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

Organization of Excitatory Inputs from the Cerebral Cortex to the Cerebellar Dentate Nucleus

Intracellular recording was made from dentate nucleus neurons (DNNs) in anesthetized cats, to investigate cerebral inputs to DNNs and their responsible pathways. Stimulation of the medial portion of the contralateral pericruciate cortex most effectively produced EPSPs followed by long-lasting IPSPs in DNNs. Stimulation of the pontine nucleus (PN), the nucleus reticularis tegmenti pontis (NRTP) and the inferior olive (IO) produced monosynaptic EPSPs and polysynaptic IPSPs in DNNs. The results indicate that the excitatory input from the cerebral cortex to DNNs is at least partly relayed via the PN, the NRTP and the 10. Intraaxonal injection of HRP visualized the morphology of mossy fibers from the PN to the DN and the cerebellar cortex. The functional significance of the excitatory inputs from the PN and the NRTP to the DN is discussed in relation to the motor control mechanisms of the cerebellum.

Three-dimensional topography of corticopontine projections from rat sensorimotor cortex: comparisons with corticostriatal projections reveal diverse integrative organization

The major cortical-subcortical re-entrant pathways through the basal ganglia and cerebellum are considered to represent anatomically segregated channels for information originating in different cortical areas. A capacity for integrating unique combinations of cortical inputs has been well documented in the basal ganglia circuits but is largely undefined in the precerebellar circuits. To compare and quantify the amount of overlap that occurs in the first link of the cortico-ponto-cerebellar pathway, a dual tracing approach was used to map the spatial relationship between projections originating from the primary somatosensory cortex (SI), the secondary somatosensory cortex (SII), and the primary motor cortex (MI). The anterograde tracers biotinylated dextran amine and Fluoro-Ruby were injected into homologous whisker representations of either SI and SII, or SI and MI. The ensuing pontine labeling patterns were analyzed using a computerized three-dimensional reconstruction approach. The results demonstrate that whisker-related projections from SI and MI are largely segregated. At some locations, the two projections are adjoining and partly overlapping. Furthermore, SI contributes significantly more corticopontine projections than MI. By comparison, projections from corresponding representations in SI and SII terminate in similar parts of the pontine nuclei and display considerable amounts of spatial overlap. Finally, comparison of corticopontine and corticostriatal projections in the same experimental animals reveals that SI-SII overlap is significantly larger in the pontine nuclei than in the neostriatum. These structural differences indicate a larger capacity for integration of information within the same sensory modality in the pontocerebellar system compared to the basal ganglia.

Projections from the Cerebral Cortex to the Red Nucleus of the Guinea-pig. A Retrograde Tracing Study

The origin and the topographic distribution of corticorubral (CR) projections in the guinea-pig were studied by using the retrograde axonal transport of a tracer, colloidal gold-labelled, enzymatically inactive horseradish peroxidase conjugated to wheat-germ agglutinin (WGAapoHRP - Au), which was injected in the red nucleus (RN). It was found that the bulk of the CR projections arise from layer V neurons of the agranular frontal cortex in both its medial (Agm) and lateral (Agl) subdivisions; in the Agm labelled neurons are preferentially located in the upper part of layer V, whereas in the Agl they are more concentrated in the central band of the layer. Fewer projections originate from areas of the granular parietal and the agranular cingulate and retrobulbar cortices. CR projections have a bilateral origin, with a large ipsilateral predominance. The pattern of retrograde cortical labelling observed after injection of WGAapoHRP - Au in different portions of the RN indicates that CR projections are distributed throughout the entire rostrocaudal extent of the nucleus, but are slightly more concentrated in the rostral parvocellular area. The morphological arrangement of CR projections in the guinea-pig, as demonstrated in the present study, shows several analogies with other mammals. The functional characteristics of the cortical areas in which CR neurons are located indicate that CR projections may play a significant role in the central organization of movement.

Salient anatomic features of the cortico-ponto-cerebellar pathway

Recent studies of the primate corticopontine projection show that the neocerebellum--in addition to connections from motor and sensory areas--receives connections from various association areas of the cerebral cortex, some of which are thought to be primarily engaged in cognitive tasks. The quantities of such connections in relation to those from more clearly motor-related parts of the cortex need to be more precisely determined, however. Furthermore, the anatomic data on origin of corticopontine fibers needs to be supplemented with physiological experiments to clarify their functional properties at the single-cell level. For example, nothing is known of the functional role of the large input from the cingulate gyrus, nor is the input from the posterior parietal cortex physiologically characterized. Finally, the scarcity of corticopontine connections from the prefrontal cortex in the monkey (and probably also in man) may not seem readily compatible with a prominent role of the neocerebellum in certain cognitive tasks. We discuss data--in particular from three-dimensional reconstructions--indicating that both corticopontine projects and pontocerebellar neurons are arranged in a lamellar pattern. Corticopontine and pontocerebellar lamellae have similar shapes and orientations but appear to differ in other respects. Corticopontine terminal fields are sharply delimited, apparently without gradual overlap between projections from different sites in the cortex, whereas pontocerebellar lamellae are more fuzzy and exhibit gradual overlap of neuronal populations projecting to different targets. In spite of the sharpness of the corticopontine projection, there may be many opportunities for convergence of inputs from different parts of the cortex. Thus, the wide divergence of corticopontine projections produces many sites of overlap, and extensive interfaces between different terminal fields enabling convergence of inputs onto each neuron. We suggest that the lamellar arrangement of corticopontine terminal fields and of pontocerebellar neurons serve to create diversity of pontocerebellar neuronal properties. Thus, each small part of the cerebellar cortex would receive a specific combination of messages from many different sites in the cerebral cortex. The spatial arrangement of cerebrocerebellar connections have to be understood both in terms of fairly simple large-scale, gradual topographic relationships and an apparently highly complex pattern of divergence and convergence. Developmental studies of corticopontine and of pontocerebellar projections together with three-dimensional reconstructions in adults suggest that the highly complex adult connectional pattern may be created by simple rules operating during development.

