Changes in thalamic nociception resulting from morphine- and meperidine-dependence in rats

Rats were injected with progressively increasing doses of morphine or meperidine during a period of 3 to 40 days. From this colony of animals individual rats were used at 3- to 4-day intervals for electrophysiologic experiments to analyze the activity of nociceptive neurons in the somesthetic thalamus. After an i.p. injection of chloralose-urethane and the appropriate preparation for a stereotaxic microelectrode penetration of the thalamus, a nociceptive neuron was identified in the nucleus ventralis posterolateralis by its unique spacing of spike potentials emitted in response to pricking the foot with a pin. In addition to the short-latency response that formed a high activity peak on poststimulus time histograms, spikes following the stimulus up to 500 ms also formed activity peaks. Single-pulse stimulation of the sciatic nerve evoked the same response as pinpricks, but innocuous stimuli (pin shielded with a piece of cork) evoked a response without the late activity peaks. Only neurons that exhibited this differential response were regarded as nociceptive. Their response and spontaneous activity were accumulated separately on a digital computer. Following this, naloxone was infused i.v. and the computer accumulations were repeated. It was found that during naloxone-precipitated narcotic withdrawal, innocuous stimuli evoked responses indicative of pain; the nociceptive system was sensitized. Furthermore, a small dose or morphine or meperidine heightened the sensitization. This action of the narcotic agents was reversed by 5-hydroxytryptophan, which assisted the narcotics in suppressing pain in morphine- or meperidine-dependent rats but had no demonstrable effect in control animals. The spontaneous tonic activity of the nociceptive neurons of the somesthetic thalamus was high in rats exhibiting narcotic dependence. Naloxone decreased the count, but not to the value of the control animals. The sensitization of nociception can be explained by a decreased action of a neural pathway that descends from the periaqueductal gray matter via the nucleus raphe magnus to the spinal cord and there blocks the excitation of the spinothalamic tract cells by A-delta and C fibers. The mechanisms that increase the spontaneous activity of the thalamic nociceptive neurons remain unclear.     

An assessment of the antinociceptive and aversive effects of stimulating identified sites in the rat brain

At many sites in the brain electrical stimulation with low current intensity is both aversive and causes antinociception. In view of the well documented antinociception caused by various types of stress and pain it is possible that in some parts of the brain the antinociception is secondary to the stress of the stimulation. At 114 sites in the rat brain the intensity of stimulation required to evoke an aversive response has been compared with the antinociceptive current intensity. Only stimulation in the dorsal hippocampus and pretectal area caused antinociception without significant aversion. Strong aversion resulted from stimulation of 46% of the sites including the central gray and nucleus raphe magnus. Antinociception was significantly correlated with the aversiveness of the stimulation although in 15% of the stimulation sites strong aversion was seen with no antinociception. It is concluded that there can be little justification in assigning a primary antinociceptive role to a brain area which evokes strong escape reactions when stimulated.     

Spinal monoamine and opiate systems partly mediate the antinociceptive effects produced by glutamate at brainstem sites

The administration of the excitatory amino acid glutamate into the periaqueductal gray (PAG) or ventromedial medulla (VM) resulted in a reliable, short lasting elevation in the tail flick and hot plate response latencies in rats. The prior intrathecal administration of methysergide (30 micrograms) or phentolamine (30 micrograms) into the lumbar space produced a significant antagonism of the elevated tail flick reflex latencies evoked by glutamate given into the PAG and VM. Intrathecal naloxone (10 micrograms) significantly antagonized the effects on tail flick produced by VM, but not PAG, injections of glutamate. No intrathecal treatment significantly antagonized the effects of intracerebral glutamate on the supraspinally organized hot plate response. These results indicate that the excitation of glutamate-receptor linked systems in the PAG and VM exert a powerful antireflexive effect on spinal processing by the activation of spinopetal monoamine pathways, but that their mechanisms do not totally overlap as the VM systems also directly or indirectly activate a naloxone sensitive link in the spinal cord. The failure to antagonize the supraspinally organized hot plate response by intrathecal antagonists indicates that aside from an effect (if any) on spinal sensory processing, these brainstem systems may also act at the supraspinal level to actively modulate the animal's response to an otherwise aversive somatic stimulus.     

