Ascending afferent regulation of rat midbrain dopamine neurons

Standard, extracellular single-unit recording techniques were used to examine the electrophysiological and pharmacological responsiveness of midbrain dopamine (DA) neurons to selected, ascending afferent inputs. Sciatic nerve stimulation-induced inhibition of nigrostriatal DA (NSDA) neurons was blocked by both PCPA (5-HT synthesis inhibitor) and 5,7-DHT (5-HT neurotoxin), suggesting mediation by a serotonergic (5-HT) system. Direct stimulation of the dorsal raphe (which utilizes 5-HT as a neurotransmitter and inhibits slowly firing NSDA neurons) inhibited all mesoaccumbens DA (MADA) neurons tested. Paradoxically, DPAT, a 5-HT_1a agonist which inhibits 5-HT cell firing, enhanced sciatic nerve stimulation-induced inhibition of NSDA neurons. MADA neurons were not inhibited by sciatic nerve stimulation and, therefore, could not be tested in this paradigm. In contrast to the dorsal raphe, electrical stimulation of the pedunculopontine tegmental nucleus preferentially excited slowly firing NSDA and MADA neurons. Thus, both excitatory and inhibitory ascending afferents influence the activity of midbrain DA neurons, and intact 5-HT systems are necessary for sciatic nerve stimulation to alter DA cell activity. However, the role that 5-HT plays in mediating peripheral sensory input remains unclear.     

Inhibition of spontaneous and evoked unit activity in the rat medial prefrontal cortex by mesencephalic raphe nuclei

The rat medial prefrontal cortex (PFC) receives a serotoninergic (5-HT) innervation which originates from the mesencephalic raphe nuclei. In the present study we determined the influence of the 5-HT ascending systems on the spontaneous and evoked activity of PFC neurons in anesthetized rats. Stimulation of the dorsal (DRN) and of the median raphe (MRN) nuclei inhibited the spontaneous activity of 35.0% and 52.8% of the PFC cells tested (mean duration of the inhibition: 75.5 and 82.2 ms, respectively). These inhibitory responses are likely mediated by the 5-HT-containing neurons since they were decreased markedly following selective destruction of ascending 5-HT pathways induced by local injections of 5,7-dihydroxytryptamine. Moreover, the inhibitory effect of MRN stimulation could be blocked by systemic administration of the 5-HT2 receptor antagonists: ketanserin and ritanserin. The effects of MRN stimulation on two types of evoked responses were studied. The excitatory responses of PFC neurons induced by the stimulation of the mediodorsal nucleus of the thalamus (MD) were inhibited by MRN stimulation applied before that of MD. Similarly, the activation of PFC cells induced by a noxious tail pinch was suppressed by a concomitant stimulation of the MRN. These results indicate that 5-HT neurons exert an inhibitory control on spontaneous or evoked activity in the rat PFC.     

Evidence for a serotonergically mediated sympathoexcitatory response to stimulation of medullary raphe nuclei

The cardiovascular role of serotonin (5-HT) containing neurons in the midline medullary raphe nuclei was studied in anesthetized cats. High frequency electrical stimulation of nucleus (n.) raphe (r.) pallidus, n.r. obscurus and n.r. magnus produced both pressor and depressor responses. Single shock stimulation of pressor sites produced an excitatory evoked potential of sympathetic nervous discharge (SND) recorded from the inferior cardiac nerve. Conversely, single shock stimulation of vasodepressor sites resulted in a computer-summed inhibition of SND. The mean conduction velocity in the sympathoexcitatory medullo-spinal pathway to sympathetic preganglionic neurons was calculated to be 1.24 m/s. The 5-HT antagonists methysergide and metergoline blocked the excitation of sympathetic activity evoked from medullary raphe nuclei. In contrast, these agents failed to alter the sympathoexcitatory response to electrical stimulation of lateral medulla pressor sites or the sympathoinhibitory response elicited by raphe stimulation. The 5-HT uptake inhibitor chlorimipramine increased the duration of the sympathoexcitatory response evoked from the raphe but not from the lateral medulla. Finally, mid-collicular transection did not effect the excitation of sympathetic activity elicited by stimulation of medullary raphe nuclei. These data suggest that serotonergic neurons in the midline medullary raphe nuclei provide an excitatory input to sympathetic neurons in the spinal cord.     