Fronto-parietal control of electrodermal activity in the cat

The aim of this work was to investigate the direct involvement of the fronto-parietal cortex in the control of spinal autonomic centers eliciting electrodermal activity (EDA). This autonomic response, linked with the activity of sweat glands, was recorded as skin potential responses (SPRs) from forepaws in the cat. Animals were paralyzed by gallamine and SPRs were obtained under halothane anaesthesia. For each animal, a transection of the medulla sparing only pyramidal tracts was carried out. SPRs were elicited by direct electrical stimulation of pericruciate and posterior parietal cortical areas before and after such a transection. Results showed that in intact preparations, stimulation of the pericruciate cortex evoked SPRs at lower thresholds than the posterior parietal cortex. After the bulbar transection, only the stimulation of pericruciate areas still elicited SPRs at low intensities. Results are interpreted as indicating that fronto-parietal control of EDA is probably mediated by a double descending system: one involving corticoreticulospinal pathways and a direct corticospinal one. We hypothesized that the somatic motor cortex initiates descending programs to autonomic centers at bulbar and spinal levels, and that these centers are involved in autonomic adjustments to somatomotor movements.

Preservation of pointing accuracy toward moving targets after extensive visual cortical ablations in cats

Impairments in reaching toward stationary and moving targets were studied in cats after restricted or extensive removal of visual cortical areas (areas 17, 18 and 19 and lateral suprasylvian visual areas). Regardless of the extent of the cortical lesion, cats were at first unable to localise and reach for a stationary target whereas they were soon able to detect and accurately point toward a mobile one. Moreover, the onset latency of such movements was dramatically increased. During post-operative re-training, the cats were unable to improve their accuracy scores when reaching towards stationary targets. In contrast, full compensation was observed for the accuracy of reaching movements directed toward moving targets. A partial recovery was observed for movement latency values that progressively decreased but left a permanent 30-40 ms impairment following extensive lesions. The role of extrageniculate messages and alternative routes involving other cortical areas in taking in charge the visuomotor activity is discussed.

Cerebro-cerebellar projections from the ventral bank of the anterior ectosylvian sulcus in the cat

1. Stimulation of the ventral bank of the anterior ectosylvian sulcus (AESv) induced marked mossy fibre (MF) and climbing fibre (CF) responses in the cerebellar posterior vermis (lobules VI-VII) and moderate sized ones in the paraflocculus, paramedian lobules and crus I and II of the cat. The relay stations for these responses to the posterior vermis were investigated morphologically and electrophysiologically. 2. It can be considered that the MF responses were relayed at least in part via the dorsolateral, peduncular and paramedian pontine nuclei, since in these nuclei there were units orthodromically responsive to AESv stimulation and antidromically responsive to stimulation of the posterior vermis. The MF responses are thought to be relayed monosynaptically, since the distribution of axon terminals labelled after injection of wheatgerm agglutinin-conjugated peroxidase (WGA-HRP) into the AESv overlapped in these pontine nuclei with that of neurons labelled after injection of WGA-HRP into the posterior vermis. 3. It is thought that the CF responses are relayed in the caudomedial part of the medial accessory olive (MAOcm), because neurons in the MAOcm were orthodromically responsive to AESv stimulation and antidromically responsive to stimulation of the posterior vermis. 4. It is suggested that the cerebro-olivary projection which transmits the orthodromic responses in the MAOcm is indirect, via the superior colliculus (SC), because injection of WGA-HRP into the AESv labelled axon terminals not in the MAOcm but in the SC, and injection of WGA-HRP into the MAOcm gave rise to retrograde labelling of cells in the SC. Synaptic connections between the axon terminals of the cerebrotectal projection and the tecto-olivary neurons were demonstrated by extracellular unit studies in the SC. 5. The hypothesis that the CF responses were transmitted via the SC was supported by the finding that the CF responses disappeared transiently after muscimol or lidocaine was injected into the SC. 6. These findings provide evidence that the MF responses are transmitted at least in part via the cerebro-ponto-cerebellar projection, while the CF responses are relayed via the cerebro-tecto-olivo-cerebellar projection. These cerebro-cerebellar pathways from the AESv are suggested to participate in conducting visual information to the posterior vermis.

Organization of cingulo-ponto-cerebellar connections in the cat

This study deals with three different aspects of the organization of connections from the cingulate gyrus to the cerebellum. (1) With the use of wheat germ agglutinin-horseradish peroxidase as a retrograde tracer, the distribution of cingulate neurons projecting to the pontine nuclei was studied. Retrogradely labeled cells were found in layer 5 in all parts of the cingulate gyrus. Average densities of cingulo-pontine cells were similar in the different cytoarchitectonic subdivisions, although some density gradients were observed. The projection was found to be remarkably strong. Average densities of corticopontine cells in the cingulate gyrus ranged from 500-700 cells per mm2 cortical surface, and the total number of neurons was in the range of 75,000-105,000 (n = 4). (2) A topographical organization of terminal fields of fibers originating in different parts of the cingulate gyrus was demonstrated with the combined use of anterograde degeneration and anterograde transport of wheat germ agglutinin-horseradish peroxidase. Terminal fibers originating in different zones of the cingulate gyrus were distributed in a patchy mosaic within a narrow band along the ventromedial aspect of the pontine nuclei. (3) We confirm, with the combined use of lesions in the cingulate gyrus and injections of wheat germ agglutinin-horseradish peroxidase in the ventral paraflocculus, that there is considerable overlap between terminal fibers originating in the cingulate gyrus, and cells retrogradely labeled from the ventral paraflocculus. The role of the ventral paraflocculus as a receiver of "limbic" input is discussed.

Visual 'cortical-recipient' and 'tectal-recipient' pontine zones play distinct roles in cat visuomotor performance

In order to test the hypothesis that visual information reaching the cerebellum through the pontine nuclei is involved in the control of visually guided movements, the effects of bilateral kainic acid pontine lesions have been analysed in cats performing a reaching movement towards a spot of light that was either stationary or moving. In 4 cats, the lesion was restricted either to the ventromedian region (cortical-recipient zone) or to the dorsolateral nucleus (tectal-recipient zone) of the pons. A major and persistent impairment was seen when the cerebellum was deprived of the pontine information influenced by the colliculus. While cats displayed no impairment when reaching towards a stationary target, they exhibited a strong accuracy deficit associated with an increased reaction time when reaching towards a moving target. In contrast, lesioning the pontine zone influenced by the visual cortex induced a transient accuracy deficit with moving targets and a transient delay in movement onset whatever the mode of target presentation. These results emphasise the involvement of visual pontine regions in the guidance of movements; they also confirm previous results showing that tectal visual information plays a more important role than that originating in the visual cortex when movements are directed towards moving targets.

