sleep deprivation increases cerebral serotonin 2a … deprivation... · sleep, vol. 35, no. 12,...

SLEEP, Vol. 35, No. 12, 2012 1615 Sleep Deprivation Increases 5-HT2A Density—Elmenhorst et al

INTRODUCTIONSeveral neurotransmitter systems are involved in the regu-

lation of the sleep-wake cycle. Prolonged wakefulness causes alterations of neurotransmitter levels that are most likely ac-companied by changes in respective receptor densities. In hu-mans only a few receptor types have been investigated in this regard,1 including adenosine A1 receptors2 and dopamine D2/D3 receptors.3 So far, there are no data available for receptors of the serotonergic system although they are decisive contributors to circadian rhythmicity and important targets for drug treat-ment of depression. Moreover, depressive symptoms can be improved by sleep deprivation.

A widely accepted theory of sleep-wake regulation is the two-process model that postulates an interplay of a specific phase of circadian rhythm and a homeostatic sleep debt that sum up the sleep propensity. Sleep deprivation of 24 h primarily affects the homeostatic regulator, which allows investigation of the driving molecular basis of this process.

The serotonergic raphe nuclei with their widespread corti-cal projections are part of the monoaminergic wake promoting system. Accordingly, cortical serotonin levels are high during wakefulness, reduced during slow wave sleep (SWS), and virtu-

SLEEP DEPRIVATION INCREASES CEREBRAL SEROTONIN 2A RECEPTOR BINDING IN HUMANShttp://dx.doi.org/10.5665/sleep.2230

Sleep Deprivation Increases Cerebral Serotonin 2A Receptor Binding in HumansDavid Elmenhorst, MD1; Tina Kroll, MD1; Andreas Matusch, MD1; Andreas Bauer, MD1,2

1Institute of Neuroscience and Medicine, INM-2, Forschungszentrum Jülich, Jülich, Germany; 2Department of Neurology, Medical Faculty, Heinrich-Heine-University Düsseldorf, Düsseldorf, Germany

Submitted for publication November, 2011Submitted in final revised form June, 2012Accepted for publication June, 2012Address correspondence to: David Elmenhorst, MD, Institute of Neuro-science and Medicine (INM-2), Forschungszentrum Jülich, 52425 Jülich, Germany; Tel: +49-(0)2461-616113; Fax: +49-(0)2461-612820; E-mail: [email protected]

Study Objectives: Serotonin and its cerebral receptors play an important role in sleep-wake regulation. The aim of the current study is to inves-tigate the effect of 24-h total sleep deprivation on the apparent serotonin 2A receptor (5-HT2AR) binding capacity in the human brain to test the hypothesis that sleep deprivation induces global molecular alterations in the cortical serotonergic receptor system.Design: Volunteers were tested twice with the subtype-selective radiotracer [18F]altanserin and positron emission tomography (PET) for imaging of 5-HT2ARs at baseline and after 24 h of sleep deprivation. [18F]Altanserin binding potentials were analyzed in 13 neocortical regions of interest. The efficacy of sleep deprivation was assessed by questionnaires, waking electroencephalography, and cognitive performance measurements.Setting: Sleep laboratory and neuroimaging center.Patients or Participants: Eighteen healthy volunteers.Interventions: Sleep deprivation.Measurements and Results: A total of 24 hours of sleep deprivation led to a 9.6% increase of [18F]altanserin binding on neocortical 5-HT2A recep-tors. Significant region-specific increases were found in the medial inferior frontal gyrus, insula, and anterior cingulate, parietal, sensomotoric, and ventrolateral prefrontal cortices.Conclusions: This study demonstrates that a single night of total sleep deprivation causes significant increases of 5-HT2AR binding potentials in a variety of cortical regions although the increase declines as sleep deprivation continued. It provides in vivo evidence that total sleep deprivation induces adaptive processes in the serotonergic system of the human brain.Keywords: [18F]altanserin, human, positron emission tomography, serotonin 2A receptor, sleep deprivationCitation: Elmenhorst D; Kroll T; Matusch A; Bauer A. Sleep Deprivation Increases Cerebral Serotonin 2A Receptor Binding in Humans. SLEEP 2012;35(12):1615-1623.