Diencephalic modulation of activities of raphe-spinal neurons in the cat

The modulatory effects of diencephalic stimulation on the activities of raphe-spinal neurons were studied extracellularly in cats. Among 240 raphe neurons recorded, 57 neurons were activated antidromically by stimulation of the cervical dorsolateral funiculus. These raphe-spinal neurons were found in the caudal raphe nuclei, i.e., the raphe magnus (43 neurons), raphe obscurus (11), raphe pallidus (2), and raphe pontis (1). All of them responded to innocuous and/or noxious peripheral mechanical stimuli with a broad receptive field. The activities of the majority of these neurons were facilitated by trains of pulse stimulation of the rostral periaqueductal gray and the thalamic relay nucleus but not of the thalamic center median nucleus. The facilitation of firing persisted for more than 3 min after the cessation of train pulse stimulation when the stimulation was applied at 20 Hz for 5 to 30 s. This facilitation was not affected by decortication of the sensorimotor area bilaterally. The facilitatory response to periaqueductal gray stimulation was markedly suppressed by systemic administration of naloxone. On the other hand, that of the thalamic relay nucleus stimulation was found to be unaffected. Based on these findings, the mechanisms of pain relief by stimulation of the rostral periaqueductal gray and thalamic relay nucleus reported in human intractable pain appear to relate, at least partly, to the activation of raphe-spinal neurons. However, the paths to raphe-spinal neurons of stimuli from the periaqueductal gray and the thalamic relay nucleus are thought to be independent from each other based on the different effects of naloxone.     

The morphological diversity and receptive fields of spiking local interneurons in the locust metathoracic ganglion

Twenty-one types of spiking local interneurons are described in a segmental ganglion of the locust. All have their cell bodies in a group at the ventral midline of the metathoracic ganglion. The interneurons are characterized by their shape as revealed by intracellular injection of dye, and by their physiology as revealed by intracellular recording. Each interneuron conforms to a basic plan, but the characteristic shape of each is derived from the elaboration of branches in some regions of the neuropil and by their absence in other regions. Some interneurons have ventral branches that extend over most of one-half of the metathoracic neuropil, whilst others have ventral branches restricted to a small region of neuropil. A few interneurons have dorsal branches that enter the first abdominal neuromere . Each type of interneuron is excited by a specific array of mechanoreceptors on the hind leg ipsilateral to its neuropilar branches. Some interneurons have a wide receptive field that encompasses most of the dorsal surfaces of the distal three parts of a leg, whilst others have a field limited to the spurs at the distal end of the tibia. The relationship between the shape of an interneuron and the size or orientation of its receptive field is discussed.     

Brainstem influences on transmission of somatosensory information in the spinocervicothalamic pathway

The inhibition of somatosensory responses of lateral cervical nucleus neurons resulting from stimulation of the brainstem has been investigated. Single unit extracellular recordings were obtained from neurons in the lateral cervical nucleus of chloralose-anesthetized cats. Electrical stimulation of the periaqueductal gray, nucleus raphe magnus, nucleus cuneiformis, and nuclei reticularis gigantocellularis and magnocellularis was found to be very effective in inhibiting the responses of lateral cervical nucleus neurons evoked by electrical or tactile stimulation of the skin. Additional experiments were performed to determine whether the inhibitory effects were mediated in the spinal cord dorsal horn or in the lateral cervical nucleus. These experiments which examined the effect of brainstem stimulation on the responses induced by stimulation of the dorsolateral funiculus or on the antidromic latency of activation of lateral cervical nucleus neurons from thalamus, revealed that most and possibly all the inhibition could be accounted for by an action on the spinal cord. These results are consistent with other studies showing that spinocervical tract cells in the spinal cord can be inhibited by stimulation of the same brainstem regions.     