Cortical influences on neurons of the lateral reticular nucleus responding to noxious stimuli

The effect of frontoparietal sensorimotor (FPSM) cortex stimulation on both the spontaneous and the noxious evoked activity of neurons in the lateral reticular nucleus (LRN) was tested in barbiturate-anesthetized rats. Ninety-three LRN neurons that responded to a noxious heat stimulus (HS) were recorded (72% antidromically fired from the cerebellum). Of these, 66 neurons altered their spontaneous firing rates in response to cortical stimulation. Two patterns of responses were found: either an excitation followed by a suppression of spontaneous activity (52 neurons), or a pure suppression of spontaneous activity lasting 50-400 msec (14 neurons). In 46 of these neurons, it was found that cortical stimulation reduced HS-evoked activity to near the baseline level. Furthermore, it was found that when applied after a prolonged cortical stimulation, the HS was ineffective. It is concluded that FPSM cortex can influence nociceptive information in LRN neurons that respond to its stimulation, possibly interfering with the mechanisms underlying stimulation-produced analgesia (SPA). In this context, it is proposed that the cortex can modulate the activity of LRN neurons that activate, through local loops, a descending antinociceptive system and also a separate projection system to the cerebellum.

Bilateral pericruciate cortical innervation of the red nucleus in cats with adult or neonatal cerebral hemispherectomy

We studied remodeling of the remaining corticorubral projections in adult cats sustaining a left cerebral hemispherectomy in adulthood or neonatally using cortical injections of [3H]leucine-proline. Injection sites and terminal fields were reconstructed from autoradiography-processed tissue. In all cats, the label filled similar extents of ares 4 gamma and 3a of the right frontal cortex. We used sections at 8 coronal planes throughout the red nucleus (RN) for computer-assisted analysis of visually estimated density and topography of distribution of terminal label, and for calculation of RN cross-sectional area. Additionally, at 3 coronal planes we further quantified terminal label using computerized procedures (number of particles for surface area). In all lesioned cats we found terminal label in the RN contralateral to the injection site with a topographic distribution similar to that of the RN ipsilateral to the injection in normal or lesioned cats and in absence of any significant shrinkage of the nucleus. The difference between the 2 age-at-lesion groups was that in the cats with neonatal ablation the density of contralateral terminal label was about double that seen in adult-lesioned subjects. However, the amount of contralateral labeling in adult-lesioned cats was substantial and represented a significant increase over the minimal labeling seen in normal cats. There were no differences between groups in labeling or size of the RN ipsilateral to the injection site. For reasons discussed, we interpret the label on the side of the hemispherectomy as representative of reinnervation of the cortically deafferented RN by crossing collaterals of fibers arising in the remaining motor cortex and not as lesion-sustained persistent prenatal connections.

Functional organization of the parvocellular red nucleus in the cat

The experiments were performed on cats under pentobarbital anesthesia. The following results were obtained. (1) Most of the neurons in the rostral part of the red nucleus (RN) were activated by stimulation of the parietal association cortex (P) and the lateral cerebellar nucleus (CN). A number of the neurons were regarded as rubro-olivary since they responded antidromically to inferior olive (IO) stimulation. (2) Stimulation of the P as well as the frontal motor cortex activated IO neurons. A longer and more variable latency in P-induced responses indicated the existence of an indirect connection from the P to the IO. The recording sites for P-induced responses were located in the rostral region of the IO, giving off projection fibers to the cerebellar hemispheral part. After horseradish peroxidase (HRP) injection through recording microelectrodes into the IO where P-induced responses were recorded, small cells labeled retrogradely with HRP were found distributed in the rostral part of the RN. (3) Effects of stimulus to or lesion of the rostral part of the RN revealed that the climbing fiber responses to P stimulation were conveyed through the rubro-olivary pathway originating in the rostral part of the RN. (4) Following HRP injection through recording electrodes into the rostral part of the RN where P-induced responses were recorded, retrogradely labeled cells were seen and located in the P and CN. From these findings, the rostral part of the RN investigated in this study could be regarded as the rubral parvocellular part. The present study suggests that the latero (dentato-)-rubro-olivo-cerebellar circuit forms an internal feedback loop and the P acts on this loop.

Cerebro-cerebellar projections from the lateral suprasylvian visual area in the cat

1. A projection from the medial bank of the lateral suprasylvian visual area, one of the targets of the cerebello-cerebral projection, back to the cerebellar cortex was demonstrated electrophysiologically in the cat. The anatomical pathways underlying this projection were investigated using orthograde and retrograde transport of wheatgerm-agglutinin-conjugated horseradish peroxidase (WGA-HRP). 2. Responses were recorded in the cerebellar cortex on stimulation of the medial bank of the lateral suprasylvian area, and were compared with those evoked by stimulation of the motor cortex and the crown part of the parietal association cortex. 3. Responses induced by stimulation of the lateral suprasylvian area were shown to consist of early mossy and late climbing fibre responses. The mossy fibre response was evoked, at a latency of 2-3 ms, predominantly in the lateral part of the contralateral cerebellar cortex (mainly, crus I, crus II, dorsal paraflocculus and paramedian lobule) and the posterior part of the vermis (mainly, lobules VII and VIII). Climbing fibre responses were elicited with the preceding mossy fibre responses and were elicited at a much longer latency than the motor cortex-induced climbing fibre response. 4. The orthograde and retrograde HRP studies suggested that the mossy fibre response is mediated by the pontine grey whereas the climbing fibre response is conveyed indirectly to the inferior olive which sends the climbing fibres to the cerebellar cortex. After WGA-HRP injections into both the medial bank of the lateral suprasylvian area and the cerebellar responsive area, orthogradely labelled terminals of cortico-pontine projection fibres and retrogradely labelled ponto-cerebellar neurones were found in the pontine grey, where distributions of the two kinds of labelling overlapped. On the other hand, retrograde neuronal labelling alone was found in the inferior olive, implying that the climbing fibre responses evoked from the lateral suprasylvian area were relayed via indirect cortico-olivary pathways.

