Lesioning or stimulating the subthalamic nucleus (STN) in patients with Parkinson's disease, or in animal models of parkinsonism, alleviates many of the symptoms and so it is tempting to think of the STN as a part of the cause of Parkinson's disease. The globus pallidus (GP) is thought to have a tonic inhibitory action on the STN. An ibotenic acid injection into the GP in rats removes the cells of the GP and, over the following 6 weeks, a progressive loss of dopamine cells (counted stereologically in sections stained for tyrosine hydroxylase) develops in substantia nigra (SN). In this investigation we show that, when animals have the STN cells destroyed by very small ibotenic acid injections, their dopamine neurons are not damaged. Furthermore, if a lesion to the GP follows a lesion of STN then the dopamine cells also survive this double insult, at least for the first 3 weeks following the lesion. The experiments provide good reason to suspect that, at least in the short term, increased activity in the STN is a contributory cause of the loss of dopamine cells which follows the lesion of the GP in rats. Whether or not this is part of the mechanism of cell loss in Parkinson's disease, the rats with GP lesions at least provide an opportunity to test strategies that might protect dopamine cells from slowly developing damage. Removing the STN seems to be neuroprotective in this new model of dopamine degeneration.

Parkinson's disease is a debilitating disorder that results from the death of dopaminergic neurones in the substantia nigra. Subthalamic nucleus neurones use glutamate as their neurotransmitter and send excitatory projections to the substantia nigra. Changes in both the mean firing rate and firing pattern of neurones of the subthalamic nucleus have been found in patients with this disease. This has led to the suggestion that hyperactivity of the subthalamic nucleus may be involved in the pathology of the dopaminergic neurones. Subthalamic nucleus lesions or treatment with glutamatergic antagonists can be neuroprotective in animal models of Parkinson's disease but until now there has been no direct evidence that hyperactivity of subthalamic nucleus neurones can lead to downstream cell death. Here we show that lesions of the rat globus pallidus (a treatment that has been shown to increase subthalamic nucleus neuronal activity) result in a significant reduction of the number of dopaminergic neurones in the substantia nigra.

In animal models of Parkinson's disease and in human patients, where dopamine cells in the substantia nigra (SN) are depleted, the mean firing rate of neurons in the subthalamic nucleus (STN) is increased. As these neurons send excitatory, glutamatergic projections to SN, it is possible that this overactivity may result in excitotoxicity, leading to further dopamine cell loss. Lesions of the STN can be neuroprotective against the dopamine cell loss caused by injections of 6-hydroxydopamine into the rat striatum. However, it has not been shown whether STN overactivity can directly cause dopamine cell death. To address this question ibotenic acid lesions were made in the rat globus pallidus, a manipulation that has been shown to increase the firing rate of STN neurons. Three weeks later a significant increase in microglial activity and astrocytic activation in the SN ipsilateral to the lesion was observed, whilst unbiased stereological counting showed the number of dopamine neurons to be reduced by 15%. By 6 weeks after the lesion there was a reduction of dopamine cells of more than 30% ipsilaterally but now contralateral to the lesion dopamine cells were also reduced by 20%. This loss of dopamine cells in the SN after a globus pallidus lesion is suggested to be due to excitotoxicity related to the increased firing of STN neurons.

In Chloralose and Urethane (1% & 10%; 1.0 ml/Kg, i.p.) anaesthetised rats we have studied the responses of single striatal neurones to stimulation of the contralateral barrel cortex and of the contralateral whisker pad. The effects of cortical stimulation are likely to be transmitted along the ‘diffuse’ pathway that projects bilaterally to the striatum. Stimulation of the whisker pad would probably activate the other corticostriatal system, which arises as collaterals from corticofugal fibres. The synchronous volley arising from the electrical stimulus to the whisker pad would have travelled via brain stem and thalamus to the barrel centres in the somatosensory cortex. It is from the barrels that the ipsilateral, topographic system arises and so the peripheral stimulus would be able to activate this system first. Later phases of the response might include activation of the septa surrounding the barrels and hence of the diffuse system. Both stimuli are able to depolarise the spiny neurones to their firing threshold. The excitatory postsynaptic potentials (EPSPs) from the two stimulus sites had similar latencies (in spite of very different path lengths and numbers of synapsesen route).However, the responses to whisker pad stimulation were rapidly rising and showed very little variation in latency to action potential generation, while the contralateral cortically derived EPSPs were slower to rise and the spike initiation latency was more variable. These characteristics might be predicted from the morphology of the two systems.