Evidence for a tonic GABAergic control of serotonin neurons in the median raphe nucleus

We examined whether serotonin (5-HT)-containing neurons of the midbrain raphe nuclei are subject to an inhibitory control by GABA. We found that injection into the median raphe nucleus of the GABA antagonists picrotoxin and bicuculline and the GABA agonist muscimol increase and decrease, respectively, the 5-HT turnover and the steady-state content of 5-hydroxyindoleacetic acid. The results provide biochemical evidence of a tonic inhibition by GABA of 5-HT neuronal activity in the median raphe nucleus; this inhibitory effect is potentiated by benzodiazepines.     

Inhibition of lateral vestibular nucleus neurons by 5-hydroxytryptamine derived from the dorsal raphe nucleus

Electrophysiological studies were performed to elucidate the effect of 5-hydroxytryptamine (5-HT) originating in the dorsal raphe nucleus (DR) on neuronal activity in the lateral vestibular nucleus (LVN) neurons, using cats anesthetized with alpha-chloralose. LVN neurons were classified into monosynaptic and polysynaptic neurons according to their responses to vestibular nerve stimulation. Conditioning stimuli applied to the DR inhibited orthodromic spikes elicited by vestibular nerve stimulation predominantly in polysynaptic neurons of the LVN. The iontophoretic application of 5-HT also inhibited orthodromic spikes of the LVN neurons. A close correlation was observed between the effects of DR conditioning stimulation and iontophoretically applied 5-HT in the same neurons. These inhibitions with both treatments were antagonized during the application of methysergide, a 5-HT antagonist. In the majority of LVN polysynaptic neurons that responded to antidromic stimulation of the ipsilateral or contralateral abducens nucleus, orthodromic spikes elicited by vestibular nerve stimulation were inhibited by DR conditioning stimulation and the iontophoretic application of 5-HT. In contrast, LVN neurons that responded to antidromic stimulation of the vestibulospinal tract were rarely affected by these treatments. These results indicate that 5-HT derived from the DR inhibits the synaptic transmission of LVN polysynaptic neurons ascending to the abducens nucleus, and suggest that 5-HT derived from the DR is involved in the regulation of the vestibulo-ocular reflex.     

Differential origin of brainstem serotoninergic projections to the midbrain periaqueductal gray and superior colliculus of the rat

Previous studies have shown that both the midbrain periaqueductal gray (PAG) and the superior colliculus receive a significant serotoninergic (5-HT) innervation. In the present study the origins of these 5-HT projections to the rodent PAG and superior colliculus were analyzed by using a combined immunohistochemical-retrograde transport technique. Thirteen brainstem regions were found to contain double-labelled 5-HT-like immunoreactive neurons following HRP injections into the PAG while only four brainstem nuclei contained double-labelled neurons following superior collicular injections. After HRP deposits into the ventral PAG, the largest percentage of double-labelled neurons was identified in nucleus raphe magnus, pars alpha of the nucleus gigantocellularis, and the paragigantocellular nucleus. The dorsal PAG, on the other hand, received the largest percentage of its 5-HT projections from nuclei raphe dorsalis, raphe obscurus, raphe pontis, and raphe medianis. The 5-HT input to the superior colliculus was found to arise exclusively from nuclei raphe dorsalis, raphe medianis, and raphe pontis and from the contralateral periaqueductal gray. Raphe nuclei were found to contribute serotoninergic projections to both the PAG and the superior colliculus while reticular nuclei contributed 5-HT projections only to the PAG. Injections of the fluorescent retrograde tracers true blue and nuclear yellow were then made into the PAG and superior colliculus to ascertain if neurons located in raphe nuclei that projected to both structures provided axon collaterals to both areas. Generally, less than 10% of raphe neurons projecting to the superior colliculus were identified as providing axon collaterals to the PAG. The present results demonstrate major quantitative and qualitative differences in the origin of 5-HT projections to the ventral PAG and superior colliculus. The origin of 5-HT input to the dorsal PAG, on the other hand, showed many similarities to the origin of 5-HT innervation of the superior colliculus. These data also indicate that approximately 35% of raphe neurons provide nonserotoninergic projections to the PAG and superior colliculus.     