Morphological features of layer V pyramidal neurons in the cat parietal cortex: an intracellular HRP study

Layer V pyramidal neurons in the cat parietal cortex (areas 5 and 7) were investigated with intracellular HRP staining. Antidromic responses were recorded intracellularly as well as extracellularly with pontine stimulation under Nembutal anesthesia. The relationship between the latency of antidromic responses and the morphology of HRP-stained neurons was analyzed. A total of 65 neurons were stained with HRP, and sixteen of these neurons were activated antidromically with pontine stimulation. Two distinct groups of layer V pyramidal neurons were detected morphologically by intracellular HRP staining; i.e., one (F type) consisted of neurons with relatively large somata (58.4 +/- 8.1 micron X 24.5 +/- 5.1 micron, N = 11) and aspiny or sparsely spinous apical dendrites, and the other (S type) consisted of neurons with smaller somata (44.6 +/- 7.6 micron X 19.3 +/- 3.9 micron, N = 22) and richly spinous apical dendrites. These two groups showed different electrophysiological properties; i.e., the former responded antidromically to pontine stimulation at a latency shorter than 1.5 ms (namely, with a conduction velocity faster than 18 m/second) and the latter responded at a latency longer than 1.5 ms. The two neuronal types in the parietal cortex corresponded respectively to fast and slow pyramidal tract neurons (PTNs) investigated in the sensorimotor cortex. Although their morphological features were almost similar to those of PTNs, the branching pattern of apical dendrites of the F-type pyramidal neuron seemed to be different from that of fast PTNs. In the parietal cortex, apical dendrites of F-type neurons showed rather frequent branching in layer I. This was similar to the pattern of branching in slow PTNs. Such a characteristic branching pattern suggested that, in the cat parietal cortex, layer V pyramidal neurons of both types are adapted to receive cerebellar inputs through the ventroanterior (VA) thalamic nucleus to the superficial cortical layers.

Electrophysiological peculiarities of cortical inputs to the cat red nucleus

Complex multicomponent EPSPs of the red nucleus rubro-spinal neurons evoked by stimulation of the sensorimotor cortex and associative fields of the parietal cortex were studied in acute pentobarbitalized cats by intracellular recording technique. Complex cortical EPSPs were recorded in 2/3 of the neurons studied. Components of the EPSPs in question were distinguished by using stimulation of various frequency and intensity. The first component of the EPSPs appearing at the lowest threshold was found to have a short and stable latency, stable rising time for depolarization and was able to follow high frequencies of stimulation. The second component was more variable, although in some EPSPs it too had a short latency, was stable enough and, like the first component could be classified as monosynaptic. The complex character of the EPSPs recorded persisted after the removal of the cerebral gray and was observed when stimulating the white matter so excluding its cortical origin. The first two components of the EPSP were evoked by corticofugal impulsation propagating at an average velocity of 18.5 m/s and 7.5 m/s being supposedly the result of activation of the slow-conducting pyramidal and cortico-rubral neurons. In some rubro-spinal neurons they were characterized by a fast rising phase being apparently an electrophysiological manifestation of the activation of axosomatic synapses.

Excitatory inputs to cerebellar dentate nucleus neurons from the cerebral cortex in the cat

1. In anesthetized cats, we investigated excitatory and inhibitory inputs from the cerebral cortex to dentate nucleus neurons (DNNs) and determined the pathways responsible for mediating these inputs to DNNs. 2. Intracellular recordings were made from 201 DNNs whose locations were histologically determined. These neurons were identified as efferent DNNs by their antidromic responses to stimulation of the contralateral red nucleus (RN). Stimulation of the contralateral pericruciate cortex produced excitatory postsynaptic potentials (EPSPs) followed by long-lasting inhibitory postsynaptic potentials (IPSPs) in DNNs. The most effective stimulating sites for inducing these responses were observed in the medial portion (area 6) and its adjacent middle portion (area 4) of the precruciate gyrus. Convergence of cerebral inputs from area 4 and area 6 to single DNNs was rare. 3. To determine the precerebellar nuclei responsible for mediation of the cerebral inputs to the dentate nucleus (DN), we examined the effects of stimulation of the pontine nucleus (PN), the nucleus reticularis tegmenti pontis (NRTP) and the inferior olive (IO). Systematic mapping was made in the NRTP and the PN to find effective low-threshold stimulating sites for evoking monosynaptic EPSPs in DNNs. Stimulation of either the PN or the NRTP produced monosynaptic EPSPs and polysynaptic IPSPs in DNNs. Using a conditioning-testing paradigm (a conditioning stimulus to the cerebral peduncle (CP) and a test stimulus to the PN or the NRTP) and intracellular recordings from DNNs, we tested cerebral effects on neurons in the PN and the NRTP making a monosynaptic connection with DNNs. Conditioning stimulation of the CP facilitated PN- and NRTP-induced monosynaptic EPSPs in DNNs. This spatial facilitation indicated that the excitatory inputs from the cerebral cortex to DNNs are at least partly relayed via the PN and the NRTP. 4. Stimulation of the contralateral IO produced monosynaptic EPSPs and polysynaptic IPSPs in DNNs. These monosynaptic EPSPs were facilitated by conditioning stimulation of the CP, strongly suggesting that the IO is partly responsible for mediating excitatory inputs from the cerebral cortex to the DN. A comparison was made between the latencies of IO-evoked IPSPs in DNNs and the latencies of IO-evoked complex spikes in Purkinje cells. Such a comparison indicated that the shortest-latency IPSPs evoked from the IO were not mediated via the Purkinje cells and suggested the pathway mediated by inhibitory interneurons in the DN.(ABSTRACT TRUNCATED AT 400 WORDS)

Brachium pontis lesions in cats partly reproduce the cerebellar dysfunction of voluntary reaching movements

The pontocerebellar pathway in the brachium pontis (BP), is known to convey signals from various cortical and subcortical visual structures to the cerebellum. Recently, a cortico-pontocerebellar pathway involving the BP has been implicated in the control of visually guided movements, on the basis of anatomical and physiological data. To further test this hypothesis, using behavioural methods, we studied the effects of a bilateral interruption of these projections in the BP, on 5 cats fully trained to perform a forepaw movement towards a moving target-light. The postoperative deficit consisted of an impairment in precision, with a strong tendency to over-reach and and increase in reaction time, contrasting with an unimpaired movement time. Although there was some initial recovery, performance soon stabilized with a permanent impairment in accuracy and reaction time. These results are discussed in relation to the various sensory signals processed at the pontine level and forwarded to the cerebellum, and compared with the effects of motor dysfunction of cerebellar origin.

