role of the dorsal paragigantocellular reticular ... - Semantic Scholar

Neuroscience 152 (2008) 849 – 857

ROLE OF THE DORSAL PARAGIGANTOCELLULAR RETICULAR NUCLEUS IN PARADOXICAL (RAPID EYE MOVEMENT) SLEEP GENERATION: A COMBINED ELECTROPHYSIOLOGICAL AND ANATOMICAL STUDY IN THE RAT R. GOUTAGNY,a,b* P.-H. LUPPI,a,b D. SALVERT,a,b D. LAPRAY,a,b D. GERVASONIa,b AND P. FORTa,b*

Key words: unanesthetized head-restrained rat, electrophysiology, tract-tracing, locus coeruleus, rapid eye movements, monoamines.

a

CNRS, UMR 5167, Institut Fédératif des Neurosciences de Lyon (IFR19), Lyon, F-69372, France

b

Université Claude Bernard, Lyon I, Villeurbanne, F-69622, France

Paradoxical sleep (PS or rapid eye movement (REM)sleep) is a sleep stage characterized by a paradoxical association of a low amplitude fast electroencephalogram (EEG) with muscle atonia and REMs. Since its discovery in the 1950s (Aserinsky and Kleitman, 1953; Jouvet et al., 1959), numerous studies had focused on brain areas involved in PS onset and maintenance. However, the neuronal network responsible for PS regulation remains to be fully identified. According to the classical “reciprocal interaction” models (Hobson et al., 1975; Sakai, 1985), the inactivation of monoaminergic wake (W)-promoting neurons within the brainstem, essentially serotonergic neurons of the dorsal raphe nucleus (DRN) and noradrenergic neurons of the locus coeruleus (LC), is a requirement for PS onset. Indeed, it has long been known that the activity of LC neurons is closely related to behavioral states since they decrease their firing during slow wave sleep (SWS) and become virtually quiescent during PS (Aston-Jones and Bloom, 1981). GABAergic neurons within the ventrolateral preoptic nucleus (VLPO) likely contribute, through reciprocal inhibitory interactions, to the firing decrease of W-promoting neurons during SWS (Sherin et al., 1996; Gallopin et al., 2000; reviews in Saper et al., 2001; Fort et al., 2004). Regarding PS, recent congruent data demonstrated that the firing cessation is due to a PS-selective tonic GABAergic inhibition. On one hand, it has been shown using microdialysis that the amount of GABA is greatly increased in the LC during PS compared with SWS or waking (Nitz and Siegel, 1997). On the other hand, using extracellular single-unit recordings in head-restrained unanesthetized rats, Gervasoni et al. (1998) showed that iontophoretic application of bicuculline, a specific GABAA-receptor antagonist, during PS, restores a firing activity in LC neurons similar to that of waking. These data highly suggest that GABA is maximally released within the LC from inputs activated specifically during PS. To identify these inputs, we combined injections in the LC of cholera toxin b subunit (CTb) as a retrograde tracer with the immunodetection of Fos protein in control rats, rats selectively deprived of PS during 3 days, and rats allowed to recover from such deprivation to obtain a longlasting PS hypersomnia (Verret et al., 2005, 2006). We showed that after PS recovery, the largest number of

Abstract—It is well known that noradrenergic locus coeruleus neurons decrease their activity during slow wave sleep and are quiescent during paradoxical sleep. It was recently proposed that their inactivation during paradoxical sleep is due to a tonic GABAergic inhibition arising from neurons located into the dorsal paragigantocellular reticular nucleus (DPGi). However, the discharge profile of DPGi neurons across the sleep–waking cycle as well as their connections with brain areas involved in paradoxical sleep regulation remain to be described. Here we show, for the first time in the unanesthetized rat that the DPGi contained a subtype of neurons with a tonic and sustained firing activation specifically during paradoxical sleep (PS-on neurons). Noteworthy, their firing rate increase anticipated for few seconds the beginning of the paradoxical sleep bout. By using anterograde tract-tracing, we further showed that the DPGi, in addition to locus coeruleus, directly projected to other areas containing wake-promoting neurons such as the serotonergic neurons of the dorsal raphe nucleus and hypocretinergic neurons of the posterior hypothalamus. Finally, the DPGi sent efferents to the ventrolateral part of the periaqueductal gray matter known to contain paradoxical sleep-suppressing neurons. Taken together, our original results suggest that the PS-on neurons of the DPGi may have their major role in simultaneous inhibitory control over the wake-promoting neurons and the permissive ventrolateral part of the periaqueductal gray matter as a means of influencing vigilance states and especially PS generation. © 2008 IBRO. Published by Elsevier Ltd. All rights reserved. *Correspondence to: R. Goutagny, Douglas Institute Research Center, McGill University, 6875 Blvd Lasalle, Verdun, Montreal, QC, Canada H4H1R3; P. Fort, CNRS UMR 5167, Faculté de Médecine RTH Laënnec, 7 rue Guillaume Paradin, 69372 Lyon Cedex 08, France. E-mail address: [email protected] (R. Goutagny), [email protected] (P. Fort). Abbreviations: AS, asymmetry index; CTb, cholera toxin b subunit; DAB, 3,3=-diaminobenzidine-4HCl; DPGi, dorsal paragigantocellular reticular nucleus; DpMe, deep mesencephalic reticular nucleus; DRN, dorsal raphe nucleus; EEG, electroencephalogram; EMG, electromyogram; EOG, electrooculogram; Fg, Fluorogold; ISI, interspike interval; LC, locus coeruleus; PBST, 0.1 M PB, containing 0.9% NaCl and 0.3% Triton X-100; PBST-Az, 0.1 M PB, containing 0.9% NaCl, 0.3% Triton X-100, and 0.1% sodium azide; PHA-L, Phaseolus vulgaris leucoagglutinin; PS, paradoxical sleep; PSB, Pontamine Sky Blue; REM, rapid eye movement; SWS, slow wave sleep; TH, tyrosine hydroxylase; vlPAG, ventrolateral part of the periaqueductal gray matter; VLPO, ventrolateral preoptic nucleus; W, wake.