ally quiescent during rapid eye movement sleep.4,5 During sleep deprivation the serotonin release is even higher than during the previous wake period, as animal findings suggest. Elevated se-rotonin levels have been measured in the hippocampus of sleep-deprived rats6 and even during the subsequent recovery period.7 Higher serotonin levels have also been found in dissected dor-sal raphe and suprachiasmatic nuclei of sleep-deprived rats.8

Serotonin acts via 14 different receptors9 including the 5-HT1 and 5-HT2 subtypes, which are important for the sleep-regulat-ing actions of serotonin. 5-HT2AR agonists (such as mescaline and psylocybin) enhance wakefulness and reduce SWS, where-as 5-HT2AR antagonists (such as eplivanserin and volianserin) increase SWS and were evaluated as hypnotic agents in phase III clinical trials.5 Blockade of the 5-HT2AR by eplivanserin mimicked certain aspects of the effect of sleep deprivation on recovery sleep.10 Additionally, there is no increase in electroen-cephalography (EEG) delta power density (an indicator of pre-vious sleep debt) after sleep deprivation in 5-HT2AR knockout mice compared with their wild-type littermates.11 We therefore used the radiofluorinated form of altanserin, a high affinity and selective 5-HT2AR antagonist for imaging of human cerebral 5-HT2ARs in vivo with position emission tomography (PET).12,13 Its low test-retest variability and high reliability14,15 are impor-tant prerequisites for the comparative study design used in the current study.

The aim of the current study was to investigate the effect of 24-h total sleep deprivation on 5-HT2AR availability in the hu-man brain to test the hypothesis that sleep deprivation induces global molecular alterations in the cortical serotonergic recep-tor system. For this purpose, volunteers were examined with [18F]altanserin PET, wake EEG, and the psychomotor vigilance


task (PVT) on two subsequent days – after normal sleep and after 24 h of sleep deprivation

MATERIALS AND METHODS

Subjects and Study DesignThe study was approved by the Ethics Committee of the

Medical Faculty of the University of Düsseldorf, Germany, the German Federal Institute for Drugs and Medical Devices, and the German Federal Office for Radiation Protection.

Eighteen healthy volunteers were recruited via advertise-ments and were reimbursed for participation. In one volunteer the metabolite analysis failed on the first examination day. In another volunteer highly differing free fractions of radioligand between the two scans were detected during image analysis. Two volunteers did not refrain from caffeine intake before the study, a finding that was revealed retrospectively by plasma caf-feine levels from 2-3 mg/L. Although there is no evidence that caffeine influences [18F]altanserin binding or 5-HT2ARs densi-ties, both volunteers were excluded from analysis to rule out possible indirect influences. Consequently, a total of 14 vol-unteers were included (eight males and six females, mean age 47.4 ± 5.2 y, range 40 – 55 yr).

After giving written informed consent, all volunteers were screened for the following exclusion criteria: history of neuro-logic or psychiatric diseases, sleep disorders, shift and/or night work, head injury, and alcohol and/or substance abuse.

Some volunteers were taking medication: levothyroxine (n = 3), iodide (n = 1), valsartan + allopurinol (n = 1), and metformin (n = 2). Medication was kept constant during the experiment. All volunteers were nonsmokers and reported to be moderate caffeine consumers with intake of 3.1 ± 1.9 cups (0.15 L) of coffee (or caffeine equivalent) per day. Caffeine intake was not allowed for at least 36 h before the first PET scan. Compliance was confirmed by determination of caf-feine plasma levels assessed by liquid chromatography/mass spectrometry. The average caffeine plasma concentration was 0.39 ± 0.4 mg/L.

All volunteers underwent two [18F]altanserin PET scans at the same time of the day on two consecutive days under iden-tical conditions; the first scan after a night with normal sleep and the second one after a night of sustained wakefulness. Volunteers were investigated in groups of two per day with a time lag of 1.5 h between measurements. Time of injection of the radioligand was between 09:33 and 13:33 on both days. Staff members monitored sleep-deprived volunteers during the entire night between the two PET scans to prevent them from falling asleep. During the 60-min period of image acqui-sition, volunteers were requested to keep their eyes open. In addition, a video system was installed to observe the volun-teers in the scanner. As soon as they closed their eyes longer than usual, volunteers were addressed and supported to stay awake. Volunteers reported sleep duration in a sleep diary for one week before the first PET scan. Volunteers were asked to sleep for at least 8 h in a regular schedule during these days. In addition, duration of sleep in the last two nights before the study was controlled by actigraphy (SomnoWatch, SOMNO-medics GmbH, Randersacker, Germany). The wristwatch-sized actigraph was worn on the nondominant wrist. Ambient

light level and movements were recorded during 48 h before the first PET scan.