The intracellular responses of single striatal neurones under Chloralose and Urethane (1% & 10%; 10ml/Kg, i.p.) anaesthesia to stimulation of the contralateral barrel cortex and whisker pad were studied in normal rats and in rats with unilateral 6-hydroxydopamine lesion of midbrain dopaminergic neurones. In normal rats both stimuli depolarise striatal cells and EPSPs in response to whisker pad stimulation have a rapid rise time, while the contralateral cortically derived EPSPs are slower to rise and the spike initiation latency more variable. The two pathways also interact at the level of a single striatal cell. In dopamine depleted animals both stimuli were also able to depolarise the spiny neurones to their firing threshold. However the EPSPs to whisker pad stimulation were significantly slower to rise compared to control animals and were similar to cortically derived EPSPs. The interaction of the two pathways was also affected by the loss of dopamine.

Microinjections of cholera toxin B subunit were made into the area of the neostriatum that receives input from the primary somatosensory barrel cortex (SI) in the rat. Studies of the cortices then allowed retrograde identification of the cortical cells supplying the striatal input. When injections were restricted to the neostriatum, retrograde labelling was found in layer V of both SI cortices. Ipsilateral to the injection, cells were retrogradely filled with toxin in all parts of the barrel field, in adjacent parietal cortex, in the motor cortex and in prefrontal areas. A similar distribution across cortical areas was seen contralaterally; however, the stained cells in the SI were between rather than within barrel columns. An earlier anterograde study suggested two inputs from the SI to the neostriatum. The present results indicate that one input to the somatosensory area of the neostriatum arises bilaterally from neurons between the barrels of the SI, while the topographic pathway from below the barrels is present only ipsilaterally. These anatomical results indicate that separate stimulation of the two corticostriatal pathways from the barrel cortex is possible. Electrical stimulation of the contralateral cortex will activate the bilateral pathway, while electrical stimulation of the whisker pads activates the barrels and hence the topographic pathway. Neurons in the somatosensory region of the striatum responded to stimuli in the contralateral cortex and in the contralateral whisker pad. In spite of very different path lengths, stimuli via the two routes gave rise to excitatory postsynaptic potentials in the striatal cells with similar latencies. The excitatory postsynaptic potentials to whisker pad stimulation had a rapid rise time and usually resulted in at least one action potential. Responses to stimulation of the contralateral cortex rose to a peak more slowly and were more variable in latency, but also gave rise to an action potential in the majority of cases. All the neurons had the physiological characteristics of medium-sized densely spiny cells and after intracellular filling with biocytin had the appropriate morphology. In summary, we propose that two corticostriatal pathways arise from layer V cells in the barrel area of the somatosensory cortex; one is bilateral and arises from cells mainly below the septa, while a topographical pathway arises from cells below the barrels. Both pathways can raise the spiny output cells of the striatum to firing threshold. The latencies from the contralateral cortex imply slowly conducting fibres with considerably more temporal dispersion than the pathway from below the barrels, which we excited from the contralateral periphery.

The cortical representations of the vibrissae of the rat form a matrix in which each whisker has its own area of cortex, called a ‘barrel’. The afferent pathways from the periphery travel first to the trigeminal nuclei and thence via the ventroposteromedial thalamus (VPM) to the cortical barrels have been described in detail. We have studied the output from barrels by filling adjacent areas of the primary somatosensory cortex (SI) with either Phaseolus vulgaris leucoagglutinin (PHA‐L) or biotinylated dextran amine (BDA) and demonstrating the course and terminations of the axons that arise within the barrel fields. The method not only dramatically illustrates the previously described corticothalamic pathway to VPM but also demonstrates a strict topography in the cortical afferents to the thalamic reticular nucleus (RT). Cells supplying the RT projection are found below the barrels in layer IV. Connections to the posterior thalamus, on the other hand, have no discernible topography and are derived from cortical areas surrounding the barrels. Thus the outputs of these ‘septal’ areas return to the region from which they receive thalamic input. The corticocortical connections are also visible in the same material. Contralateral cortical connections arise from the cells of the septa between barrels. The projections to secondary somatosensory area (SII) are mirror images of the barrel pattern in SI with rather more overlap but nonetheless a recognisable topography.