0306-4522/08$32.00⫹0.00 © 2008 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2007.12.014

849

850

R. Goutagny et al. / Neuroscience 152 (2008) 849 – 857

CTb/Fos double-labeled cells was found in the dorsal paragigantocellular reticular nucleus (DPGi), suggesting that this medullary nucleus contains GABAergic neurons responsible for the tonic inhibition of LC neurons during PS. Supporting this hypothesis, it has been previously shown in anesthetized rats that the electrical or chemical stimulation of the DPGi area induced a firing inhibition of LC neurons. Furthermore, this inhibition is GABAergic in nature since it is blocked by LC application of bicuculline (Ennis and Aston-Jones, 1989). Finally, the electrical stimulation of the DPGi area is followed by an increase in PS quantities that is blocked by concomitant LC application of picrotoxin (a specific GABAA-receptor antagonist, Kaur et al., 2001). Despite the large body of experimental data supporting the contribution of DPGi to the PS-specific inhibition of LC neurons, the presence in this nucleus of neurons activated specifically during PS (PS-on neurons) remains to be directly demonstrated in vivo. Furthermore, determining the anatomical place of the DPGi within the complex neuronal network involved in PS should put complementary insight regarding its contribution to the vigilance regulation. To fill this gap and to precisely determine the role of the DPGi in PS mechanisms, we performed first, extracellular recordings of DPGi neurons across the sleep–waking cycle by using the head-restrained rat model (Souliere et al., 2000) and second, an anterograde tract-tracing study with Phaseolus vulgaris leucoagglutinin (PHA-L) to identify the brain areas receiving direct inputs from the DPGi.

EXPERIMENTAL PROCEDURES Electrophysiology The head-restrained rat method. The procedure (fixation of the head-restraining system and chronic implantation for the polygraphic recordings) has been previously described in detail (Boissard et al., 2002; Gervasoni et al., 1998). All experiments were conducted in agreement with the Guide for the Care and Use of Laboratory Animals (NIH Publication 80-23; authorization no. 03505 of the French Ministry of Agriculture) and every effort was made to minimize the number of animals used and their suffering. Briefly, male Sprague–Dawley rats (280 –320 g, n⫽9; IFFA Credo, L’arbreslie, France) were anesthetized with chloral hydrate (400 mg/kg, i.p) and mounted conventionally in a stereotaxic frame (David Kopf, CA, USA). Five electrodes were fixed in the skull bilaterally above the frontal (Bregma ⫹4 mm AP and ⫾2 mm L) and parietal (Bregma ⫺3 mm AP and ⫾3 mm L) cortices, and unilaterally above occipital (Bregma ⫺9 mm AP and ⫺3 mm L) cortex to monitor the EEG. Two wire electrodes were inserted into the neck muscles and two electrodes were inserted behind each ocular globe to monitor electromyogram (EMG) and electrooculogram (EOG), respectively. The head-restraining system was then put in place. After recovery (2 days), the rats were habituated to the restraining and recording system for 8 –10 days. At the end of the training, they could stay calm for 5– 6 h daily sessions during which active waking (with movements, AW), quiet waking (without movement, W), SWS and PS were routinely observed. After the habituation process and before the first single-unit recording session, rats were anesthetized with chloral hydrate (320 mg/kg, i.p., additional doses as needed), and a 4 mm hole was drilled over the DPGi. Daily recording sessions were typically performed over a maximum of 7–10 days, each session lasting ⬃4 – 6 h. The brain

surface was cleaned under local lidocaine anesthesia at the beginning of each daily recording session. Recording and analysis of neuronal activity. Extracellular recordings of DPGi neurons were performed using single-barrel glass micropipettes (external tip diameter, 2–3 ␮m) filled with Pontamine Sky Blue (PSB, 2% in sodium acetate 0.5 M, pH 7.5) or PHA-L (2.5% in 0.01 M phosphate-buffered saline; Vector Laboratories, Burlingame, CA, USA). Electrode impedances measured at 10 Hz ranged between 7 and 15 M⍀. Filtered (AC, 0.3–10 kHz) and unfiltered (DC) electrode signals were amplified (P16; Grass Instruments, RI, USA) and fed to storage oscilloscopes (2211 Tektronix, OR, USA) and an audio monitor (AM8, Grass). Single-unit activity (signal-to-noise ratio of at least 3:1) was isolated with an amplitude spike discriminator (Neurolog Spike Trigger, Digitimer Ltd., UK), collected and stored on a personal computer via a Cambridge Electronic Design (Cambridge, UK) interface using the Spike 2 software, in parallel with analog-to-digital samplings of amplified (Alvar, Reega, Paris, France) polygraphic signals (EEG and EMG; sample rate, 500 and 250 Hz, respectively), the AC trace at a rate of 15 kHz and a video acquisition of the rat behavior. Neurons from the DPGi were identified 1) by their stereotaxic location (bregma, anteroposterior, ⫺11.4 to ⫺12 mm; lateral, 0.4 – 0.6 mm; ventral, 6.8 – 8 mm); 2) on line by comparison with the unit recordings in adjacent brain areas during the electrode penetration through the cerebellum (dorsally), the gigantocellular reticular nucleus (ventrally), medial vestibular and prepositus hypoglossi nuclei (dorso-laterally) and the hypoglossal motor nucleus (caudally to the DPGi); and 3) off line after injection of PSB or PHA-L for localization of the recording sites. The mean discharge rates of individual neurons during W, SWS and PS were calculated by averaging spike counts made for at least 30 s continuous recordings in one given vigilance state. The effects of behavioral states on the discharge of each class of recorded neurons were assessed with one-way analysis of variance followed by a Tukey’s multiple comparison test. A P value inferior to 0.05 was considered statistically significant. The discharge pattern of neurons was appreciated by first-order interspike interval (ISI) histograms (ISHs), displaying the distribution of intervals between consecutive spikes, that were built with 500 bins of 1 ms width. For each vigilance state, an asymmetry index (AS) was defined as the ratio of the mode (the most frequent ISI) to the mean ISI (the reciprocal of the mean firing rate). Thus, an AS near the unit reveals a relatively regular discharge pattern, whereas the more the index differs from the unit, the more irregular is the spike train. For PS-on neurons, the onset of the firing increase and decrease was defined as the time when the firing rate during three consecutive 1-s bins was at least 2 S.D. greater or smaller than the mean rate during 30 – 60 s before the state transition (corresponding to a 30 s window). Finally, to measure the spike duration, action potentials were averaged 20 –30 times. All data are expressed as mean⫾S.E.M. and the significance level for all statistical analyses was set at P⬍0.05.