Waking EEGAt least one hour before PET scanning, volunteers were

placed in a quiet environment and supplied with electrodes in the C3, C4, A1, and A2 positions for EEG and above and be-low the lateral angulus oculi for electrooculography. Volunteers rested on a bed in a slightly elevated position for approximately 20 min before a waking EEG was performed, which consisted of 3 min with eyes closed and 5 min with eyes open. Patients were instructed to relax and to fixate on a spot on the ceiling in 2.5-m distance during the eyes-open period. An observer monitored the session continuously and addressed the patients when signs of drowsiness were detected (e.g., rolling eye movements or a reduction in alpha activity). Signals were recorded with a poly-graphic amplifier (Somnoscreen PSG EEG-10-20, Somnomed-ics, Germany), conditioned by a high-pass filter (0.3 Hz) and a low-pass filter (30 Hz) and digitized (128 Hz, 16 bit). The Dom-ino software (version 2.1, Somnomedics) was used to export raw signals in the EDF+ data format. EDF+ data were then imported into MATLAB (The Math Works Inc., Natick, MA, USA) with routines from the BIOSIG Toolbox (A. Schlögl, BIOSIG:a free and open source software library for biomedical signal process-ing, 2003 – 2010, available online: http://BIOSIG.SF.NET) and further analyzed by using tools from the signal processing tool-box of MATLAB. For each 2-sec epoch, power spectra were cal-culated by fast Fourier transform routine. Data were previously conditioned by applying a Hanning window and linear detrend-ing. For the 5-min periods with eyes open, activity was averaged from each artifact-free 2-sec epoch as follows: alpha (power 8-12 Hz), lower alpha (power 7-9 Hz16), theta (power 5-8 Hz), and delta (power 0.5-5 Hz). Artifacts were identified by an automated blink detection algorithm and subsequent visual inspection. Au-tomatic artifact detection was performed by squaring the electro-oculography signal and excluding every epoch containing a data point that was more than four times higher than the mean of the squared signal. Relative difference in theta and delta activity of the average of the C3/A2 and C4/A1 derivations were expressed as (sleep deprivation – baseline) / baseline activity. One partici-pant with an extraordinarily high eye blink frequency was ex-cluded from EEG analyses because more than 90% of the epochs of both days had to be discarded due to resulting artifacts.

Cognitive Performance MeasurementsParticipants conducted a 10-min reaction-time task (PVT)17

immediately before PET scanning. In the morning of the first test day participants were trained once in conducting the task. They were asked to perform tests as fast and correct as possible.

Participants had to respond to a LED signal lighting up in ir-regular intervals (1.5-10 s) on a handheld pocket personal com-puter by pressing a key as quickly as possible. After each trial the reaction time (in ms) was displayed. The number of present-ed signals depended on the reaction times of each participant. In this sample the number of presented signals averaged 69.7 ± 10.8 per 10-min trial. Reaction times equal to or longer than 500 ms were regarded as lapses and excluded from analysis. Furthermore, reaction times that were shorter than 130 ms were most probably reactions without stimulus (false starts) and


therefore also excluded from analysis. Reaction speed (equals: 1/reaction time) for each reaction was assessed and 10% fastest responses were averaged.

Before and after each PET scan as well as during the nights, participants’ sleepiness was screened with the Stanford Sleepi-ness Scale (SSS) and a visual analog scale (VAS). The SSS is a 7-point Likert scale ranging from very awake to nearly asleep. The VAS ranged from 0 to 1 (very sleepy to very awake). Me-dian SSS and VAS values obtained within the last 15 min before the start of each scan were used for further analysis.