GABAergic circuitry in the rostral ventral medulla of the rat and its relationship to descending antinociceptive controls

This study used postembedding immunocytochemistry to examine the organization of GABA-immunoreactive synapse in the rostral ventral medulla (RVM) of the rat. To determine whether the outflow neurons of the RVM are under GABAergic control, we examined the distribution of GABA-immunoreactive synapses upon bulbospinal projection neurons that were labelled by retrograde transport of wheatgerm agglutinin-HRP from the cervical spinal cord. To study the possible convergence of GABAergic and periaqueductal gray (PAG) synaptic inputs to RVM neurons, we also made lesions in the PAG and examined the relationship between degenerating PAG axons and GABA-immunoreactive terminals. Approximately 45% of all synapses in the RVM, which includes the midline nucleus raphe magnus and the nucleus reticularis paragigantocellularis lateralis, were GABA-immunoreactive. The vast majority of GABA-immunoreactive terminals contained round, clear, and pleomorphic vesicles and made symmetrical axodendritic synapses; axoaxonic synapses were not found. Almost 50% of the retrogradely labeled dendrites in the NRM were postsynaptic to GABA-immunoreactive terminals. Several examples of convergence of degenerating PAG terminals and GABAergic terminals onto the same unlabelled dendrite were also found. These data indicate that the projection neurons of the RVM are under profound GABAergic inhibitory control. The results are discussed with regard to the hypothesis that the analgesic action of narcotics and electrical stimulation of the midbrain PAG involves the regulation of tonic GABAergic inhibitory controls that are exerted upon spinally-projecting neurons of the nucleus raphe magnus.     

Inhibition of the tooth pulp-evoked jaw opening reflex by stimulation of raphe nuclei in the rat

Effects of conditioning electrical stimuli applied to nucleus raphe dorsalis (NRD), nucleus raphe medialis (NRMed), and nucleus raphe magnus (NRMag) on the jaw opening reflex (JOR) were compared using Wistar rats. Activity of the digastric musle after tooth-pulp stimulation was used as an index of the magnitude of the JOR which was considered a noxious response. The JOR was suppressed to about 30 to 40% of the control after a single-pulse stimulation of the NRD, NRMed, or NRMag. None of these raphe-induced suppressions was reversed by naloxone, a specific opiate antagonist, suggesting that these effects are not related to the endogenous opiate system. By contrast, 50-Hz electrical pulse trains of 0.1-ms duration for 1 min applied to the NRD and NRMed suppressed the reflex to about 30% of the control, an effect sustained for 10 to 20 min, whereas this stimulation of the NRMag suppressed the reflex to about 20% of the control and the suppressive effects were maintained for more than 30 min. In these cases, NRMed- and NRd-induced suppression was partially antagonized and the NRMag-induced suppression was almost completely antagonized by naloxone, suggesting that this long-term suppression, especially that induced by NRMag, is closely related to endogenous opiates. These results show that raphe-induced suppression may be produced by complex actions of a short-term neural mechanism and a long-term opiate-related mechanism.     

Postsynaptic inhibition of cat medullary dorsal horn neurons by stimulation of nucleus raphe magnus and other brain stem sites

Sensory responses of neurons in the medullary and spinal cord dorsal horn can be inhibited by stimulation of a number of brain stem regions. These regions include the nucleus raphe magnus (NRM), the nucleus reticularis gigantocellularis (NGC), the nucleus reticularis magnocellularis (NMC), the periaqueductal gray (PAG), and the nucleus cuneiformis (CU). The purpose of this study was to determine whether or not this inhibition is mediated by postsynaptic processes. Experiments were carried out on chloralose-anesthetized cats. The responses of 29 medullary dorsal horn (trigeminal subnucleus caudalis) cells were recorded with carbon-fiber microelectrodes. Included were cells which responded to noxious stimulation (nine cells) as well as cells which responded only to nonnoxious input. The presence of postsynaptic inhibition was tested by two indirect techniques. We studied the effects of conditioning stimulation of the five regions on the latency of antidromically activated cells and also on the firing rate of neurons excited by iontophoretically applied glutamate. Conditioning stimulation was associated with a block or increased latency of antidromic activation in 15 of 18 nociceptive and nonnociceptive neurons. These effects reflect membrane hyperpolarization, presumably resulting from postsynaptic inhibition. Furthermore, conditioning stimulation of these regions inhibited the glutamate-evoked firing of all 11 cells tested, also indicating a postsynaptic type of inhibition of medullary dorsal horn cells. Thus these results indicate that at least part of the inhibition induced by stimulation of the NRM, NGC, NMC, PAG, and the CU probably results from postsynaptic inhibitory mechanisms.     