Response of nucleus raphe magnus neurons to electrical stimulation of nucleus cuneiformis: Role of acetylcholine

There is considerable evidence that cells in the ventral medulla which includes nucleus raphe magnus (NRM) and nucleus magnocellularis are involved in a descending pain inhibitory system. Anatomical studies indicate a strong projection from nucleus cuneiformis (NCF) to the ventral medulla and histochemical studies suggest that many NCF neurons are cholinergic. Therefore, we investigated the effect of NCF stimulation on NRM unit activity and explored the possible role of acetylcholine (ACh) in this interaction. Of 180 NRM neurons examined, 43% were excited and 14% were inhibited by NCF stimulation. The average latency to the peak excitatory response was about 14 ms with a range of 5-32 ms. There was a tendency for the response latencies to cluster around 5 and 14 ms. Inhibitory responses were between 10 and 65 ms in duration. The anatomical specificity of the effective stimulation site was assessed by determining the response of a given NRM neuron to stimulation of areas dorsal and ventral as well as within NCF. The most reliable and intense responses of NRM neurons was observed with electrode placements within NCF. The most effective NCF region for activating NRM neurons corresponded to that region of NCF that contains a large population of neurons that project directly to NRM as seen in the present histochemical studies. The involvement of ACh in the interaction between NCF and NRM was examined with iontophoretic application of ACh and its antagonists. Of NRM neurons that responded to ACh, 79% were excited, an effect which was blocked by scopolamine.(ABSTRACT TRUNCATED AT 250 WORDS)     

Electrophysiological evidence for convergence of inputs from the medial prefrontal cortex and lateral habenula on single neurons in the dorsal raphe nucleus

Neuronal projections to the dorsal raphe nucleus (DRN) from the medial prefrontal cortex (mPFC) and lateral habenula nucleus (LHb) provide the two key routes by which information processed by mood regulatory, cortico-limbic-striatal circuits input into the 5-HT system. These two projections may converge as it appears that both activate local GABAergic neurons to inhibit 5-HT neurons in the DRN. Here we have tested this hypothesis by measuring the effect of stimulation of the mPFC and LHb on the activity of 5-HT and non-5-HT, putative gamma-amino butyric acid (GABA) neurons in the DRN using extracellular recordings in anaesthetized rats. A total of 119 5-HT neurons (regular, slow firing, broad spike width) and 21 non-5-HT, putative GABA neurons (fast-firing, narrow spike width) were tested. Electrical stimulation of the mPFC or LHb caused a poststimulus inhibition (30 ms latency) of 101/119 5-HT neurons, of which 61 (60%) were inhibited by both the mPFC and LHb. Electrical stimulation of the mPFC or LHb also caused a short latency (12-20 ms) poststimulus facilitation of 10/21 non-5-HT neurons, of which 5 (50%) were activated by both the mPFC and LHb. These data indicate that a significant number of 5-HT neurons and non-5-HT neurons in the DRN are influenced by both the mPFC and LHb. Moreover, the data are compatible with the hypothesis and that there is a convergence of mPFC and LHb inputs on local circuit GABAergic neurons in the DRN which in turn inhibit the activity of 5-HT neurons.     

Influence of inhibitory and excitatory inputs on serotonin efflux differs in the dorsal and median raphe nuclei

The dorsal (DRN) and median raphe nuclei (MRN) are two major sources of serotonergic projections to forebrain that are involved in regulation of behavioral state and motor activity, and implicated in affective disorders such as depression and schizophrenia. To investigate afferent influences on serotonergic neurons, this study compared the role of endogenous GABA and glutamate in the DRN and MRN using microdialysis and measurement of locomotor activity in freely behaving rats. Local infusion of the GABA(A) receptor antagonist bicuculline increased serotonin (5-HT) efflux in the DRN but not the MRN. In contrast, infusion of glutamate receptor antagonists produced larger decreases in 5-HT efflux in the MRN compared with the DRN. Moreover, glutamate receptor antagonists attenuated the increase in 5-HT efflux produced by GABA receptor blockade in the DRN. Thus, the disinhibitory effect of GABA blockers could be ascribed in part to an enhanced influence of glutamate. Measurements of locomotor activity indicate that changes in 5-HT were not simply correlated with behavioral activity induced by drug infusion. In summary, the role of inhibitory and excitatory afferents was strikingly different in the DRN and MRN. GABA afferents were the predominant tonic influence on serotonergic neurons in the DRN. In contrast, glutamatergic but not GABAergic afferents had a strong tonic influence on serotonergic neurons in the MRN.     