Neuroanatomical experiments Surgery. PHA-L injections into the DPGi. Male Sprague– Dawley rats (280 –320 g, IFFA Credo; n⫽7) were anesthetized with chloral hydrate (400 mg/kg, i.p.) and mounted conventionally in a stereotaxic frame (David Kopf, Epinay-sur-Seine, France) with ear bars and a head holder. The bone was exposed and cleaned and a hole was made under the DPGi stereotaxic coordinates. A single-barrel glass micropipette (external tip diameter: 2–3 ␮m) filled with PHA-L (2.5%) was lowered into the brain with a piezoelectric micromanipulator. The localization of the DPGi was assessed by the recording of neuronal activity throughout the electrode penetration: 500 ␮m ventral to the last neurons recorded in the cerebellum, 500 ␮m dorsal to the first neurons recorded in the gigantocellular nucleus and 500 ␮m rostral to the hypoglossal


851

90 min at room temperature; and iii) an ABC-HRP solution (1:1000 in PBST; Elite kit, Vector Laboratories) for 90 min at room temperature. Finally, the sections were immersed for 15 min at room temperature in DAB solution without nickel. As a result, immunoreactive cell bodies were colored brown well contrasting with black coloration of PHA-L-immunopositive fibers and terminal-like structures. For the retrograde tract-tracing experiments, free-floating sections were incubated in: 1) a goat antiserum to Fg (1:5000 in PBST-Az; Euromedex, Souffelweyersheim, France) or CTb (1: 40,000 in PBST-Az; List Biological) over 3 days at 4 °C; 2) a biotinylated rabbit antigoat IgG (1:2000 in PBST; Vector Laboratories) for 90 min at room temperature; and 3) an ABC-HRP solution (1:1000 in PBST; Elite kit, Vector Laboratories) for 90 min at room temperature. Then the sections were immersed in a 0.05 M Tris–HCl buffer (pH 7.6) containing 0.025% DAB (Sigma), 0.003% H2O2 and 0.6% nickel ammonium sulfate for 10 min at room temperature. The histochemical reaction was stopped by extensive rinses in PBST-Az. Finally, the sections were mounted on gelatin-coated slides, dried, dehydrated, and coverslipped with Depex. To provide semiquantitative data, four rats with a PHA-L injection centered to the DPGi were selected. The analysis was performed on the same number of sections per structure containing anterogradely-labeled fibers (Paxinos and Watson, 1997).

nucleus. Once the electrode tip was in place within the DPGi, PHA-L was iontophoretically injected (⫹5 ␮A, 7 s on, 7 s off, during 30 min). Animals were allowed 10 days’ recovery from surgery in individual Plexiglas containers before being killed and subsequent immunohistochemical experiments. Retrograde tracer injections into the vlPAG. Male Sprague– Dawley rats (280 –320 g, n⫽3; IFFA Credo) were anesthetized with chloral hydrate (400 mg/kg, i.p) and mounted conventionally in a stereotaxic frame (David Kopf, CA, USA). The skull was placed at a 15° angle (nose tilted down) and an additional angle of 15° was applied to the micromanipulator (leading to a total angle of 30°) to avoid the transverse sinus overlying the ventrolateral part of the periaqueductal gray matter (vlPAG) during the subsequent electrode penetrations. CTb (0.5% in 0.1 M PB, pH 6; List Biological Laboratories, Campbell, CA, USA) or Fluorogold (Fg, 4% in saline) were iontophoretically injected into the vlPAG (⫹5 ␮A, 7 s on, 7 s off, during 30 min). Animals were allowed 3 days’ recovery from surgery in individual Plexiglas containers before being killed and subsequent neuroanatomical experiments. Immunohistochemistry. Under deep anesthesia, rats were perfused first with saline solution followed by a fixative solution composed of 4% paraformaldehyde. The brains were then stored at 4 °C for at least 2 days in 30% sucrose in 0.1 M PB. They were rapidly frozen with CO2 gas and 25-␮m-thick coronal sections were cut on a cryostat. The free-floating sections were rinsed and stored until use in 0.1 M PB, containing 0.9% NaCl, 0.3% Triton X-100 (PBST), and 0.1% sodium azide (PBST-Az). For PHA-L immunohistochemistry, free-floating sections were incubated in: 1) a goat antiserum to PHA-L (1:5000 in PBST-Az; Abcys SA, Paris, France) over 3 days at 4 °C; 2) a biotinylated rabbit antigoat IgG (1:2000 in PBST; Vector Laboratories) for 90 min at room temperature; and 3) an ABC-HRP solution (1:1000 in PBST; Elite kit, Vector Laboratories) for 90 min at room temperature. Then the sections were immersed in a 0.05 M Tris–HCl buffer (pH 7.6) containing 0.025% 3,3=-diaminobenzidine-4HCl (DAB; Sigma, France), 0.003% H2O2, and 0.6% nickel ammonium sulfate for 10 min at room temperature. The histochemical reaction was stopped by extensive rinses in PBST-Az. In some cases, PHA-L immunohistochemistry was combined with immunodetection of neurotransmitters involved in sleep–waking regulation. For this purpose, the free-floating sections pretreated for PHA-L were submitted to a second immunohistochemical process with sequential incubations in i) a rabbit antiserum to tyrosine hydroxylase (TH) (1:10,000 in PBST-Az, Institut Jacques Boy, France), 5-HT (1:5000 in PBST-Az; Oncogene, Cambridge, MA, USA) or hypocretin (1:10,000 in PBST-Az; Phoenix Pharmaceutical, St. Joseph, MO, USA) over 3 days at 4 °C; ii) a biotinylated goat antirabbit IgG (1:1000 in PBST; Vector Laboratories) for

RESULTS Electrophysiology Classification of neurons by their behavior. The localization of each recorded neuron was estimated postmortem by assessing on brainstem sections the location of PSB or PHA-L deposit loci, corresponding to the area where cells were recorded. In this context, a sample of 41 neurons was recorded in the DPGi (n⫽9 rats, three PHA-L and six PSB injection sites) during at least one complete sleep–waking cycle (with W, SWS and PS). Neurons recorded outside the DPGi, i.e. in the vestibular nucleus or the gigantocellular reticular nucleus ventral to the DPGi, were not considered further. Among the 41 neurons of the DPGi, 11 were classified as phasic since i) they did not exhibit a continuous discharge during at least one vigilance state and ii) they were characterized by an AS lower than 0.25 during all vigilance states (Table 1). All these neurons depicted a significant