Relative contributions of the nucleus raphe magnus and adjacent medullary reticular formation to the inhibition by stimulation in the periaqueductal gray of a spinal nociceptive reflex in the pentobarbital-anesthetized rat

Ammonia intoxication decreases the hyperpolarizing action of postsynaptic inhibition. This study examines the metabolic state of the spinal cord during this effect of ammonia intoxication on spinal motoneurons. ATP, ADP, AMP, the adenylate energy charge, glucose, PCr, pyruvate, alpha-ketoglutarate and glutamate were unchanged during the effect of ammonia on the hyperpolarizing action of postsynaptic inhibition. NH4+, glutamine and lactate were increased. Ammonia intoxication affected postsynaptic inhibition without changes of the resting membrane potential, the neuron input resistance, the action potential and EPSPs. The encephalopathy caused by ammonia intoxication is known to occur without an alteration of the tissue energy state. The effect of ammonia intoxication on postsynaptic inhibition can be considered as a cause of the encephalopathy because postsynaptic inhibition is altered without a change of the tissue energy state, the resting membrane potential, the whole neuron resistance, the action potential and EPSPs.     

Influences of locus ceruleus, raphe dorsalis, and periaqueductal gray matter on somatosensory-recipient thalamic nuclei

Spontaneous and evoked discharge of neurons in the nucleus ventralis posterolateralis (VPL) and spontaneous discharge of neurons in the posterior group and nucleus lateralis posterior (LP) were conditioned by brief trains of stimuli to the locus ceruleus (LC), raphe dorsalis (RD), and periaqueductal gray matter (PAG) in cats anesthetized with pentobarbital or ketamine. Stimulation of LC and RD was without effect on VPL neurons, but induced a long-latency, long-lasting inhibition of LP neurons. Stimulation of the PAG induced marked inhibition of the firing of neurons in all three thalamic nuclei. No differences were found between cats anesthetized with ketamine or pentobarbital.     

Comparison of the effects of ventral medullary lesions on systemic and microinjection morphine analgesia

The effects of electrolytic lesions of the nucleus raphe magnus (NRM), nucleus reticularis paragigantocellularis (PGC) and nucleus raphe alatus (NRA) on analgesia elicited in the rat from systemic morphine and morphine microinjection into the periaqueductal gray (PAG) were evaluated using the tail flick test. No consistent change in baseline pain sensitivity was observed following lesions of the NRM, PGC or NRA. To determine the effect of ventral medullary lesions on systemic porphine analgesia, pain sensitivity was assessed prior to and 40 min after 6 mg/kg morphine administration (i.p.) at 2 days preceding lesioning and 5, 12 and 19 days post-lesion. NRM and PGC lesions produced only slight reductions in analgesia at 5 days after surgery. It was observed that large NRM, large PGC, and NRA lesions significantly attenuated analgesia evaluated at 12 days post-lesion. Smaller lesions confined within the NRM or PGC were reliably less effective than the larger lesions in reducing analgesia. In a subsequent study, 5 μg morphine in 0.5 μl saline was microinjected into the ventral PAG at the level of the dorsal raphe. Identical testing procedures were used and the analgesia was assessed at 2 days before lesioning and 5 and 12 days post-lesion. In contrast to the previous study, large NRM lesions abolished analgesia as early as 5 days following lesioning. Small NRM lesions were less effective and PGC lesions were generally ineffective in attenuating analgesia induced by morphine microinjection. We conclude that the NRA may act as a functional unit in the mediation of systemic morphine analgesia. In contrast, analgesia elicited from intracerebral (PAG) morphine microinjection is mediated via the NRM.     

Midbrain nuclei projecting to the medial medulla oblongata in the monkey

To identify the midbrain nuclei that project to the medial part of the lower brainstem in the monkey, labeled cells were mapped in the midbrain following the injection of horseradish peroxidase into the medial medulla oblongata. After the general distribution of labeled cells was observed in three animals with large injections, more discrete injections of HRP were made in different locations in six additional animals. The small injections were centered in the nucleus raphe magnus, nucleus reticularis giganto-cellularis, or nucleus medullae oblongatae centralis. The five labeled mid-brain nuclei were the periaqueductal gray, nucleus cuneiformis, deep layers of the superior colliculus, nucleus of Darkschewitsch, and the interstitial nucleus of Cajal. In addition, the parvocellular division of the red nucleus and the posterior pretectal nucleus contained large numbers of cells when the injection spread into the inferior olive. No major differences in the distribution of labeled cells between different injection sites were found with the exception that the superior colliculus did not contain any labeled cells when the injection was restricted to midline structures. The functional implications of these anatomical findings are discussed in relation to the descending control of pain.     