PET Acquisition[18F]Altanserin PET was performed as previously de-

scribed12,18 using a bolus/infusion schedule. Scanning took place with participants in the supine position and a quiet ambi-ence. Participants’ heads were immobilized in the canthomeatal orientation by a vacuum cushion. PET data were acquired in three-dimensional mode on a Siemens ECAT Exact HR+ scan-ner (Siemens-CTI, Knoxville, TN, USA) equipped with a cir-cular lead shield to reduce scatter radiation from outside the field of view. [18F]Altanserin was administered as a combina-tion of bolus (2 min) and constant infusion (178 min, bolus/infusion ratio of Kbol = 2.1 h) to rapidly approach an equilibrium of the radioligand in blood and brain. Emission was recorded between 120–180 min after injection in 10-min time frames. Head movements of the participants were recorded with a Po-laris optical tracking system (Northern Digital Inc, Waterloo, Ontario, Canada). Head positions were permanently monitored and, if necessary, manually corrected under guidance of a video system and reference skin marks. Data sets were fully corrected for random coincidences, scatter radiation and attenuation (10 min 68Ge/68Ga-transmission scan directly before emission scan), Fourier rebinned into two-dimensional sinograms, and reconstructed by filtered back projection (Shepp filter, 2.5 mm width) with a voxel size of 2 × 2 × 2.43 mm³ (63 slices).

Heparinized venous plasma samples (antecubital vein cath-eter) were taken at 2, 5, 10, 20, 30, 45, 60, and 120 min after injection and every 10 min during PET emission acquisition. Additional venous blood samples were taken before injection to assess the fraction of free [18F]altanserin in plasma (unbound to proteins, denoted by fP), the recovery of pure [18F]altanserin, and plasma caffeine levels. Radioactivity in whole blood and plasma samples (separated from whole blood samples by 3 min centrifugation at 1,000 g) was measured in an automated gam-ma counter (1480 WIZARD, Wallac-ADL GmbH, Freiburg, Germany) that was cross-calibrated weekly with the PET scan-ner. The fraction of nonmetabolized [18F]altanserin was deter-mined in all samples by selective liquid-liquid extraction with quantification of the recovery of total radioactivity followed by thin-layer chromatography.19

Image ProcessingTo exclude structural brain abnormalities and to define ana-

tomic regions, individual high-resolution magnetic resonance imaging (MRI) data sets were acquired (Magnetom Trio 3T scanner, Siemens, Erlangen, Germany) using a three-dimension-al T1-weighted MPRAGE sequence (voxel size 1 mm3). MRI was oriented according to the anterior-posterior commissure (AC-PC line) using the MPI-Tool software (version 6.42, ATV

GmbH, Germany). The MRI was segmented (SPM5, Wellcome Trust Centre for Neuroimaging, London, UK) and the gray mat-ter compartment was converted into a binary mask by assigning all voxels with a probability higher than 0.1 to gray matter.

A realignment of later PET frames to the first frame was performed with SPM2 (Wellcome Trust Centre for Neuroim-aging, London, UK). Summed PET images of the first scan were manually coregistered to individual MRI data using the MPI-Tool software. Summed PET data of the second scan were aligned to the coregistered PET images of the first scan using the mutual information algorithm of the MPI-Tool software. All transformations were checked visually and applied to each 10-min frame.

Binding PotentialThe proposed outcome parameter of [18F]altanserin PET is

the binding potential (BPP, mL/cm³)20 related to the plasma com-partment activity of [18F]altanserin (CP). The BPP was calculated according to BPP = (Cvoxel – Creference) / CP (with Creference being the average concentration in the reference region cerebellum) for each voxel (Cvoxel) and frame.13 The average of all frames resulted in the final parametric image of [18F]altanserin BPP.

As a substantial fraction of radiometabolites crosses the blood-brain barrier during [18F]altanserin bolus plus constant infusion experiments, the binding potential related to the non-displaceable compartment (BPND) or distribution volume (VT) cannot be determined reliably as the fraction of radiometabo-lites is included in the model. BPP is therefore the only accurate outcome parameter of 5-HT2AR densities in [18F]altanserin bo-lus plus constant infusion experiments.

Region of Interest AnalysisThe volume-weighted average BPP values for the neocortex

served as primary outcome. According to the atlas template im-plemented in PMOD software (PMOD v.2.95, PMOD Group, Zurich, Switzerland, AAL atlas [Anatomical Automatic Label-ing, merged version]21), neocortex was anatomically defined as being composed of the following side-averaged regions of interest (ROI): precentral and postcentral gyrus (gy), rolandic operculum, supplementary motor area, superior, medial and inferior frontal gy, rectal gy, insula, calcarine sulcus, cuneus, lingual gy, occipital lobe, fusiform gy, supramarginal gy, angu-lar gy, precuneus, paracentral lobule, Heschl gy, parietal lobe, and temporal lobe. BPP images were normalized to the template space and BPP values were individually read out from the para-metric image voxels classified as gray matter on each partici-pants’ MRI. In detail, individual 1 mm3 voxel size MRI data sets were spatially normalized to the Montreal Neurological Institute (MNI)/International Consortium for Brain Mapping (ICBM) 152 T1 template as supplied with SPM2. The MRI-based nonlinear transformation matrix was transferred to the parametric BPP images.