Table 1. Electrophysiological characteristic of the three types of neurons recorded in the DPGi Mean firing rate (Hz) W Phasic neurons (n⫽11) Tonic state-indifferent neurons (n⫽12) Tonic PS-on neurons (n⫽18)

SWS

Asymmetry index PS

Spike duration (ms)

W

SWS

PS

8.03⫾1.9

6.6⫾1.6

17.7⫾4.5

0.14⫾0.01***/#

0.15⫾0.03***/#

0.16⫾0.02***/###

0.57⫾0.02***/###

27.98⫾9.42

28.11⫾8.75

32.28⫾10.12

0.95⫾0.03°°°

0.77⫾0.08°°°

0.70⫾0.07°

0.35⫾0.03

4.24⫾0.9

7.34⫾1.1

46.06⫾5.1$$$

0.47⫾0.09

0.40⫾0.05

0.52⫾0.05

0.32⫾0.04

$$$ P⬍0.001, One way repeated measure ANOVA followed by a Tukey’s multiple comparison test indicating an intra-group difference. *** P⬍0.001, Phasic neurons vs. tonic state-indifferent neurons. One way ANOVA followed by a Tukey’s multiple comparison test. ### P⬍0.001, Phasic neurons vs tonic PS-on neurons. One way ANOVA followed by a Tukey’s multiple comparison test. # P⬍0.05, Phasic neurons vs tonic PS-on neurons. One way ANOVA followed by a Tukey’s multiple comparison test. °°° P⬍0.001, Tonic state-indifferent neurons vs. tonic PS-on neurons. One way ANOVA followed by a Tukey’s multiple comparison test. ° P⬍0.05, Tonic state-indifferent neurons vs. tonic PS-on neurons. One way ANOVA followed by a Tukey’s multiple comparison test.

852


Fig. 1. Discharge of DPGi neurons across sleep–wake states. (A) Phasic neuron exhibited a discharge rate closely related to eye movements during PS as shown by the EOG recording. (B) Firing of a DPGi neuron not related to the sleep–waking cycle.

firing activity during periods with eye movements (positive correlation with the EOG amplitude, r⫽0.48 range from 0.12– 0.92, t⫽7.22, P⬍0.001) or whisker twitching during W and PS assessed by video recordings (Figs. 1A, 3C). This subtype of DPGi neurons was not further considered. The 30 remaining DPGi neurons were classified as tonic and segregated into two sub-groups according to their firing rate during the sleep–waking cycle. Neurons

from the former subgroup (n⫽12) were characterized by a tonic and regular discharge pattern but did not show any modification of their firing rate through the sleep–waking cycle (Figs. 1B, 3C, Table 1). The latter group (n⫽18) encompassed neurons exhibiting a significant tonic increase in their mean firing rate during PS compared with W and SWS (Fig. 2A). They were therefore classified as tonic PS-on neurons and were mainly founded in the caudal part

Fig. 2. Electrophysiological characteristics of PS-on neurons from the DPGi. (A) A PS-on neuron showing a great firing rate increase during PS vs. W and SWS. This neuron did not present phasic activation correlated to eye movements. (B) The majority of the PS-on neurons started firing before the onset of PS identified by the association of a desynchronized EEG with a muscular atonia (data are expressed as the normalized mean frequency⫾S.E.M.). (C) These PS-on neurons decreased its firing rate prior the end of PS bout and became virtually quiescent at the onset of W (data are expressed as the normalized mean frequency⫾S.E.M.).


853

Fig. 3. Electrophysiological characterization and localization of the three neuronal populations within the DPGi. (A) Typical examples of a tonic-PS-on neuron (red), a tonic state-independent neuron (blue) and a phasic neuron (green) average extracellular waveforms. (B) Distribution of spike duration of all recorded neurons. (C) Anatomical reconstruction of the localization of each neuron from the three categories recorded in the DPGi and used in the present cell database: Tonic PS-on neurons in red; tonic state-independent neurons in blue; phasic neurons in green. List of abbreviations: Accessory facial nucleus Acs7; A5 noradrenalin cells A5; dorsal cochlear nucleus, deep core DCDp; dorsal cochlear nucleus, molecular layer DCMo; dorsal cochlear nucleus, fusiform layer DCFu; dorsomedial spinal trigeminal nucleus DMSp5; gigantocellular reticular nucleus Gi; gigantocellular reticular nucleus, alpha part GiA; genu of the facial nucleus g7; granule cell layer of cochlear nuclei GrC; intermediate reticular nucleus, alpha part IRtA; lateral paragigantocellular nucleus LPGi; medial longitudinal fasciculus mlf; medial vestibular nucleus, magnocellular part MVeMC; medial vestibular nucleus, parvicellular part MVePC; matrix region of the medulla Mx; parvicellular reticular nucleus PCRt; perifacial zone P7; periolivary nucleus PO; prepositus hypoglossi Pr; raphe magnus nucleus RMg; raphe obscurus nucleus Rob; raphe pallidus nucleus RPa; nucleus of the solitary tract Sol; spinal vestibular nucleus SpVe; superior vestibular nucleus SuVe; ventral cochlear nucleus, anterior part VCA; ventral cochlear nucleus, posterior part VCP.

of the DPGi (Fig. 3C). They showed a non-significant increase in their mean firing rate during SWS compared with W (n⫽18, Table 1, Fig. 2A). In contrast, they exhibited a significantly faster mean discharge rate during PS vs. SWS or W (n⫽18, P⬍0.001) accompanied by a decrease in discharge irregularity (Table 1). Tonic PS-on neurons from the DPGi were recorded during transitions SWS/PS and PS/W. For the majority of these neurons (n⫽13/15), the firing rate increased gradually during SWS, anticipating for about 14.6⫾1.3 s the PS onset, determined as the time

point when EEG theta rhythm became predominant (Fig. 2B). Similarly, PS-on neurons began to decrease their firing rate 2.7⫾0.8 s before awaking, defined as the time of disappearance of the predominant theta rhythm (Fig. 2C). Finally, we determined whether the two subtypes of tonic recorded neurons could be differentiated by their spike characteristics (i.e. extracellular spike waveform and duration). Extracellular action potentials were usually recorded as a positive deflection followed by a negative one. According to a previous classification for brainstem neurons

854


Fig. 4. Photomicrographic illustrations of anterogradely-labeled fibers in different brain areas after a PHA-L injection centered on the DPGi (A). Numerous PHA-L-labeled fibers (colored in black) were found in close vicinity and sometimes apposing TH-immunoreactive neurons of the LC (colored in brown, B), 5-HT-immunoreactive neurons of the DRN (in brown, C), hypocretin-immunoreactive neurons of the posterior hypothalamus (in brown, D) and within the vlPAG (E). (G, H) Photomicrographs showing the high number of retrogradely-labeled neurons in the DPGi following a Fg injection into the vlPAG/DPMe (F). Scale bar⫽150 ␮m A; B, C and E: 25 ␮m, C and H: 20 ␮m and F and G: 100 ␮m. For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.