The role of acetylcholine in the function of the nucleus raphe magnus and in the interaction of this nucleus with the periaqueductal gray

The nucleus raphe magnus (NRM) plays an important role in the inhibition of pain. Although this region receives afferents from several areas of the brain, the afferent input from the periaqueductal gray (PAG) has been shown to have significant physiological importance. Together, these two sites constitute the major component of a descending network involved in pain inhibition. In this study the role of acetylcholine (ACh) in the function of the NRM was investigated and the possibility that ACh may be a transmitter between the PAG and the NRM was tested. ACh was applied iontophoretically. Scopolamine and gallamine were used to test the type of cholinergic receptors that are present in the NRM. The results of this study shows the following. (1) The majority of the cells in the NRM are excited by ACh. (2) This response to ACh is partially or totally blocked by scopolamine whereas gallamine does not block the response. (3) There is no correlation between the excitatory response to stimulation of PAG and to ACh. There are cells that respond to PAG stimulation by inhibition but are excited by ACh and there are a few cells that are inhibited by ACh but are excited by PAG stimulation. (4) Scopolamine, at a dose that blocks the ACh response, does not block the response to PAG stimulation. (5) There is no correlation between the response to ACh and the type of projection (direct or indirect) to the spinal cord, as tested by stimulation of the dorsolateral funiculus. From these results it is concluded that ACh is an excitatory transmitter at the NRM region but this transmitter does not mediate the interaction between the PAG and NRM.     

Dorsal mesencephalic projections to pons, medulla, and spinal cord in the cat: limbic and non-limbic components.

The vertebrate dorsal mesencephalon consists of the superior colliculus, the dorsal portion of the periaqueductal gray, and the mesencephalic trigeminal neurons in between. These structures, via their descending pathways, take part in various behavioral responses to environmental stimuli. This study was undertaken to compare the origins and trajectories of these pathways in the cat. Injections of horseradish peroxidase into the cervical spinal cord and upper medullary medial tegmentum retrogradely labeled cells mainly in the contralateral intermediate and deep superior colliculus, and in the ipsilateral dorsal and lateral periaqueductal gray and adjacent tegmentum. Only injections in the medullary lateral tegmental field labeled mesencephalic trigeminal neurons ipsilaterally. Autoradiographic tracing results, based on injections across the dorsal mesencephalon, revealed three efferent fiberstreams. A massive first fiberstream (limbic pathway), consisting of thin fibers, descended ipsilaterally from the dorsal and lateral periaqueductal gray and adjacent superior colliculus through the mesencephalic and pontine lateral tegmentum, terminating in these areas as well as in the ventral third of the caudal pontine and medullary medial tegmentum. A few fibers from the dorsal periaqueductal gray matter (PAG) were distributed bilaterally to the dorsal vagal, solitary, and retroambiguus nuclei. The second fiberstream (the predorsal bundle) descended contralaterally from the superior colliculus (SC) and consisted of both thick and thin labeled fibers. The thin fibers terminated bilaterally in the dorsomedial nucleus reticularis tegmenti pontis and the medial half of the caudal medial accessory inferior olive. The thick fibers targeted the contralateral dorsal two thirds of the caudal pontine and medullary medial tegmental fields, and the facial, abducens, lateral reticular, subtrigeminal, and prepositus hypoglossi nuclei. A few fibers recrossed the midline to terminate in the ipsilateral medial tegmentum. Caudal to the obex, fibers terminated laterally in the tegmentum and upper cervical intermediate zone. From the lateral SC, fibers terminated bilaterally in the lateral tegmental fields of the pons and medulla and lateral facial subnuclei. The third fiberstream (mesencephalic trigeminal or Probst tract) terminated in the supratrigeminal and motor trigeminal nuclei, and laterally in the tegmentum and upper cervical intermediate zone. In summary, neurons in the PAG and in the deep layers of the SC give rise to a massive ipsilateral descending pathway, in which a medial-to-lateral organization exists. A similar topographical pattern occurs in the crossed SC projections. The possibility that these completely different descending systems cooperate in producing specific defensive behaviors is discussed.     

