatomical tracers in multiple locations of PPC to reveal the origins and targets of thalamic connections of four PPC domains for different types of hindlimb, forelimb, or face movements. Thalamic connections of all four domains included nuclei of the motor thalamus, ventral anterior and ventral lateral nuclei, as well as parts of the sensory thalamus, the anterior pulvinar, posterior and ventral posterior superior nuclei, consistent with the sensorimotor functions of PPC domains. PPC domains also projected to the thalamic reticular nucleus in a somatotopic pattern. Quantitative differences in the distributions of labeled neurons in thalamic nuclei suggested that connectional patterns of these domains differed from each other.     

Projections from the parietal cortex to the brain stem nuclei in the cat, with special reference to the parietal cerebro-cerebellar system

Direct projections from the anterior portions of the parietal cortex of the cat to the brain stem nuclei, especially those sending fibers to the cerebellum, were investigated by the Nauta-Gygax and Fink-Heimer methods. Following unilateral lesions of the anterior portions of the middle suprasylvian and/or lateral gyri, a significant amount of pericellular degeneration was found almost entirely ipsilaterally in the rostral levels of the red nucleus and its vicinities, and in the pontine nuclei. Projection fibers to the pontine nuclei appeared to extend over several longitudinal, columnar zones in the pontine gray. Fibers from the anterior portion of the lateral gyrus were observed mainly in the paramedian and lateral nuclei, and those from the middle suprasylvian gyrus in the ventral, paramedian and lateral nuclei. Degeneration in the nucleus reticularis tegmenti pontis of Bechterew was slight, and found bilaterally with ipsilateral predominance. The significance of the anterior portion of the parietal cortex of the cat as a link of cerebro-cerebellar loops was discussed.     

Thalamo-telencephalic connections: new insights on the cortical organization in reptiles

Tracer injections into the dorsal tier of the lacertilian dorsal thalamus revealed an extensive innervation of the cerebral cortex. The medial cortex, the dorsomedial cortex, and the medial part of the dorsal cortex received a bilateral projection, whereas the lateral part of dorsal cortex and the dorsal part of the lateral cortex received only an ipsilateral thalamic projection. Thalamocortical fibers were found superficially in all cortical regions, but in the dorsal part of the lateral cortex, varicose axons within the cellular layer were also observed. The bilateral thalamocortical projection originates from a cell population located throughout the dorsolateral anterior nucleus, whereas the ipsilateral input originates mainly from a rostral neuronal subpopulation of the nucleus. This feature suggests that the dorsolateral anterior nucleus consists of various parts with different projections. The dorsal subdivision of the lateral cortex displayed hodological and topological (radial glia processes) features of a dorsal pallium derivative. After tracer injections into the dorsal cortex of lizards, we found long descending projections that reached the striatum, the diencephalic basal plate, and the mesencephalic tegmentum, which suggests that it may represent a sensorimotor cortex.     

The contribution of the dorsal lateral geniculate nucleus to the total pattern of thalamic terminations in striate cortex of the Virginia opossum

These studies were carried out to show the manner of projection of the dorsal lateral geniculate nucleus and other thalamic nuclei to striate cortex in the Virginia opossum. In order to demonstrate these projections, lesions were made in the dorsal lateral geniculate nucleus, in the ventral lateral geniculate nucleus, in most of the thalamus on one side except for the dorsal lateral geniculate nucleus, and in the entire unilateral thalamus. Following various survival times, usually seven days, the brains were appropriately prepared and stained with procedure I of the Fink-Heimer technique. Dorsal lateral geniculate neurons project in a topographical manner only to certain layers of striate cortex. These projections from the lateral geniculate are compared with the same system in other mammals, and it is concluded that it is similar in all mammals studied, except for the cat. In the cat the lateral geniculate projects beyond the border of striate cortex, but even in the cat the layers of termination within striate cortex are apparently similar. The ventral lateral geniculate nucleus does not project to visual cortex. Dorsal thalamic nuclei other dian the lateral geniculate project to peristriate cortex and to layers VI and I of striate cortex. The finding that thalamic nuclei, other than the lateral geniculate nucleus, project to striate cortex has never been described as part of the visual pathways in other mammals. It is suggested that these additional projections arise mainly from the lateral nuclear group of the thalamus in the opossum, and must be considered in relation to any response characteristics and organization of striate cells determined from physiological studies. These multiple thalamic projections can provide the substrate for more than one representation or “map” of sensory information in striate cortex.     

Connections of the parietal lobe

The efferent connections of the posterior parietal cortex were studied in rhesus monkeys subjected to selective lesions of the superior and inferior parietal lobules, which correspond approximately to Brodmann's areas 5 and 7, respectively.

Following ablations of either the superior or inferior parietal lobule, axon degeneration, stained with the Nauta and Fink-Heimer methods, was traced into the extreme, external, and internal capsules, and into the cerebral peduncle. This degeneration extended into the ipsilateral insular cortex, cingulate gyrus, prefrontal and premotor cortices, and the precentral and postcentral gyri. In addition to these connections, the superior lobule sends fibers to the ipsilateral inferior parietal lobule and superior temporal gyrus, and via the corpus callosum to the contralateral superior and inferior parietal lobules, whereas the inferior parietal lobule sends fibers to the ipsilateral superior parietal lobule and to the contralateral superior and inferior parietal lobules. A prominent fiber system to the ipsilateral temporal lobe degenerates following lesions in the inferior parietal lobule (area 7); in such cases fiber degeneration appears in the superior, middle and inferior temporal convolutions, and in the fusiform and parahippocampal gyri.

Both lobules evidently project to the claustrum and body of the caudate nucleus. Both, moreover, have massive efferent connections with the dorsal two-thirds of the putamen. By contrast, no evidence of projections from the parietal cortex to the globus pallidus was found in any of the cases studied.

A further subcortical projection from the posterior parietal cortex involves the nucleus reticularis thalami and the nucleus lateralis posterior thalami. The inferior lobule projects directly to the nucleus lateralis dorsalis and to the mediodorsal region of the nucleus lateralis posterior that closely adjoins two thalamic cell groups: the n. lateralis dorsalis and the intralaminar nucleus centralis lateralis. The superior parietal lobule, by contrast, projects massively to a ventrolateral district of the nucleus lateralis posterior.

Parietosubthalamic connections could be traced from areas 5 and 7 to the zona incerta and fields H₂ and H of Forel, but evidence for terminal connections with the n. subthalamicus (Luys) could not be foud.

Both areas 5 and 7 project massively to the pretectal area and the deeper layers of the superior colliculus. This parieto-mesencephalic connection is amplified by a fiber connection from the inferior parietal lobule (area 7) to the lateral, densocellular region of the circumaqueductal gray matter. No evidence of parietal corticonigral fibers connections was found. Finally, both parietal lobules were found to project to the pontine nuclei.

Speculations regarding the associative functions of the parietal lobules at the cortical and subcortical levels are presented, with particular emphasis upon the possible significance of the projections from the inferior parietal lobule to insular, cingulate and temporal regions of the cortex.