Collaterals of corticospinal and pyramidal fibres to the pontine grey demonstrated by a new application of the fluorescent fibre labelling technique

Selective visualization of collaterals of corticospinal and pyramidal fibres to the pons in cat was obtained by retrograde transport of the fluorescent tracer fast blue (FB) through the stem fibres. Unilateral FB injections in the cervical cord and the pyramidal tract respectively produced soft blue fluorescent labelling of pyramidal fibres and of fibres and structures resembling 'terminals' in the pontine grey: contralateral to the spinal injections and ipsilateral to the pyramidal injections. These labelled elements were concluded to represent collaterals of corticospinal and pyramidal fibres because (a) their distribution corresponded to that of the pericruciate corticopontine fibres, (b) their labelling was prevented when the FB injections were preceded by a transection of either the cerebral peduncle or the pyramidal tract which lesions also prevented the FB labelling of the distal parts of the transected axons. Similar findings were obtained when using wheat germ agglutinin-horseradish peroxidase. In other experiments FB-labelling of pyramidal collaterals was combined with retrograde labelling of pontine neurones projecting to the contralateral anterior lobe of the cerebellum using diamidino yellow dihydrochloride as the second tracer. The distributions of the retrogradely labelled neurones and of the pyramidal collaterals in the pontine grey showed an almost complete overlap indicating that these collaterals mainly establish connections with the cerebellar anterior lobe.

Cortical control of oculomotor functions. II. Vestibulo-ocular reflex and visual-vestibular interaction

The cortical control of the vestibulo-ocular reflex (VOR) and visual suppression of VOR was studied in 13 adult cats with unilateral lesions. VOR was tested in the dark by sinusoidal rotations of the animal at different frequencies. Visual suppression of VOR was tested in the light by keeping the visual field stationary with respect to the animal. No deficits of VOR and visual suppression of VOR appeared following unilateral ablations of visual cortex. Unilateral lesions of different parts of the suprasylvian cortex were made in the posterior and middle suprasylvian cortex involving area 7 and the lateral suprasylvian area (LSA). After middle suprasylvian cortex damage (particularly area 7), all the animals exhibited a VOR asymmetry due mainly to a gain decrease of slow phases directed towards the side of the lesion. In two animals, transitory spontaneous nystagmus was present in the dark with the fast phase directed toward the side of the lesion. Only when LSA was destroyed, could an asymmetry of the visual suppression of VOR be observed with a loss of the visual suppression during ipsilateral rotations. The VOR deficit was transient: spontaneous nystagmus disappeared within the first postoperative week, the vestibular asymmetry and the loss of visual suppression of VOR were no longer present after 2-3 weeks. We conclude that the middle suprasylvian cortex, particularly area 7, exerts an ipsilateral control on the VOR.

Organization of corticopontocerebellar connections to the paramedian lobule in the cat

To reveal the organization and relative magnitude of connections from various parts of the cerebral cortex to the cerebellar paramedian lobule via the pontine nuclei, horseradish peroxidase conjugated to wheat germ agglutinin was injected in the paramedian lobule in conjunction with injection of the same tracer in various parts of the cerebral cortex in 14 cats. Termination areas of cortical fibres (anterogradely labelled) and pontine neurons projecting to the paramedian lobule (retrogradely labelled) were carefully plotted in serial sections through the pons. On the average 89% of all labelled cells were found in the pontine nuclei contralateral to the cerebellar injection, 11% in the ipsilateral pontine nuclei. The highest degree of overlap between anterograde and retrograde labelling was found after injections in the posterior sigmoid gyrus (SmI), while less overlap was found after injections of the anterior sigmoid gyrus (MsI). Injections of the second somatosensory area (SmII) and the parietal association cortex (areas 5 and 7) gave moderate degrees of overlap. Very little or no overlap was found after injections of the premotor cortex (area 6), the visual areas 17, 18 and 19 and the auditory cortex (AI and AII). It is concluded that a major cortical input to the paramedian lobule arises in the posterior sigmoid gyrus (SmI), but that additional contributions arise in the anterior sigmoid gyrus (MsI), the parietal areas 5 and 7 and the second somatosensory cortex (SmII). Among the latter regions probably the parietal areas contribute most. Overlap between terminal regions of cortical fibres and cells projecting to the paramedian lobule takes place at numerous discrete sites at virtually all rostrocaudal levels of the pons. Cerebrocortical afferents via the pontine nuclei to the intermediate zone of the posterior lobe are organized according to the same principles as described previously for cortical afferents to the hemispheral parts of the posterior lobe (crus I and II).

Organization of afferent connections to the lateral and interpositus cerebellar nuclei from the brainstem relay nuclei: a horseradish peroxidase study in the cat

Afferent projections to the lateral (dentate) and interpositus cerebellar nuclei from the brainstem relay nuclei were studied in cats using the horseradish peroxidase (HRP) method. In the first series of experiments, HRP was injected into the brachium pontis. Mossy fiber terminals were anterogradely labeled, predominantly in the lateral (hemispherical) part, moderately in the intermediate part, and slightly in the vermal part of the cerebellum. Besides these terminals in the cerebellar cortex, axon terminals labeled anterogradely were also found in the cerebellar nuclei. The labeled terminals appeared almost exclusively in the lateral nucleus and rarely in the interpositus nucleus. Cells labeled retrogradely were found both in the pontine nuclei and the tegmental reticular nucleus, but not in other brainstem nuclei. In the second series of experiments, HRP was injected into the lateral and interpositus nuclei, and retrograde labeling was examined in the brainstem relay nuclei. After HRP injection into the lateral nucleus, the number of labeled cells was significantly large in the pontine nuclei, but fairly small in the reticular or vestibular nuclei. The number of labeled cells was generally large in the inferior olive, mainly in the principal olive. After HRP injection into the interpositus nucleus, the number of labeled cells was moderate in the reticular or vestibular nuclei, but small in the pontine nuclei. The number of labeled cells in the inferior olive was also large, being distributed mainly in the accessory olives. These results indicate that the pontine nuclei and the principal olive provide major afferent inputs to the lateral nucleus, whereas the reticular nuclei, the vestibular nuclei and the accessory olives are the major afferent sources to the interpositus nucleus.

Cortical control of oculomotor functions. I. Optokinetic nystagmus

The cortical control of horizontal optokinetic nystagmus (OKN) has been studied in 13 adult cats with unilateral lesions. OKN was induced by rotating the visual field around the animals in both binocular and monocular conditions. (1) No deficits of OKN appeared following unilateral ablations of visual cortex. (2) Lesions of different parts of suprasylvian cortex were made: the posterior and the middle suprasylvian cortex involving area 7 and the lateral suprasylvian area (LSA). Only the middle suprasylvian cortex damage produced on OKN asymmetry due to a decrease of the slow-phase velocity directed toward the side of the lesion. The deficits were compensated for within about 10 days. We conclude that the middle suprasylvian cortex and particularly LSA regulate the ipsilateral slow phases of OKN.