Morphine microinjected into the periaqueductal gray has differential effects on 3 classes of medullary neurons

The effects of microinjection of 5–10 μg of morphine into the midbrain periaqueductal gray (PAG) on the activity of neurons in the rostral ventral medulla (RVM) were studied in lightly anesthetized rats. Based on the relationship between changes in neuronal activity and the occurrence of the tail-flick reflex (TF), RVM neurons were divided into 3 groups: off-cells, on-cells and neutral cells. The off-cells exhibited an abrupt pause and the on-cells an acceleration beginning just prior to the occurrence of the TF. Neutral cell firing did not change at the time of the TF. Microinjections of morphine into the PAG which inhibited the TF had differential effects on the spontaneous activity of the 3 groups of neurons in RVM. Off-cells showed an increase and on-cells a decrease in spontaneous activity which preceded the inhibition of the TF. These microinjections also reduced the TF-related responses of off- and on-cells. The effects on cell activity were reversed by systemically administered naloxone and were not seen following microinjections which failed to block the TF. Neutral cell activity was unchanged following microinjection of morphine into the PAG. These results support the hypothesis that off- and on-cells in the RVM mediate the effects of microinjection of morphine into the PAG on spinal nociceptive reflexes.     

Effect of interruption of bulbospinal pathways on lordosis, posture, and locomotion

The effect of interruption of descending bulbospinal projections arising from (i) the medial medullary reticular formation: nucleus reticularis gigantocellularis (NGC) and nucleus reticularis magnocellularis (NMC), (ii) nucleus raphe magnus, and (iii) nucleus subceruleus on the lordosis reflex, posture, and locomotion was examined. After interruption of descending NGC-NMC fibers, time-dependent deficits in lordosis were observed during the first two postoperative weeks with subsequent reflex recovery. A syndrome of postural and locomotor deficits occurred after NGC-NMC lesions. Postoperatively, postural reflexes were dominated by increased extensor tone which resulted in limb hyperextension. Specific time-dependent deficits in the control of extensor and flexor muscle groups involved in locomotion resulted in a motor syndrome that is detailed. After interruption of NMC fibers alone, no postural or locomotor deficits occurred and effects on lordosis were limited to the first few postoperative days. The NMC data suggested that the majority of the functional deficits seen after NGC plus NMC lesions were due to destruction of the NGC descending projections. Interruption of NSC bulbospinal fibers resulted in a significant depletion of spinal cord noradrenaline (NA) but not serotonin (5-HT). Functional deficits, when present, were limited to the first few postoperative days. Although interruption of descending NRM fibers resulted in a significant depletion of spinal cord 5-HT concentrations, greater depletions occurred in the NGC plus NMC group. No functional deficits were observed in the NRM group. These data suggest that interruption of NGC-NMC fibers destroyed the descending output from neurons which integrate the hormonal and sensory information necessary for lordosis.     

Stimulation of the mesencephalic central gray increases blood prolactin in ovariectomized rats

Systemic or intraventricular administration of opiate compounds induces not only analgesia but also the reciprocal responses of increased prolactin and decreased luteinizing hormone (LH) release. Electrical stimulation of the ventral periaqueductal gray (VPAG) or dorsal raphe nucleus (DRN) also inhibits both pain perception and LH release. In the present experiments prolactin release in response to stimulation of the VPAG-DRN was studied in ovariectomized rats. Stimulation was delivered through chronically implanted bipolar electrodes for 2 h. Blood samples collected at 10-min intervals through an indwelling jugular cannula were assayed for prolactin by radioimmunoassay. Increased release of prolactin followed increased intensity of stimulation (0.5, 1.0, 2.0 mA) in animals pretreated (10 min before initial blood sample) with saline. Naltrexone (3.7 mg/kg, i.v.) pretreatment was ineffective in antagonizing the effect of stimulation. Response to stimulation of the posterior VPAG-DRN was significantly greater than that to stimulation of more anterior VPAG-DRN loci at a current of 1.5 mA but not at 2.0 mA. Stimulation with a current of 0.5 mA was ineffective in releasing prolactin when applied to the VPAG-DRN, but when delivered to the periaqueductal gray directly lateral to the cerebral aqueduct, marked aversive and prolactin responses resulted. Taken together these data indicate that the VPAG in the region of the DRN contains a specific system capable of stimulating prolactin release and that opiate receptors are not interposed between the mesencephalic sites of activation of this pathway and the hypothalamic-hypophyseal portal system.