(Koyama et al., 1998), spikes with a positive deflection lasting more than 0.75 ms followed by long negative deflection were classified as broad spikes; spikes with other characteristics were classified as brief spikes. Here, both subtypes of tonic neurons depicted brief spikes, and could not be differentiated. In contrast, phasic neurons exhibited significantly longer spikes than the tonic neurons (P⬍0.001, Table 1, Fig. 3A–B). Neuroanatomy Among seven PHA-L injection sites obtained, four were used here since they showed no or a very limited tracer diffusion into adjacent areas (i.e. the prepositus hypoglossi nucleus, Fig. 4A) and centered on the caudal part of the DPGi, thus matching with the location of PS-on neurons. In all injection cases, a consistent and reproducible tracing was observed. Anterogradely-labeled fibers were seen bilaterally with a highly ipsilateral predominance to the injection locus (Table 2). At the ponto-medullary levels, extensive PHA-L-labeled varicose fibers and punctate terminallike structures were present throughout the gigantocellular nucleus, abducens nucleus, caudal part of the pontine reticular formation, vestibular nuclei, sensory part of the V nucleus, Kolliker-Füse nucleus and parabrachial nucleus. A moderate number of PHA-L-immunoreactive fibers and terminal-like structures were observed in lateral paragigantocellular and facial nuclei. Scattered labeled fibers were detected within the inferior olivary complex and the lateral part of the pontine reticular nucleus. As expected, a moderate number of varicose fibers emanating from the DPGi were labeled in the LC. PHA-L/TH double-immunostaining further indicated that anterogradely-labeled fibers were distributed within the core of the LC, often in close vicinity of the cell bodies of TH-immunopositive neurons (Fig. 4B). At the mesencephalic level, a substantial number of varicose fibers were localized in the laterodorsal tegmental and DRN. In this latter nucleus, PHA-L-labeled fibers were localized closely to 5-HT-immunoreactive cells (Fig. 4C).

Fibers originating from the DPGi were also found in the vlPAG and the adjacent deep mesencephalic reticular nucleus (DpMe; Fig. 4E). To confirm this projection, we targeted Fg or CTb injections to the vlPAG/DpMe areas (Fig. 4F, n⫽2 CTb and n⫽2 Fg injection sites). As expected, a large number of neurons retrogradely-labeled from the vlPAG were observed in the DPGi (Fig. 4G and H). More rostrally, dense and tightly confined plexus of PHA-L-labeled terminals and fibers was found in the Darkschewitsh nucleus, interstitial nucleus of Cajal, ventral part of the periaqueductal gray matter, superior colliculus, deep mesencephalic nucleus and periventricular gray area. At the forebrain level, a moderate plexus of clustered terminal-like structures was labeled in the perifornical area of the posterior hypothalamus, closed and sometimes apposed to cell bodies of hypocretin-immunoreactive neurons (Fig. 4D). Finally, few PHA-L-labeled fibers were seen in the ventrolateral thalamic nuclei, zona incerta and medial septum/diagonal band complex area.

DISCUSSION We show for the first time that the DPGi contains neurons selectively activated during PS (tonic PS-on neurons). Moreover, using anterograde tract-tracing, we demonstrated that neurons of the DPGi simultaneously project to the main structures known to be involved in PS regulation, including the wake-promoting areas and the vlPAG. These original results highlight the role of the DPGi in PS regulation. Noradrenergic neurons of the LC are silent during PS, due to a tonic GABAergic inhibition (Gervasoni et al., 1998). A number of arguments, including the present data, suggest that the DPGi is the best candidate for this inhibition. First, the DPGi contained a large number of GADimmunoreactive cells projecting to the LC (Luppi et al., 1999) while the inhibition of LC neurons by electrical stimulation of the DPGi is abolished by bicuculline, a GABAA-

R. Goutagny et al. / Neuroscience 152 (2008) 849 – 857 Table 2. Efferents projections from the DPGi Structure

Ipsilateral labeling

Contralateral labeling

Path fibers

Medial septum Horizontal and diagonal band of Broca Septofimbrial nucleus Ventral part of the lateral preoptic area Basal nucleus of Meynert Amygdala complex Stria medullaris Zona incerta Perifornical area Thalamus paratenial Thalamus ventrolateral Posterior paraventricular thalamus Ventromedian hypothalamus Posterolateral hypothalamus Tuberomamillary nucleus Periventricular grey area Deep mesencephalic nucleus Superior colliculus vlPAG Oculomotor nucleus Interstitial nucleus of Cajal Darkschewitsch nucleus Dorsal raphe nucleus vlPAG Laterodorsal tegmental nucleus Pedunculopontine tegmental nucleus Lateral part of the pontine reticular nucleus Sublaterodorsal nucleus Medial and lateral part of the parabrachial nucleus Kolliker-Füse nucleus Sensory part of nV Locus coeruleus Vestibular nuclei Caudal part of the pontine reticular formation Abducens nucleus Noyau facial Gigantocellular nucleus Ventral gigantocellular nucleus Lateral paragigantocellular nucleus Inferior olivary complex