Variation in response to stimulation of central 5-hydroxytryptamine mechanisms in two strains of albino rat

The responses of neurons in the fronto-parietal cortex to iontophoretically applied 5-hydroxytryptamine (5-HT) and to stimulation of the nucleus raphe medianus were determined and compared in Wistar and Sprague-Dawley (SD) rats. In the Wistar strain there were many more excitatory responses to both 5-HT and raphe stimulation than in the SD rats. The duration of the excitatory responses to stimulation was significantly greater in the Wistars than in SD rats and the duration of inhibitory effects significantly less. These results may help to explain divergent findings in different laboratories concerning the action of 5-HT in the cerebral cortex. We were also able to demonstrate a difference in the behavioral consequences of central 5-HT receptor activation between the two strains. Following administration of the precursor, 5-hydroxytryptophan, Wistar rats showed a much greater incidence of ‘wet dog shake’ responses than SD rats, but the SD strain displayed a more severe tremor.     

Serotoninergic dorsal raphe neurons possess functional postsynaptic nicotinic acetylcholine receptors

Very few neurons in the telencephalon have been shown to express functional postsynaptic nicotinic acetylcholine receptors (nAChRs), among them, the noradrenergic and dopaminergic neurons. However, there is no evidence for postsynaptic nAChRs on serotonergic neurons. In this study, we asked if functional nAChRs are present in serotonergic (5-HT) and nonserotonergic (non-5-HT) neurons of the dorsal raphe nucleus (DRN). In rat midbrain slices, field stimulation at the tegmental pedunculopontine (PPT) nucleus evoked postsynaptic currents (eEPSCs) with different components in DRN neurons. After blocking the glutamatergic and GABAergic components, the remaining eEPSCs were blocked by mecamylamine and reduced by either the selective alpha7 nAChR antagonist methyllycaconitine (MLA) or the selective alpha4beta2 nAChR antagonist dihydro-beta-eritroidine (DHbetaE). Simultaneous addition of MLA and DHbetaE blocked all eEPSCs. Integrity of the PPT-DRN pathway was assessed by both anterograde biocytin tracing and antidromic stimulation from the DRN. Inward currents evoked by the direct application of acetylcholine (ACh), in the presence of atropine and tetrodotoxin, consisted of two kinetically different currents: one was blocked by MLA and the other by DHbetaE; in both 5-HT and non-5-HT DR neurons. Analysis of spontaneous (sEPSCs) and evoked (eEPSCs) synaptic events led to the conclusion that nAChRs were located at the postsynaptic membrane. The possible implications of these newly described nAChRs in various physiological processes and behavioral events, such as the wake-sleep cycle, are discussed.     

Possible projections from regions of paraventricular and supraoptic nuclei to the spinal cord: electrophysiological studies

The existence of monosynaptic connections between neurons in the paraventricular nucleus (PVN) of the hypothalamus and the intermediolateral cell column (ILC) of the spinal cord was studied by electrophysiological techniques in chloralose-anesthetized cats. Sympathetic preganglionic discharges (recorded from the 2nd or 3rd thoracic white ramus) were evoked by microstimulation of certain regions in or near the PVN with short train of pulses and below 50 microA current. By recording responses of 'identified' and 'non-identified' neurosecretory cells in the PVN and supraoptic nucleus (SON) to stimulation of the ILC of the thoracic cord, it was possible to identify antidromically evoked action potentials in 9 out of 297 neurons tested. Among them, 2 neurons were also antidromically excited by the pituitary stalk stimulation, 5 were orthodromically excited by the same stimulus and the remaining 2 were not excited by the stalk stimulation. Our results indicate that some PVN neurons, though small in number, send axons directly to the ILC of the cord, and that a very few neurons among these also send their axons to the pituitary gland.     

Estrogen enhances light-induced activation of dorsal raphe serotonergic neurons

The serotonergic system has been implicated in the modulation of physiological processes including circadian rhythms, learning, memory, mood and food intake. In females, cessation of ovarian function produces deleterious changes in all of these processes and estrogen treatment often ameliorates these conditions. Estrogen may produce these effects by acting on the midbrain raphe, an estrogen-sensitive region that receives direct projections from sensory systems. Here we examined the ability of estradiol to modulate neuronal responses of neurons within raphe nuclei to photic stimulation. Ovariectomized rats treated with estradiol or cholesterol were killed 1 h after the normal onset of light (Zeitgeber time 0) or after a 2-h phase advance (Zeitgeber time 22). In a second study, estradiol-treated ovariectomized rats under constant dark conditions were exposed to light 2 h before the subjective onset of circadian time [(CT)22] and killed 1 h later (CT23). The brains from all animals were processed for Fos and/or serotonin (5-HT) immunocytochemistry. Comparisons showed that the phase shift increased Fos immunoreactivity in all dorsal raphe nucleus (DRN) regions. Although estradiol did not alter the overall number of Fos-positive nuclei, it significantly increased the number of Fos/5-HT double-labelled cells in the medial and lateral DRN. In contrast, neither a phase shift nor estradiol altered the number of Fos-immunoreactive cells or the proportion of Fos-positive 5-HT cells in the median raphe nucleus. Results reveal that the DRN 5-HT system responds to changes in the light : dark cycle and that these responses are modulated by estrogen.     