For detailed regional analysis, ROI were defined on indi-vidual, not normalized, MRI data using the PVElab software (PVElab pipeline program, v. 2.0, PVEOut project, Neurobi-ology Research Unit, Rigshospitalet, Copenhagen University Hospital, Denmark).22

As the field of view of PET was smaller than that covered by MRI, ROI were adjusted manually with the PMOD software if


ROI size protruded from one of the two corresponding PET fi eld of views. Within the cerebellar ROI only PET voxels classifi ed as gray matter were used to generate time-activity curves (Creference). A subset of cortical ROI defi ned by PVElab was merged, side averaged, and used for readout of regional BPP values from the parametric image voxels previously classifi ed as gray matter.

Statistical AnalysisAll results are reported as average values (mean ± standard

deviation) unless otherwise noted. Statistical signifi cance of BPP was assessed with a mixed model (proc MIXED, SAS In-stitute Inc. Cary, North Carolina, USA) repeated measure anal-ysis of variance (rmANOVA) with condition (baseline versus sleep deprivation) and ROI as within-subject factors and sex as a between-subject factor. Paired t-tests were used for pairwise

comparisons (baseline versus sleep deprivation) of BPP, PET imaging characteristics, EEG recording, PVT, and self-ratings (except for SSS ratings where the nonparametric Wilcoxon test was used). Pearson product moment correlations were used to assess relationships between relative changes ((sleep depri-vation – baseline) / baseline) in BPP and EEG power spectral analysis, PVT, and self-ratings on VAS (baseline – sleep de-privation). Analysis of association between self-ratings on SSS and BPP was done with Spearman rank correlation. To correct for multiple comparisons the false discovery rate (FDR) pro-cedure23 (proc MULTTEST, SAS) was applied to the P values obtained by the regional BPP analyses.

RESULTS

Experimental FrameworkThe average sleep duration the night before the fi rst scan was

6.6 ± 1 h according to the self-rating of the participants and 6.6 ± 1.2 h according to actigraphy. At the time of scan-start, par-ticipants were awake on average for 6.9 ± 1.1 h at baseline and for 30.6 ± 0.8 h after sleep deprivation. Self-reported wake-up times and actigraphy times differed only by a few minutes.

PET scans at baseline and sleep deprivation condition were not signifi cantly different with regard to time of day of scanning, injected activity or mass of injected altanserin, fP value, or the fraction of parent compound in plasma. An overview of these data is given in Table 1. There was no signifi cant difference be-tween the participants who underwent scanning fi rst or second during the day (BPP neocortical region: 1.45 vs. 1.37, P = 0.45).

The free fraction of ligand in plasma was determined by ul-trafi ltration. It is well known that this method is error-prone and highly variable for ligands with a high fraction of protein bind-ing such as [18F]altanserin. As reported pre viously,13,15 the accu-racy of the data did not improve by correcting with the fP value: BPF (which is the binding potential in relation to the free frac-tion corrected plasma activity) correlated signifi cantly with fP

(r = -0.62, P < 0.00001), whereas BPP did not (r = 0.18, P = 0.2).

Effects of Sleep DeprivationMean parametric images of BPP before and after sleep de-

privation are depicted in Figure 1. Sleep deprivation signifi -

Table 1—Statistical data for baseline and sleep deprivation with regard to PET imaging characteristics, self ratings, performance measurements, and wake EEG recordings, respectively