⫹ ⫹

⫺ ⫺

⫺ ⫺

⫺ ⫺

⫺ ⫺

⫺ ⫺

⫺

⫺

⫺

⫺ ⫹⫹ ⫺ ⫺ ⫹⫹ ⫹

⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹

⫺ ⫺ ⫹ ⫺ ⫺ ⫺ ⫺

⫹

⫺

⫺

⫹

⫺

⫺

⫺ ⫹⫹ ⫹⫹⫹

⫺ ⫹ ⫹

⫺ ⫺ ⫹

⫹⫹⫹ ⫹⫹⫹ ⫺ ⫹⫹⫹

⫹ ⫹⫹⫹ ⫺ ⫹

⫺ ⫺ ⫺ ⫺

⫹⫹⫹ ⫹⫹ ⫹ ⫹⫹

⫹ ⫹⫹ ⫹ ⫹

⫺ ⫺ ⫺ ⫺

⫹

⫺

⫺

⫹⫹

⫹⫹

⫹

⫹⫹⫹⫹ ⫹⫹⫹⫹

⫹⫹⫹ ⫹⫹

⫹ ⫺

⫹⫹⫹⫹ ⫹⫹⫹⫹ ⫹ ⫹⫹⫹⫹ ⫹⫹⫹⫹

⫹⫹ ⫹⫹⫹ ⫹ ⫹⫹⫹⫹ ⫹⫹⫹

⫺ ⫺ ⫺ ⫺ ⫺

⫹⫹⫹⫹ ⫹⫹⫹ ⫹⫹⫹⫹ ⫹⫹⫹

⫹⫹⫹⫹ ⫹⫹⫹ ⫹⫹⫹⫹ ⫹⫹⫹

⫺ ⫺ ⫺ ⫺

⫹⫹⫹

⫹⫹⫹⫹

⫺

⫹⫹

⫹⫹

⫺

Localization and relative number of anterogradely-labeled fibers (from sparse (⫺) to dense labeling (⫹⫹⫹⫹), n⫽4 rats) following PHA-L injection into the DPGi. In addition to varicosities, path fibers are indicated in the third column.

855

receptor antagonist (Ennis and Aston-Jones, 1989). Secondly, it has been shown that electrical stimulation of the DPGi increases PS, an effect reversed by the administration in the LC of the GABAA-receptor antagonist picrotoxin (Kaur et al., 2001). Third, by combining the injection of CTb in the LC with the immunohistochemistry of the protein Fos, in control rats, rats selectively deprived of PS and rats allowed to recover from such deprivation, Verret et al. (2006) recently showed that the DPGi contained numerous CTb/Fos double-labeled cells only in the PS recovery animals. These crucial data highly support the presence within the DPGi of neurons projecting to the LC and selectively activated during PS, therefore becoming suitable candidates for mediating the PS-specific inhibition of LC neurons. In perfect agreement with these functional data, we show for the first time in unanesthetized rats that a large contingent of neurons recorded within the DPGi presented an activity correlated to the vigilance states and was primarily strongly activated during PS. A fundamental characteristic of these newly identified brainstem PS-on neurons relates to their profile of activation during transitions between SWS to PS, a crucial period for mechanisms of PS generation. Indeed, we showed that the discharge rate increase of the PS-on neurons precedes by a few seconds the beginning of PS bouts. Therefore, as for other contingents of PS-on cells previously recorded in cat brainstem (Sakai et al., 1981; Sakai, 1985), the DPGi PS-on neurons satisfy the selectivity, tonicity and PS-latency criteria necessary for being considered PS-executive neurons. It is likely that the DPGi PS-on neurons are GABAergic in nature. These neurons present the narrow spike duration and it has been shown in the cortex that neurons with short duration action potentials are mainly GABAergic interneurons (Mountcastle et al., 1969; Simons, 1978; McCormick et al., 1985). Further, preliminary functional data obtained in our team by combining Fos immunohistochemistry with GAD in situ hybridization indicated that a large proportion of the DPGi neurons Fos-labeled in PS recovery rats are GABAergic in nature (Sapin et al., 2007). Albeit the GABAergic nature of the PS-on neurons recorded from the DPGi remains to be directly provided, our results strongly suggest that the PS-on neurons recorded in the DPGi may play a key role in the early phase of PS onset by simultaneously inhibiting the widely distributed wake-promoting neurons, including primarily LC noradrenergic neurons. To support this hypothesis, we showed that, in addition to the LC, the DPGi send efferent projections to other brain areas containing neurons involved in W promotion, such as the serotonergic neurons of the DRN and the hypocretinergic neurons of the posterior hypothalamus. These results are in agreements with previous retrograde tracing studies in rats showing direct projections from DPGi either to DRN or LC (Aston-Jones et al., 1986; Luppi et al., 1995; Gervasoni et al., 2000). Further, it has been recently evidenced by paired injections of retrograde tracers that neurons from the DPGi that send collateral projections to both DRN and LC nuclei (Lee et al., 2005a). Connections between DPGi and the posterior hypothalamus have also been reported in both anterograde and retrograde tract-

856


tracing works in rats (Ohtake, 1992; Abrahamson and Moore, 2001). These W-promoting neurons, as the LC neurons, exhibit in rats an exactly inverse pattern of discharge to that of PS-on neurons from DPGi, with a maximum of activity during W, decreasing during SWS to become inactive during PS (Aston-Jones and Bloom, 1981; Lee et al., 2005b; Lydic et al., 1987; Mileykovskiy et al., 2005). It is noteworthy that LC neurons stop firing 10 –20 s prior to PS onset (Gervasoni et al., 2000; Aston-Jones and Bloom, 1981). In the same time window, PS-on neurons from the DPGi start firing (14.6⫾1.3 s before PS onset). This close time relationship is consistent with a monosynaptic GABAergic inhibitory pathway from the DPGi to the LC during PS, as demonstrated in anesthetized rats (Ennis and Aston-Jones, 1989). Finally, the PS-off W-promoting neurons are inhibited by GABA during PS (Alam et al., 2005; Gervasoni et al., 1998, 2000; Goutagny et al., 2005). Based on all these congruent data, we propose that the cessation of activity of the W-promoting neurons during PS involves the presumably GABAergic PS-on neurons of the DPGi. Besides, the extended VLPO may also synergically contribute to these PS-specific inhibitory processes since it contains some presumed GABAergic neurons selectively activated during PS and sending projections to the Wpromoting neurons (Lu et al., 2002). Another major anatomical result obtained in this study is the significant projection from DPGi to vlPAG and the adjacent DpMe as shown by both anterograde and retrograde tract-tracing. This result is in agreement with a previous study showing a direct projection from the DPGi area to the vlPAG (Ohtake, 1992). The vlPAG region contains GABAergic neurons responsible for PS inhibition (Boissard et al., 2002, 2003; Lu et al., 2006). Moreover, these GABAergic neurons could be PS-off in nature, i.e. active during W and SWS and silent during PS as suggested by recent Fos experiments (Verret et al., 2005, 2006; Sapin et al., 2007). Finally, the lesion with neurotoxic ibotenic acid of vlPAG or its reversal inactivation with muscimol, a specific GABA-A receptor agonist, induced a strong increase in PS quantities in cats and rats (Lu et al., 2006; Sastre et al., 1996). Therefore, the inactivation of vlPAG, a required step for PS onset, may also originate from the PS-on neurons from the DPGi. Our results are in partial agreement with those of Lu et al. (2006), showing a “flip-flop switch” between two GABAergic populations, namely the vlPAG and the sublaterodorsal nucleus. Indeed, these authors concluded from their recent experiments that i) the lesion of either the LC or DRN did not produce any effect on the sleep–wake behavior and ii) an extensive lesion of the medial medullary reticular formation including the DPGi, did not reduce PS atonia. Several factors may explain the apparent discrepancy between these results and ours. First, it was not tested in the Lu et al. (2006) study whether the simultaneous inhibition of the LC, DRN and hypocretinergic neurons affected PS quantities. In addition, orexin-B–IgG– saporin was used as a neurotoxic drug to perform neuronal lesion, a method that may not produce complete lesion since some neurons may be devoid of orexin-B receptors.