Principles of organization of the corticopontocerebellar projection to crus II in the cat with particular reference to the parietal cortical areas

In 13 cats injections of horseradish peroxidase-wheat germ agglutinin in various parts of the cerebral cortex were combined with injections in the cerebellar crus II in the same animal in order to study the cortical regions that may influence the crus II via the pontine nuclei. In 2 cats lesions in the cerebral cortex were combined with horseradish peroxidase injections in the crus II. In the pons terminal regions (anterogradely labelled from the cerebral cortex or containing terminal degeneration) and cell groups retrogradely labelled from crus II were carefully plotted. The pontocerebellar projection to crus II is mainly crossed, on the average 26% of the labelled cells were found in the ipsilateral pons. Some overlap between sites of ending of cortical fibres and sites of origin of fibres to crus II was present in all cases, but the degree of overlap varied considerably, depending on which cortical region was injected. Typically, partial overlap between terminal patches and groups of labelled cells occurred at multiple sites in the pontine nuclei. A major input to crus II appears to come from the parietal region. Experiments with bilateral cortical injections showed that the pontine projection from the parietal region is topographically organized in a precise mosaic pattern of adjacent but apparently non-overlapping patches of termination. Area 6 also has strong connections with crus II, while only very few of the corticopontine fibres from the sensorimotor region overlap with cell groups labelled from crus II. The second somatosensory area and the visual cortex both seem able to influence a small but significant proportion of cells projecting to crus II. In contrast to other cortical regions, the auditory cortex appears to send fibres mainly to cell groups projecting to the ipsilateral crus II. It is concluded that the input to crus II originates in wide areas of the cerebral cortex. Small subgroups of neurons projecting to crus II can be differentiated on the basis of their cortical afferents. It appears likely that each subgroup receives fibres mainly or in some instances only from one cortical site. The corticopontocerebellar projection to crus II probably exhibits a high degree of spatial order providing a specific pattern of convergence and divergence in the cerebellar cortex, in agreement with recent physiological evidence from micromapping studies.

Electrophysiological properties of cortical synaptic inputs of rubro-spinal neurons

Excitatory postsynaptic potentials (EPSPs) of the red nucleus neurons evoked by stimulation of the cerebellar nucleus interpositus as well as the sensorimotor and association parietal regions of the cerebral cortex were studies in acute cats. As for the first two structures, a monosynaptic connection of the association cortex with rubro-spinal neurons was shown to exist. The analyses of the time characteristics of the unitary EPSPs suggested a localization of synapses of fibers from the association cortex closer to the soma when compared with those which originated from axons of the sensorimotor cortical cells.

The corticopontocerebellar pathway to crus I in the cat as studied with anterograde and retrograde transport of horseradish peroxidase

In 13 cats injections of horseradish peroxidase (HRP) in various parts of the cerebral cortex were combined with injections of HRP in the cerebellar crus I in the same animal in order to study the cortical regions which may influence the crus I via the pontine nuclei. In the pons terminal regions anterogradely labeled from the cerebral cortex and cell groups retrogradely labeled from crus I were carefully plotted. Overlap between sites of ending of cortical fibers and sites of origin of fibers to crus I was relatively modest in all experiments. On the other hand, partial overlap was usually found at multiple sites. The largest single input to crus I appears to come from the parietal region, particularly the anterior part of the suprasylvian gyrus, while the sensorimotor region contributes much less. Area 6, the second somatosensory cortex and the orbital gyrus, all seem able to influence pontine cells projecting to crus I. Least overlap is found after injections of the visual cortex. The size and orientation of the dendritic fields of pontine cells were studied in Golgi-impregnated material. The dendritic fields average 187 X 339 microns in the transverse plane and are so small that they will only moderately increase the overlap. It is concluded that small subgroups of neurons projecting to crus I receive somewhat different sets of cortical afferents. The input to crus I must originate in wide areas of the cerebral cortex, and probably exhibits a high degree of spatial order, with an intricate pattern of specific divergence and convergence. The present results are compatible with previous physiological evidence from micromapping studies of a precise and complicated mosaic pattern of connections to the cerebellar hemispheres.

Projections from the pontine nuclei proper and reticular tegmental nucleus onto the cerebellar cortex in the cat. An autoradiographic study

After injections of 0.5 microliter of tritiated leucine and/or proline into various parts of the pontine nuclei proper or the pontine tegmental reticular nucleus (N.r.t.) of 34 cats, labeled terminals of pontocerebellar fibers were found in the cerebellar cortex. Fibers from the pontine nuclei and N.r.t. terminate as mossy fibers in the granular layer of the cerebellum, and no evidence is obtained of labeled fibers in the molecular layer. The pontocerebellar projection is, in general, bilateral with a contralateral preponderance, and a complex organization has been shown to exist in the cat. Clear evidence of divergence of this projection from a small pontine area has been demonstrated. Thus, the dorsolateral nucleus has a heavy projection to lobule VII, besides modest projections to lobules VI, VIII, and IX, crus I and II, paraflocculus, and paramedian lobule. On the other hand, a particular cerebellar region receives afferent fibers from several pontine regions, confirming previous HRP studies. For example, lobule VII receives heavy projections from parts of the dorsolateral, peduncular, and paramedian nuclei, less heavy projections from the lateral part of the lateral nucleus, and some from other parts of the pontine nuclei. This is a convergent feature of the pontocerebellar projections. In addition, small adjoining areas within a pontine subdivision have different patterns of cerebellar projections, shwing preferential sites of terminations. This suggests some degree of localization within the pontine nuclei. The cerebellar projection from the N.r.t. shows an essentially similar organization as the projection from the pontine nuclei proper, an apparent difference being only that the former is more extensive in the fields of termination than the latter. Some evidence for a parasagittal termination of pontocerebellar projections to the paramedian lobule has been found in this study. However, this is not as clear-cut as such patterns in the cerebellar projections from the spinal cord, cuneate nucleus, and lateral reticular nucleus shown recently in rat and cat.