Baseline Sleep deprivation P valueDaytime of scanning in h:min 13:10 ± 1:07 12:48 ± 0:50 0.12Injected activity in MBq 202 ± 44 211 ± 29 0.50Injected amount in nmol 4.23 ± 2.6 3.48 ± 2.3 0.21f P value in % 1.13 ± 0.6 1.45 ± 1.4 0.20Fraction of unmetabolized ligand in % (mean 120-180 min) 0.78 ± 0.04 0.76 ± 0.04 0.16Visual analog scale (0: very sleepy, 1: very awake) 0.62 ± 0.2 0.26 ± 0.1 < 0.0001Stanford Sleepiness Scale (median 0: awake, 7: nearly sleeping) 2 4 0.001Psychomotor vigilance task – speed 10% fastest in m/s 4.91 ± 0.7 4.75 ± 0.7 0.08Psychomotor vigilance task – number of lapses in % of reactions 3.07 ± 3.0 3.43 ± 5.6 0.83

Paired t-test (except for Stanford Sleepiness Scale: Wilcoxon test). EEG, electroencephalography; f P, fraction of free [18F]altanserin in plasma (unbound to proteins); MBq, megabecquerel; PET, positron emission tomography.

Figure 1—MRI (A) and parametric images representing binding potential (BPP) of [18F]altanserin (5-HT2A receptor binding) at baseline (B) and after sleep deprivation (C). Note the overall cerebral increase of BPP values after sleep deprivation. Spatially normalized average images (n = 14).

R L BPP2.4

0

A

B

C


cantly increased the binding potential of [18F]altanserin in the neocortical ROI (AAL) by 9.6% from 1.36 ± 0.26 to 1.49 ± 0.27 mL/cm³ (paired t-test, P = 0.024). The further explorative analysis of the automatically delineated ROI (PVElab) with a rmANOVA showed a significant main effect of condition (P < 0.0001) and ROI (P < 0.0001) but not of sex (P = 0.08) or of the condition*ROI interaction (P = 0.99). Significant in-creases (FDR corrected paired t-test) were found in the fol-lowing ROI: medial inferior frontal gyrus, insula, anterior cingulate, parietal, sensomotoric, and ventrolateral prefrontal cortices. The respective BPP of the investigated ROI are re-ported in Table 2. Average BPP increased in all ROI and in most participants. For example, in the ventrolateral prefrontal cortex 11 out of 14 participants showed an increase of indi-vidual BPP. Individual BPP values for this region are plotted in Figure 2. An overview of the relative changes of BPP rang-ing from 5% (orbitofrontal cortex) to 15% (insula) is given in Figure 3.

There were clear trends with regard to reaction times of the PVT (prolonged after sleep deprivation by 5.8 ± 19 ms) and the 10% fastest reaction speeds (reduced by 3.4 ± 6.1%; P = 0.067) (Table 1).

During the waking EEG recordings the power density of the delta-frequency range was significantly increased ((sleep deprivation – baseline) / baseline) after sleep deprivation by 33.7 ± 32% (P = 0.003, n = 13). Power density in the theta range showed a significant increase of 25.6 ± 40% (P = 0.038, n = 13). No significant differences were found in the alpha or lower alpha band.

Questionnaires (VAS and SSS) revealed a significant increase of sleepiness after sleep deprivation (Table 1). The correlation analysis for changes in self-ratings (sleep deprivation – baseline) versus EEG power-density changes (see previous text) gave a positive significant correlation (theta: r = 0.68, P = 0.014; delta: r = 0.6, P = 0.039, n = 13). No significant correlations were found between the self-reports and the EEG changes toward the cognitive performance measures.

Correlative Analysis of Receptor Densities and Sleep-Related Measures

The change in 5-HT2AR availability correlated significantly with the time spent awake at time of scanning in some regions, and in several regions with the relative change in PVT per-formance. The respective significance levels and correlation coefficients are reported in Table 3. A representative plot for the superior frontal gyrus atlas region is depicted in Figure 4. No significant correlations were found between the subjective sleepiness rating (VAS) or the delta and theta EEG power den-sity versus the relative changes in BPP.