Accordingly, it has been demonstrated that this orexin receptor subtype is heterogeneously located within the medulla oblongata (Marcus et al., 2001). It remains thus possible that PS-on neurons of the DPGi were spared from the neurotoxic-induced lesion if they do not express functional orexin-B receptors. Finally, we recorded in the DPGi phasic neurons with a discharge rate modulated by both postural and ocular movements during waking and PS. This is in line with a previous electrophysiological study in alert monkeys (McFarland and Fuchs, 1992) showing that the DPGi may contribute to the control of fine eye movements or processes of visuomotor coordination. Accordingly, the major inputs to the DPGi originated from neurons belonging to the saccadic system, i.e. the nucleus of Cajal, the Darkschewitsch and vestibular nuclei (Moschovakis et al., 1996, 1998). Here we further demonstrate that DPGi neurons send efferent projections back to these brain areas. It should be finally noted that the PS-on neurons within the DPGi do not present an activity related to the REMs characteristic of PS and thus do not seem be involved in their control.

CONCLUSION We highlighted here for the first time that the DPGi contains a population of PS-on neurons that may play a key role in the executive processes of PS generation, by inhibiting simultaneously the wake-promoting neurons and the vlPAG, two permissive systems for PS onset. Additional studies, such as the confirmation of the GABAergic phenotype of the PS-on neurons or the effects on PS of DPGi lesion, are necessary to test our hypothesis and to determine the mechanisms of activation of DPGi neurons required for PS generation. Acknowledgments—This work was supported by CNRS (FRE2469 and UMR5167) and Université C. Bernard Lyon I. R. Goutagny received a PhD grant from the “Région Rhône-Alpes” and the “Fondation pour la Recherche Médicale.” We thank S. Williams for his critical reading of the manuscript and R. Boissard and L. Leger for their help.

REFERENCES Abrahamson EE, Moore RY (2001) The posterior hypothalamic area: chemoarchitecture and afferent connections. Brain Res 889:1–22. Alam MN, Kumar S, Bashir T, Suntsova N, Methippara MM, Szymusiak R, McGinty D (2005) GABA-mediated control of hypocretinbut not melanin-concentrating hormone-immunoreactive neurones during sleep in rats. J Physiol 563:569 –582. Aserinsky E, Kleitman N (1953) Regularly occurring periods of eye motility and concomitant phenomena during sleep. Science 118: 273–274. Aston-Jones G, Bloom FE (1981) Activity of norepinephrine-containing locus coeruleus neurons in behaving rats anticipates fluctuations in the sleep-waking cycle. J Neurosci 1:876 – 886. Aston-Jones G, Ennis M, Pieribone VA, Nickell WT, Shipley MT (1986) The brain nucleus locus coeruleus: restricted afferent control of a broad efferent network. Science 234:734 –737. Boissard R, Fort P, Gervasoni D, Barbagli B, Luppi PH (2003) Localization of the GABAergic and non-GABAergic neurons projecting to the sublaterodorsal nucleus and potentially gating paradoxical sleep onset. Eur J Neurosci 18:1627–1639.