The distribution of pontine projection cells in visual and association cortex of the cat: an experimental study with horseradish peroxidase

The projections from the visual and association areas of the cat's neocortex to the pons were investigated with horseradish peroxidase as retrograde tracer. Small injections were made into the pars basalis of the pons, along its entire rostrocaudal extent. The cortical areas considered were areas 17, 18, 19, 20, 21, and the lateral suprasylvian areas (LSA); the posterior (PMSA), and the anterior middle suprasylvian association area (AMSA), the anterior lateral association area (ALA) and the anterior suprasylvian association area (ASA). A pontine projection was found for all the areas investigated; however, areas differ in the relative strength of their projection, in their intraareal distribution of projection cells, and in the location of their projection zones within the pons. A low to moderate density of projection cells is seen in the areas 17, 18, 19, 20, 21, and in PMSA. The posterior part of LSA contains only a few projection cells, whereas in more anterior parts of LSA the density of projection cells is moderate to high. A relatively dense distribution of projection cells also appears in AMSA, ALA, and ASA. In those areas which are retinotopically organized (17, 18, 19, LSA) the representation of the center of gaze contains far fewer projection cells than the representation of peripheral vision. In the association areas the distribution of projection cells appears even. The projection zones from areas 17, 18, and 19 overlap with the zones from LSA in the anterior half of the basal pons. The projection zones from areas 20 and 21 and from ALA and ASA are located in the middle third and the projection zones from PMSA and AMSA spread throughout the entire rostrocaudal extent of the basal pons. Our findings indicate that efferent impulses from the visual cortical areas and from the association areas on the middle suprasylvian gyrus are relayed to the cerebellum exclusively via the basal pontine nuclei. The findings further suggest that the visual corticopontine projections carry a map of the visual field in which the cortical magnification factor is reduced.

The pontocerebellar system in the rat: an HRP study. II. Hemispheral components

The projection of basilar pontine neurons to the cerebellar hemispheres was studied to pigmented rats by means of the retrograde transport of horseradish peroxidase. Injections of horseradish peroxidase were restricted to the lateral aspects of the lobulus simplex (11 cases), crus I (26 cases), crus II (23 cases), and paramedian lobule (18 cases). The main focus of labeled neurons following lobulus simplex injections of horseradish peroxidase was located in the ventral pons, at rostral levels. Interestingly, the majority of labeled cells were distributed ipsilateral to the injection site. After crus I injections, however, labeled neurons were most evident contralaterally , although labeled ipsilateral cells were conspicuous rostrally. The majority of labeled cells were characteristically distributed along the medial, ventral, and lateral perimeters of the pontine gray. This pattern of labeling contrasts with that in cases of crus II injections, in which the main focus of labeled somata occupied more central regions of medial and ventral portions of the pons. Similarly, the pattern of labeling following injections into the paramedian lobule largely avoided the medial and lateral perimeters of the pontine gray, while numerous labeled somata occupied the central region of the pons. In addition to the pontine regions described above, labeled cells were observed in various cases in the dorsal peduncular region, the lateral and dorsolateral areas, and the nuclear reticularis tegmenti pontis (NRTP) where three separate zones of labeling could be discerned in various cases. Several general organizational features were derived from these studies. Although specific quantitation procedures were not applied, the number of ipsilaterally labeled neurons was impressive in some cases, as was the mirror-image location of certain ipsi- and contralateral cell clusters. It was also noted that certain, similarly located clusters of labeled pontine neurons were present in cases in which injections were made into different cerebellar lobules, at least raising the possibility that some pontine neurons might give rise to divergent projections of multiple cerebellar locations, Moreover, it was evident that the location of certain clusters of labeled neurons was congruent with terminal zones of various pontine afferent systems, particularly those of the sensorimotor cortex. Combining the latter finding with the preceeding notion regarding pontocerebellar divergence suggests a mechanism by which sensorimotor information might be transmitted to several different cerebellar locations.

The parieto-rubro-olivary pathway in the cat

Stimulation of the parietal association cortex as well as the frontal motor cortex elicited clearly extracellular unitary activities or field potentials in the ipsilateral inferior olive in the cat. The parietal-induced responses came out generally at a longer and more variable latency than the frontal-induced ones. This suggested the existence of an indirect pathway from the parietal association cortex to the inferior olive. The recording sites for the parietal-induced responses were located not only in the dorsal lamella but also in the ventral lamella of the principal olive and in the medial accessory olive. Such olivary sites were exclusively in the rostral half of the inferior olive, and these areas in the olive were considered to give projection fibres predominantly to the hemispherical parts of the cerebe-lar cortex (neocerebellum). Small neuronal cells were labelled with horseradish peroxidase (HRP) homolaterally in the midbrain tegmentum, after HRP was injected through recording glass microelectrodes into the inferior olive where only the parietal-induced responses were evidently recorded. These small cells were distributed in the rostral one-third of the red nucleus and/or around the adjacent midbrain reticular formation close to the lateral border of the red nucleus. In referring to recent anatomical and physiological data, such small neurones labelled with HRP could be identified as the parvocellular red nucleus neurones. The present results indicate the existence of the parieto-rubro-olivary pathway system in the cat and suggest, in association with our previous studies, that the parvocellular red nucleus neurones participate in control of highly co-ordinated posture and movement predominantly through the neocerebellum.

Motor responses evoked by microstimulation of restiform body in the cat

Motor effects produced by microstimulation of restiform body (RB) were studied in acute unanesthetized cats, using tungsten electrodes for stimulating the peduncle and bipolar steel electrodes for recording muscular activity (EMG). The main results were the following. 1. Threshold microstimulation (18.24 microA +/- 8.77 S.D.) of effective foci within RB elicited single muscle contractions of ipsilateral limbs, primarily of forelimb; overthreshold activation (32.83 microA +/- 9.25S.D.) of the same points produced complex movements in 61.54% of cases that involved muscles of shoulder, neck, and trunk. 2. Single muscle contractions exhibited a mean latency (20.09 msec +/- 2.04 S.D.) which was significantly longer than that shown by complex movements (10.00 msec +/- 3.10 S.D.). Furthermore, a decrease in frequency of stimulating train below 300 Hz and a reduction in duration below 30 msec caused a steep rise of threshold for single muscle responses that was not observed when studying complex movements. 3. Acute RB interruption between stimulating electrode and cerebellum abolished single muscle contractions; conversely, complex movements remained unmodified even when the RB was lesioned in cats chronically submitted to interruption of brachium conjunctivum (BC). 4. The pathway involved in promoting RB induced single muscle activation includes interpositus nucleus, BC and rubrospinal tract. Possible modalities of RB afferent participation to the motor control are briefly discussed.