Connections of hypothalamic paraventricular neurons with the dorsal medial thalamus and neurohypophysis: an electrophysiological study in the rat

Connections between the hypothalamic paraventricular nucleus (PVN) and thalmic paraventricular nucleus (PVT) were examined using electrophysiological methods. Efferent projections of adjacent PVN cells were defined on the basis of antidromic activation from either PVT (n= 12) or neurophyphoseal (n= 38) stimulation; antidromic activation from both sites (n= 3) suggested that some PVN cells project both to the PVT and to the neurohypophysis. PVT stimulation evoked only weak orthodromic responses from 21% of PVN neurohypophyseal neurons, whereas short latency, high probability orthodromic responses were observed from 43% of PVN non-neurosecretory neurons. These data indicate reciprocal PVN-PVT connections and suggest that PVT afferents preferentially innervate non-neurosecretory PVN cells.     

Dorsal raphe nuclear stimulation of SCN serotonin release and circadian phase-resetting

Serotonin (5-HT) is strongly implicated in the regulation of mammalian circadian rhythms. However, little is known of the functional relationship between the circadian clock located in the suprachiasmatic nucleus (SCN) and its source of serotonergic innervation, the midbrain raphe nuclei. In previous studies, we reported that electrical stimulation of the dorsal or median raphe nuclei (DRN and MRN, respectively) induced 5-HT release in the SCN. Notably, DRN- but not MRN-stimulated 5-HT release was blocked by the 5-HT(1,2,7) antagonist, metergoline, suggesting that the DRN signals to the SCN indirectly via the activation of a 5-HT-responsive multisynaptic pathway. In the present study, pretreatment with the 5-HT(2,7) antagonist, ritanserin, also significantly inhibited DRN-electrically stimulated SCN 5-HT release. However, pretreatment with the 5-HT(1A) antagonist, NAN-190, or the 5-HT(2) antagonists ketanserin and cinanserin had little suppressive effect on this DRN-stimulated 5-HT release. In complementary behavioral trials, electrical stimulation of the DRN during subjective midday caused a 1.3-h advance in the free-running circadian activity rhythm under constant darkness, which was inhibited by metergoline. Collectively, these results are evidence that: (1) DRN-stimulated 5-HT release in the SCN requires the activation of an intermediate target with receptors having 5-HT(7) pharmacological characteristics; (2) electrical stimulation of the DRN induces phase-resetting of the circadian activity rhythm; and (3) activation of 5-HT receptors is necessary for this DRN-stimulated circadian phase-resetting. In view of the dynamic changes in DRN neuronal activity incumbent with the daily sleep-activity cycle, and its functional linkages to the SCN and intergeniculate leaflet, the DRN could serve to provide behavioral/arousal state information to various sites comprising the brain circadian system.     

Serotonin- and substance P-containing projections to the nucleus tractus solitarii of the rat

The objective of the present study was to determine the location of the neurons that give rise to serotonin- and substance P-containing terminals in the nucleus tractus solitarii. This was done by injecting rhodamine-filled latex microspheres into the nucleus tractus solitarii of rats to retrogradely label neuronal cell bodies and by processing sections from the brains of these animals to determine whether the labelled neurons contained serotonin or substance P immunoreactivity. Serotonin-immunoreactive neurons that projected to the nucleus tractus solitarii were found in the nucleus raphe magnus, nucleus raphe obscurus, nucleus raphe pallidus, and in the ventral medulla, lateral to the pyramidal tract. Substance P-immunoreactive neurons that projected to the nucleus tractus solitarii were found in similar areas but were proportionately less numerous in the nucleus raphe magnus and proportionately more numerous in the nucleus raphe pallidus. It is concluded that neurons in the medullary raphe nuclei, some of which presumably utilize serotonin or substance P as a neurotransmitter, could regulate autonomic function via direct projections to the nucleus tractus solitarii.     