DISCUSSIONThe current study demonstrates that a single night of total

sleep deprivation leads to a global increase of specific [18F]al-

Table 2—Regional [18F]altanserin binding potentials (BPP) before and after 1 night of sleep deprivation (n = 14)

Condition t-testRegion Baseline Sleep deprivation P P FDR corrected

Orbitofrontal ctx 1.48 ± 0.3 1.55 ± 0.41 0.358 NSMedial inferior frontal gyrus 1.47 ± 0.31 1.61 ± 0.34 0.023 < 0.05Anterior cingulate ctx 1.51 ± 0.31 1.7 ± 0.32 0.007 < 0.05Insula 1.48 ± 0.32 1.67 ± 0.31 0.010 < 0.05Superior temporal gyrus 1.63 ± 0.33 1.77 ± 0.35 0.047 NSParietal ctx 1.36 ± 0.32 1.51 ± 0.32 0.022 < 0.05Medial inferior temporal gyrus 1.74 ± 0.34 1.86 ± 0.35 0.091 NSSuperior frontal gyrus 1.33 ± 0.3 1.44 ± 0.33 0.051 NSOccipital ctx 1.44 ± 0.3 1.57 ± 0.3 0.040 NSSensomotoric ctx 1.03 ± 0.27 1.14 ± 0.29 0.016 < 0.05Posterior cingulate ctx 1.31 ± 0.36 1.45 ± 0.39 0.041 NSDLPFC 1.49 ± 0.32 1.61 ± 0.34 0.077 NSVLPFC 1.41 ± 0.29 1.55 ± 0.36 0.015 < 0.05

Paired two-tailed t-test. Ctx, cortex; DLPFC, dorsolateral prefrontal cortex; FDR, false discovery rate; NS, not significant; P, significance level; VLPFC, ventrolateral prefrontal cortex. Bold numbers represent P values that exceed the 0.05 threshold.

Figure 2—Individual changes of binding potential (BPP) of [18F]altanserin (5-HT2A receptor binding) in the atlas neocortical region (mean of left and right). Straight lines indicate participans with increased BPP values, dotted lines indicate participants with decreased BPP values.

Baseline Sleep deprivation

2.2

2.0

1.8

1.6

1.4

1.2

1.0

BPP


tanserin binding in the human neocortex. This finding points to an upregulation of 5-HT2AR density because the PET outcome parameter BPP is directly proportional to the concentration of available 5-HT2AR as has been shown by comparison with in vitro methods.14 To our knowledge this is the first investigation of the effect of sleep deprivation on the cerebral density of the 5-HT2AR. So far, no direct in vivo or in vitro data are available with regard to prolonged wakefulness and its consequences for cerebral 5-HT2AR in humans or animals. Findings in rodents suggest that during sleep deprivation the serotonin release is higher than during the previous wake period (as discussed in the introduction). Sleep deprivation might therefore alter sero-

tonin levels, which in turn changes both the actual occupancy and the medium- to long-term expression of 5-HT2ARs. In mo-lecular terms, 5-HT2ARs are upregulated throughout the period of sleep deprivation as shown in the current study.

The observed effect size of approximately 10% change in receptor availability is in accordance with earlier findings of studies investigating behavioral or pathologic influences on neuroreceptor regulation. For example, a significant 6.4% de-crease of dopamine D2 receptor availability in humans in cau-date was reported after 24 h of sleep deprivation using [11C]raclopride PET.3 These data were recently confirmed by the same group reporting a 5.1% decrease in ventral striatum.24

Figure 3—Average relative changes (sleep deprivation – baseline) / baseline of the binding potential (BPP) of [18F]altanserin (5-HT2A receptor binding) in different brain regions indicating the effect of sleep deprivation (n = 14). Error bars denote SEM. Ctx, cortex; DLPFC, dorsolateral prefrontal cortex; VLPFC, ventrolateral prefrontal cortex. *Significantly different from zero, paired t-test, P < 0.05 controlled for false discovery rate.

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

DLPFC

VLPFC

Sup fro

ntal gy

rus

Med inf

fronta

l gyrus

Orbitofr

ontal c

tx

Post cin

gulate

ctx

Ant cing

ulate c

txIns

ula

Med inf

temp g

yrus

Sup tem

p gyru

s

Sensom

otoric

ctx

Parietal

ctx

Occipita

l ctx

Relat

ive ch

ange

s in

BPP

* * *

**

*

*

**

*

Table 3—Correlation between the relative changes of the [18F]altanserin binding potential BPP and the time spent awake (n = 14) or relative changes in psychomotor vigilance task (PVT, n = 14) between baseline and sleep deprivation