R. Goutagny et al. / Neuroscience 152 (2008) 849 – 857 Boissard R, Gervasoni D, Schmidt MH, Barbagli B, Fort P, Luppi PH (2002) The rat ponto-medullary network responsible for paradoxical sleep onset and maintenance: a combined microinjection and functional neuroanatomical study. Eur J Neurosci 16:1959 –1973. Ennis M, Aston-Jones G (1989) GABA-mediated inhibition of locus coeruleus from the dorsomedial rostral medulla. J Neurosci 9: 2973–2981. Fort P, Luppi P-H, Gallopin T (2004) In vitro identification of the presumed sleep-promoting neurons of the ventrolateral preoptic nucleus (VLPO). In: Sleep, circuits and functions (Luppi P-H, Simon SA, Nicolelis MAL, eds), pp 43–64. Frontiers in Neuroscience, USA: CRC Press. Gallopin T, Fort P, Eggermann E, Cauli B, Luppi PH, Rossier J, Audinat E, Muhlethaler M, Serafin M (2000) Identification of sleeppromoting neurons in vitro. Nature 404:992–995. Gervasoni D, Darracq L, Fort P, Souliere F, Chouvet G, Luppi PH (1998) Electrophysiological evidence that noradrenergic neurons of the rat locus coeruleus are tonically inhibited by GABA during sleep. Eur J Neurosci 10:964 –970. Gervasoni D, Peyron C, Rampon C, Barbagli B, Chouvet G, Urbain N, Fort P, Luppi PH (2000) Role and origin of the GABAergic innervation of dorsal raphe serotonergic neurons. J Neurosci 20:4217–4225. Goutagny R, Luppi PH, Salvert D, Gervasoni D, Fort P (2005) GABAergic control of hypothalamic melanin-concentrating hormone-containing neurons across the sleep-waking cycle. Neuroreport 16:1069 –1073. Hobson JA, McCarley RW, Wyzinski PW (1975) Sleep cycle oscillation: reciprocal discharge by two brainstem neuronal groups. Science 189:55–58. Jouvet M, Michel F, Courjon J (1959) [On a stage of rapid cerebral electrical activity in the course of physiological sleep]. C R Seances Soc Biol Fil 153:1024 –1028. Kaur S, Saxena RN, Mallick BN (2001) GABAergic neurons in prepositus hypoglossi regulate REM sleep by its action on locus coeruleus in freely moving rats. Synapse 42:141–150. Koyama Y, Honda T, Kusakabe M, Kayama Y, Sugiura Y (1998) In vivo electrophysiological distinction of histochemically-identified cholinergic neurons using extracellular recording and labelling in rat laterodorsal tegmental nucleus. Neuroscience 83:1105–1112. Lee MG, Hassani OK, Jones BE (2005b) Discharge of identified orexin/hypocretin neurons across the sleep-waking cycle. J Neurosci 25:6716 – 6720. Lee HS, Kim MA, Waterhouse BD (2005a) Retrograde double-labeling study of common afferent projections to the dorsal raphe and the nuclear core of the locus coeruleus in the rat. J Comp Neurol 481:179–193. Lu J, Bjorkum AA, Xu M, Gaus SE, Shiromani PJ, Saper CB (2002) Selective activation of the extended ventrolateral preoptic nucleus during rapid eye movement sleep. J Neurosci 22:4568 – 4576. Lu J, Sherman D, Devor M, Saper CB (2006) A putative flip-flop switch for control of REM sleep. Nature 441:589 –594. Luppi PH, Aston-Jones G, Akaoka H, Chouvet G, Jouvet M (1995) Afferent projections to the rat locus coeruleus demonstrated by retrograde and anterograde tracing with cholera-toxin B subunit and Phaseolus vulgaris leucoagglutinin. Neuroscience 65:119 –160. Luppi PH, Gervasoni D, Peyron C, Barbagli B, Boissard R, Fort P (1999) Norepinephrine and REM sleep. In: Rapid eye movement sleep (Mallick BN, Inoue S, eds), pp 107–122. New Delhi: Norosa Publishing House. Lydic R, McCarley RW, Hobson JA (1987) Serotonin neurons and sleep. I. Long term recordings of dorsal raphe discharge frequency and PGO waves. Arch Ital Biol 125:317–343. Marcus JN, Aschkenasi CJ, Lee CE, Chemelli RM, Saper CB, Yanagisawa M, Elmquist JK (2001) Differential expression of orexin receptors 1 and 2 in the rat brain. J Comp Neurol 435:6–25.

857

McCormick DA, Connors BW, Lighthall JW, Prince DA (1985) Comparative electrophysiology of pyramidal and sparsely spiny stellate neurons of the neocortex. J Neurophysiol 54:782– 806. McFarland JL, Fuchs AF (1992) Discharge patterns in nucleus prepositus hypoglossi and adjacent medial vestibular nucleus during horizontal eye movement in behaving macaques. J Neurophysiol 68(1):319 –332. Mileykovskiy BY, Kiyashchenko LI, Siegel JM (2005) Behavioral correlates of activity in identified hypocretin/orexin neurons. Neuron 46:787–798. Moschovakis AK, Kitama T, Dalezios Y, Petit J, Brandi AM, Grantyn AA (1998) An anatomical substrate for the spatiotemporal transformation. J Neurosci 18:10219 –10229. Moschovakis AK, Scudder CA, Highstein SM (1996) The microscopic anatomy and physiology of the mammalian saccadic system. Prog Neurobiol 50:133–254. Mountcastle VB, Talbot WH, Sakata H, Hyvarinen J (1969) Cortical neuronal mechanisms in flutter-fibration studied in unanesthetized monkeys. Neuronal periodicity and frequency discrimination. J Neurophysiol 32:452–484. Nitz D, Siegel JM (1997) GABA release in the locus coeruleus as a function of sleep/wake state. Neuroscience 78:795– 801. Ohtake T (1992) Ascending projections from the gigantocellular reticular and dorsal paragigantocellular nuclei of the medulla oblongata in the rat: an anterograde PHA-L tracing study. Neurosci Res 14:96 –116. Paxinos G, Watson C (1997) The rat brain in stereotaxic coordinates. Sydney: Academic Press. Sakai K, Sastre JP, Kanamori N, Jouvet M (1981) State-specific neurons in the ponto-medullary reticular formation with special reference to postural atonia during paradoxical sleep in the cat. In: Brain mechanisms and perceptual awareness (Pompeiano O, Ajmone Marsan C, eds), pp 405–429. New York: Raven Press. Sakai K (1985) Neurons responsible for paradoxical sleep. In: Sleep: neurotransmitters and neuromodulators (Wauquier A, Janssen Research Foundation, eds), pp 29–42. New York: Raven Press. Saper CB, Chou TC, Scammell TE (2001) The sleep switch: hypothalamic control of sleep and wakefulness. Trends Neurosci 24:726–731. Sapin E, Berod A, Hanriot L, Ravassard P, Salvert D, Fort P, Luppi PH, Leger L (2007) Localization of GABAergic neurons activated after paradoxical (REM) sleep deprivation and hypersomnia in rats. Society for Neuroscience annual meeting. Sastre JP, Buda C, Kitahama K, Jouvet M (1996) Importance of the ventrolateral region of the periaqueductal gray and adjacent tegmentum in the control of paradoxical sleep as studied by muscimol microinjections in the cat. Neuroscience 74:415– 426. Sherin JE, Shiromani PJ, McCarley RW, Saper CB (1996) Activation of ventrolateral preoptic neurons during sleep. Science 271:216–219. Simons DJ (1978) Response properties of vibrissa units in rat somatosensory neocortex. J Neurophysiol 50:798 – 820. Souliere F, Urbain N, Gervasoni D, Schmitt P, Guillemort C, Fort P, Renaud B, Luppi PH, Chouvet G (2000) Single-unit and polygraphic recordings associated with systemic or local pharmacology: a multipurpose stereotaxic approach for the awake, anaesthetic-free, and head-restrained rat. J Neurosci Res 61:88–100. Verret L, Fort P, Gervasoni D, Leger L, Luppi PH (2006) Localization of the neurons active during paradoxical (REM) sleep and projecting to the locus coeruleus noradrenergic neurons in the rat. J Comp Neurol 495:573–586. Verret L, Leger L, Fort P, Luppi PH (2005) Cholinergic and noncholinergic brainstem neurons expressing Fos after paradoxical (REM) sleep deprivation and recovery. Eur J Neurosci 21: 2488 –2504.

(Accepted 6 December 2007) (Available online 15 December 2007)