The sensory input to the neostriatum from groups of cortical cells related to individual facial vibrissae has been investigated at both light- and electron-microscopic resolution. The purpose of the study was to establish the extent to which corticostriatal input maintains the anatomical coding of spatial information that is present in cortex. A double anterograde tracing method was used to identify the output projections from groups of adjacent neurons in different barrel columns, so that the anatomical relationships between two groups could be studied throughout their length. Adjacent whiskers are represented in adjoining cortical barrels and an examination of corticostriatal projections from these reveals two patterns of projection. In one, the anatomical topography is partially preserved; the barrels are represented in adjoining, discrete, areas of the somatosensory neostriatum. In the second projection pattern, the neostriatal innervation is diffuse and adjacent barrels are represented in overlapping regions of the neostriatum. Moreover, the fibres are thinner, have smaller boutons, and are present in both the ipsilateral and contralateral neostriatum. The two systems also enter the neostriatal neuropile separately. The discrete topographic system enters the adjacent neostriatum as collaterals which leave the descending corticofugal fibres at right angles, while the diffuse system enters directly from the corpus callosum at an acute angle. Examination of the neostriatal terminal fields by correlated light and electron microscopy, shows that characteristic axospinous terminals on spiny neurons are made by both groups of cortical fibres, although they differ in their size and morphology. It is concluded that at least two corticostriatal pathways arise from the barrel cortex. One connection maintains some of the anatomical code implicit in the barrel pattern of primary somatosensory cortex, but another, more diffuse, system is overlaid upon it which may carry different information from this complex area of cortex.

In addition to the spiny cells of the neostriatum, which compose more than 95% of the population of neurones, there are a number of interneurones whose chemical identity is known but whose roles in striatal neurophysiology are still controversial.

Embryonic nigral or raphé grafts implanted in the dopamine-depleted neostriatum contain serotonergic neurones and give rise to a serotonergic hyperinnervation of the host striatum. No similar pattern of hyperinnervation is observed in rats with striatal tissue grafts, rats with nigral grafts implanted in the intact striatum, or rats with lesions alone. These observations suggest that striatal dopamine denervation induces some target-derived trophic factors that stimulate fibre growth from serotonergic as well as dopaminergic neurones.

Injections of the anterograde tracer Phaseolus vulgaris-leucoagglutinin within the ventromedial thalamic nucleus resulted in many filled fibres in the frontal areas of rat cerebral cortex. The fibres were restricted to the upper part of layer I except in a small area of motor cortex where terminals were also found in deeper layers. Terminals were also seen in the striatum, in parts of the mesencephalic reticular formation and occasionally in the contralateral ventromedial nucleus. There is some topographical order in the projection with medial and dorsal areas well represented in medial cortex while lateral parts of ventromedial nucleus are more directly related to the cortical area that receives the ventrolateral thalamic nucleus projection. Electron microscopic examination showed the terminals in layer I of cortex making synaptic contact with dendritic spines and small dendritic profiles that showed a very dense postsynaptic specialization. Neurons in the ventromedial nucleus could be antidromically driven from electrode positions along strips of cortex which could not be easily related to any known organizational pattern in the cortex. Thalamic neurons responding antidromically to only one stimulation site were more common when the stimulation was within motor cortical areas, suggesting that in this region a more restricted pattern of termination is the rule.

In the hypothalamus of androgen-insensitive testicular feminized (Tfm) mice the normal pattern of immunohistochemical staining for glial fibrillary acidic protein (GFAP) is markedly different from normal. Along the borders of the third ventricle and in the dorsomedial and arcuate nuclei, the numbers of stained astrocytes are increased. The usual ordered array of tanycytic processes is obscured by a tangle of GFAP-stained stellate glial cells. GFAP immunostaining in other regions of the Tfm forebrain is similar to that in normal mice. These results suggest that the distribution of reactive glia in the hypothalamus may have been changed as a consequence of the genetic defect in Tfm mice.