Principal neurons of the basilar pons as the source of a recurrent collateral system

The basilar pontine nuclei in the opossum are composed of two general categories of neurons, intrinsic cells and the principal or projection neurons. Observations from Golgi material indicate that principal neurons whose primary axons project to the cerebellar cortex may also give rise to recurrent branches distributing within the pontine gray. Such collaterals were observed to arise near the soma and at some distance from the cell body of the parent axon. The electron microscopic correlate of such a system was identified in the basilar pontine neuropil in animals subjected to lesions of the cerebellar cortex. These lesions destroyed mossy terminals and their parent axons and thus initiated a retrograde reaction in basilar pontine projection neurons which manifested itself in the form of morphologic alterations observed in somata, dendrites, and a class of axonal boutons. Similar altered axon terminals were not observed in control material and did not correspond to the terminals of cerebello-pontine axons described in previous work. It was therefore suggested that such boutons represented the terminals of the recurrent collateral system observed in Golgi material.

The neurons and the synaptic endings in the primate basilar pontine gray

Two types of neurons, projection and intrinsic, previously identified in Golgi preparations of the adult monkey (Macaca mulatta) basilar pontine gray (Cooper and Fox, '76) were observed electronmicroscopically in Macaca mulatta and the squirrel monkey Saimiri sciureus. The cell body of the projection neuron measures up to 37 micrometer and its cytoplasm is rich in organelles. The Goli apparatus, ribosomes, and mitochondria are disposed around the nucleus, while rough endoplasmic reticulum though abundant is usually confined to one half of the cell body. The cell body of the intrinsic neuron measures less than 20 micrometer and its cytoplasm displays prominent ribosomes, but a paucity of other organelles. Five types of synaptic profiles have been identified in the neuropil of the basilar pons; one measures up to 5 micrometer and the rest 2 micrometer or less. They are: (1) a large profile (MSV) containing medium size vesicles (500A) and a central core of mitochondria and neurofilaments; (2) a profile (SSV) containing small round vesicles (250-500 A) which is the most abundant and ubiquitous; (3) a profile (F) containing flattened or pleomorphic vesicles; (4) a profile (LSV) containing large oval egg shaped vesicles (750 A); and (5) a pale profile (PP) that contains oval and occasionally pleomorphic vesicles. MSV, SSV, and LSV terminals form asymmetrical contacts and F terminals form symmetrical contacts with both dendritic and vesicle-containing, pale profiles. The vesicle-containing, pale profile is both pre- and post-synaptic and participates in serial synapses. Following unilateral cortical ablations both dark and filamentous degeneration were observed in the ipsilateral basilar pontine gray.

Responses of pontocerebellar neurones to stimulation of the parietal association and the frontal motor cortices

The corticopontine projections from the parietal association cortex (the anterior portion of the middle suprasylvian gyrus) were electrophysiologically investigated and compared with those from the frontal motor cortex (the anterior sigmoid gyrus) in cats under light Nembutal anaesthesia. It was indicated by field potential study that the pontine nucleus (PN) neurones receive a significant amount of the direct corticopontine fibres from both the parietal and frontal cortical areas. In extracellular unitary study, out of 107 PN neurones identified by antidromic activation due to the brachium pontis stimulation, 33 responded with firings to stimulation of the parietal association area and 64 to the frontal motor area. Only 10 of them were excited by both parietal and frontal stimulations, but they were not estimated to receive the dual monosynaptic projections from both cortical areas. There were found data suggesting that the pontocerebellar neurones with faster conduction velocities respond at shorter latencies to the cortical stimulation and those with slower conduction velocities fire at longer latencies on the cortical stimulation. No remarkable difference was observed between the topographical localization of the PN neurones receiving the projection fibres from the parietal association and the frontal motor cortical areas. It was concluded that a vast majority of the pontocerebellar neurones possibly receive monosynaptic contacts differentially with the corticopontine fibres originating from the parietal association and the frontal motor areas.

Electrophysiological studies of the projections from the parietal association area to the cerebellar cortex

1. Responses evoked in the cerebellar cortex by stimulation of the parietal association cortex (rostral portions of the middle suprasylvian gyrus) were recorded and analysed in cats, and were compared with those by stimulation of the motor cortex (anterior sigmoid gyrus). 2. The parietal stimulation elicited early mossy fibre and late climbing fibre responses in the cerebellar cortex. The mossy fibre responses appeared at a latency of 2.0--2.5 msec and predominantly in the lateral (hemispherical) part of the contralateral cerebellum (mainly crus I, crus II and paramedian lobules). Cutting of the inferior cerebellar peduncle produced little or no influence upon the mossy fibre responses, which suggests that the mossy fibre responses are mediated chiefly by the pontine nuclei. 3. The climbing fibre responses were recorded at a latency of 17--19 msec and markedly in the contralateral intermediate and medial parts of IV--VI lobules. The responses were easily sppressed by anaesthesia and depended on the conditions of experimental animals. The unstable appearance of the responses and their longer latencies than those of the climbing fibre responses due to stimulation of the motor cortex imply indirect pathways from the parietal association cortex to the inferior olive. 4. The predominant projection of the parietal-induced mossy fibre responses to the lateral part of the cerebellum was compared with the mossy fibre projection from the motor cortex and was discussed as an important component in the cerebrocerebellar loops.

Afferents to prefrontal cortex from the thalamic mediodorsal nucleus in the rhesus monkey

Thalamic afferents to Macaque prefrontal cortex from the mediodorsal nucleus were examined by techniques specific for anterograde degeneration and axoplasmic transport. The sampling procedure employed permits establishing the extent of topographic projections to cortex from subcortical foci for the same brain which was surveyed subsequently in tracing specific neuronal connections by electron microscopy. Topographic and general laminar distribution of thalamic terminals are presented in terms of 3 subareas of prefrontal cortex. The dorsolateral and ventral (orbital) surfaces of prefrontal cortex receive respectively projections from the lateral and medial subdivision of the mediodorsal nucleus. In addition, the medial wall of the frontal lobe, including the dorsomedial part of the lateral convexity, heretofore regarded as athalamic, receives input from the caudal-dorsomedial aspect of the mediodorsal nucleus. Preliminary evidence suggests that axons from the mediodorsal nucleus terminate in the head of caudate nucleus, as Sachs-81 described 65 years ago in the first orthograde study of thalamo-prefrontal cortex connections.