Raphe and periaqueductal gray induced suppression of non-nociceptive neuronal responses in the dorsal column nuclei and trigeminal sub-nucleus caudalis

Electrical stimulation of the feline periaqueductal gray matter and nucleus raphe magnus was found to inhibit the firing of trigeminal sub-nucleus caualis nocipeptive neurons, as has been previously reported. However, stimulation at these same sites using similar current intensities was also found to be equally effective in inhibiting the responses of non-nociceptive neurons in both nucleus caudalis and in the dorsal column nuclei.     

Chloral hydrate anesthesia blocks the excitatory response of dorsal raphe neurons to phasic auditory and visual stimuli in cats

Serotonergic neurons within the dorsal raphe nucleus were driven, with similar temporal characteristics, by phasic auditory and visual stimuli in freely moving cats. Chloral hydrate (300–350 mg/kg, i.p) produced little change in the spontaneous activity of raphe neurons (−18.7%), but completely blocked the responsiveness of these units to sensory stimulation. Based on these and other data, we hypothesized that the anesthetic inactivates a general excitatory input to the dorsal raphe nucleus.     

Excitation of supraoptic vasopressin cells by stimulation of the A1 noradrenaline cell group: failure to demonstrate role for established adrenergic or amino acid receptors

The effects of adrenergic and excitatory amino acid antagonists on supraoptic nucleus (SON) neurosecretory cell responses to stimulation of the A1 noradrenaline (NA) cell group were examined in anaesthetized male rats. As in previous studies, delivery of cathodal pulses (100 microA, 1 ms pulses, 1 Hz) to the A1 region of the caudal ventrolateral medulla excited spontaneously active, antidromically identified neurosecretory cells, the majority of which were identified as arginine vasopressin (AVP) secreting on the basis of basal discharge patterns and responses to abrupt increases in arterial blood pressure. Administration of alpha- and beta-adrenoreceptor antagonists, by systemic or intracerebroventricular delivery of a bolus, or by direct pressure injection into the SON, did not alter neurosecretory cell responses to A1 stimulation, even when doses applied exceeded that required for blockade of excitations elicited by local application of NA. Application of the broad spectrum excitatory amino acid antagonist kynurenic acid (5-40 mM) blocked the excitatory effects of locally applied glutamate (100 microM) and transiently inhibited spontaneous activity, but failed to alter the excitatory effects of A1 region stimulation on SON cells. Identical effects were obtained with a selective kainate/quisqualate receptor antagonist. These data indicate that neurosecretory cell responses to activation of the A1 cell group are unaltered by antagonists of alpha- and beta-adrenoreceptors, or excitatory amino acid receptors.(ABSTRACT TRUNCATED AT 250 WORDS)     

Extent of colocalization of serotonin and GABA in the neurons of the rat raphe nuclei

Previous investigations of the distribution of neurons containing both serotonin and GABA in the brainstem raphe nuclei have yielded discrepant results amongst different authors. This study attempted to clarify the distribution as well as the proportions of raphe and other brainstem neurons that contain both neurotransmitters. All the nine serotonergic cell groups known to be present in the brainstem were examined with an indirect immunofluorescence method using antibodies against serotonin and glutamic acid decarboxylase in colchicine-treated rats. Sections were incubated either simultaneously or sequentially for the two immunolabels. Brainstem neurons that were labelled for both markers were generally infrequent. Of all the serotonin cell groups in the brainstem, the nucleus raphe magnus contained the most double-labelled cells (a mean of 3.6% of a total of 625–1155 serotonin-immunoreactive cells counted in this nucleus), followed by the nucleus raphe obscurus (1.5% of a total of 220–550 serotonin-immunoreactive neurons counted). The dorsal, median and pontine raphe nuclei as well as the supralemniscal nucleus (the B9 group) contained very few double-labelled cells, which comprised a mean of 0.1–0.7% of all serotonin-immunoreactive cells in each of these nuclei. No double labelled cells were present in the caudal linear raphe nucleus or the nucleus raphe pallidus, nor in the B4 group. These results suggest that only a very small percentage of serotonergic neurons in the medullary raphe nuclei (raphe magnus and raphe obscurus) also contain GABA, whereas such cells are virtually absent in the midbrain raphe nuclei or in the non-raphe serotonergic cell groups in the brainstem.