Anatomical studies in several species have demonstrated a pallidostriatal pathway. We employed electrophysiological and anatomical methods to distinguish the neurones of this pathway from the two pallidocortical groups reported in the rat. Injection of fluorescent retrograde tracers, combined with immunohistochemistry for choline acetyltransferase, provided anatomical evidence of the distinction between neurones of the cholinergic pallidocortical projection and pallidal cells retrogradely labelled from the striatum. Extracellular recordings made in the globus pallidus of halothane-anaesthetised rats provided an electrophysiological description of the pallidostriatal pathway. In some animals neurones of this pathway were distinguished from pallidocortical neurones as a stimulating electrode, situated in the crus cerebri, permitted identification of neurones which projected through this region as well as to the striatum. Neither pallidocortical pathway is reported to have descending axons travelling in the crus cerebri at this point. A preliminary electrophysiological study was carried out in rats in which the dopamine-containing cells of substantia nigra had been destroyed by injections of 6-hydroxydopamine at least 6 months prior to the recording. In the globus pallidus on the lesioned side the mean firing rate of neurones was increased compared with controls. No specific change in firing pattern was noted but the neurones were more responsive to striatal stimulation suggesting that long-term dopamine denervation alters the sensitivity of neurones of globus pallidus.

Iontophoretic injections of the Phaseolus Vulgaris leucoagglutinin (PhAL) in the region of the ventromedial thalamic nucleus of the rat resulted in a complete “Golgi-like” filling of small groups of neurones. The gross distribution of the terminals across the surface of the cortex confirms earlier studies with other tracing methods. The detail available in the PhAL material suggests a rather more restricted domain of influence for individual cells. Some terminal axons were limited to a small area in the motor cortex with boutons in both deeper and surface layers. Other terminals ran for several millimeters, parallel to the surface and immediately below it. In related electrophysiological experiments two groups of ventromedial cells were encountered. One group was antidromically activated by threshold stimulation only in a small area of cortex. Another was characterised by activation from several points on the cortical surface along a strip with a maximum width of 0.5mm but which could be as long as the exposed area of cortex.

Nuclear yellow injections into the striatum of rats led to staining of both neuronal and glial nuclei even after less than 24 h ‘transport time’. The distribution of some of the glial nuclei suggested that the dye was being transported along the striato-nigral pathway. Staining in these areas was greatly reduced if kainic acid was co-injected with the Nuclear yellow.

This study of the relationships between cells identified by their catecholamine fluorescence and their less fortunate neighbours became possible with the advent of autoradiographic tracing methods. A major output from the neostriatum returns to the substantia nigra where it fills the pars reticulata. Outputs from this area of substantia nigra are present on both sides of the brain in the thalamus, in parts of parafascicular, intralaminar, and mediodorsal nuclei, and the superior colliculi in the deeper layers. Mainly unilateral pathways reach the ventromedial nucleus of thalamus and also pass under the lateral part of the colliculus to reach the region of the nucleus pendunculo-pontinus among the fibres of the brachium conjunctivum. The roles of those areas in the transmission of the output of the basal ganglia to the motor system of the animal, however, remain obscure.

Radioactive proteins seen in autoradiographs of the rat striatum after injection of radioactive leucine into the substantia nigra are not evenly distributed. Areas of more intense labelling are regularly spaced in the nucleus in rows parallel to the edge of the striatum.

Subsequent of their achieving stable rates of responding for electrical stimulation through electrodes implanted in the region of the brachium conjunctivum, rats were anaesthetized and 6-hydroxydopamine injected through an implanted guide tube into the area of lateral hypothalamus in which the ascending axons of dopamine-containing neurones are found. The resultant destruction of these axons had similar temporary effects on the behaviour of the animals whether it was ipsilateral or contralateral to the electrode. It is concluded that nigral or mesolimbic dopaminergic systems are not involved in pontine intracranial self-stimulation behaviour.

In animals pretreated with 6-hydroxydopamine, the descending efferent connections of the substantia nigra are restricted to a bilateral projection to the superior colliculus and a pathway which stops in the dorso-lateral reticular formation of the pons.