BPP (day 2 – day 1) / day1 Time spent awake at scantime (min) PVT speed 10% fastest (day 2 – day 1) / day1Orbitofrontal ctx NS r = -0.66, P = 0.010Med inf frontal gyrus NS r = -0.63, P = 0.016Ant cingulate ctx NS r = -0.58, P = 0.028Insula NS r = -0.65, P = 0.013Sup temp gyrus r = -0.58, P = 0.031 r = -0.58, P = 0.029Parietal ctx NS r = -0.63, P = 0.016Med inf temp gyrus NS r = -0.60, P = 0.024Sup frontal gyrus r = -0.64, P = 0.014 r = -0.65, P = 0.013Occipital ctx NS NSSensomotoric ctx NS r = -0.66, P = 0.011Post cingulate ctx r = -0.55, P = 0.039 r = -0.56, P = 0.037DLPFC r = -0.62, P = 0.018 r = -0.58, P = 0.031VLPFC NS NS

Ant, anterior; ctx, cortex; DLPFC, dorsolateral prefrontal cortex; inf, inferior; med, medial; NS, not significant; P, significance level under the null hypothesis of zero correlation; post, posterior; r, Pearson correlation coefficient; sup, superior; temp, temporal; VLPFC, ventrolateral prefrontal cortex.


Otherwise it was shown with [11C]DASB PET that depending on the season 5-HTtransporter BPND in putamen varies between winter and summer by approximately 11%.25 In a study of pa-tients with epilepsy, an approximately 8% increase of opioid receptor availability in the epileptic focus 8.5 h after a spon-taneous seizure was measured with [11C]diprenorphine PET, with a gradual return to normal levels.26 Participants suffering from a major depressive disorder showed a decrease of receptor availability from 8% to 30% using [18F]altanserin PET.27 Given these pathologic data, we did not expect that alterations caused by sleep deprivation were in the same range as those changes caused by severe neurologic disorders. Instead, an overall dif-ference of 10% is more likely to reflect a physiologic effect.

An intriguing finding of this study is that correlations of sleep deprivation duration versus BPP suggest that, at least in the investigated time span, 5-HT2AR availability seemed to de-crease with more time awake. This finding is counterintuitive and opposite of the observed increase in 5-HT2AR availability in the sleep deprivation group as a whole.

Basal serotonin concentrations in the rat brain (microdialy-sis) have been reported to be in the range of 0.5 nM28 and to rise several-fold with sleep deprivation.7 In turn, higher serotonin levels will result in downregulation of the 5-HT2AR, which is a constant observation for G-protein coupled receptors. How-ever, stimulation of 5-HT2ARs with serotonin or equivalent agonists showed a more complex reaction pattern.29 Internaliza-tion of the membrane-bound, green-flourescent-protein-fused 5-HT2AR upon serotonin stimulation (greater than 100 nM) is completed within approximately 10 min in transfected HEK239 cells30 and no internalization was observed at lower serotonin concentrations, even at longer time scales. In contrast, an in vitro study of the effect of serotonin stimulation (10 µM) on rat cerebellar granule cells showed that the binding of [3H]ketan-serin and 5-HT2A mRNA increased within hours, indicating an upregulation of 5-HT2AR.31 Moreover, atypical receptor regu-

lations were also reported for 5-HT2AR antagonists: a chronic blockade of 5-HT2AR was often followed by a “paradoxical” downregulation. The increase of 5-HT2AR density in this study is in line with some of the previously mentioned results. How-ever, as 5-HT2AR was investigated at only one time point, it cannot be excluded that there is an early decrease followed by a compensatory increase of 5-HT2ARs just at the time of the PET scan. It is therefore important to establish the specific condi-tions in humans and in vivo.

Regarding the study design, we decided to investigate all participants in the same order: ’baseline’ PET scan followed by sleep deprivation and repeated PET scan. Previous studies evaluating the reliability of [18F]altanserin PET in a test-retest design did not find any evidence for a carryover or order ef-fect. Furthermore, the PET tracer methodology makes it highly unlikely to produce any pharmacologic or background (radio-activity) effects with a given time shift of 24 h.

It is also important to ensure that the experimental setting does not induce systematic behavioral changes. To deal with this issue, the behavior of the participants was monitored during the week before the investigation using daily questionnaires about their sleep-wake behavior, sleep quality, fatigue, and overall per-formance. During the last two days before the experiment par-ticipants were controlled by actigraphy. None of the measured items was significantly different the day before the experiment compared with the preceding period. Data from previous studies in which we recorded sleep EEG are in accordance with reports from other groups showing that the first night spent at a sleep laboratory is usually subjectively and objectively disturbed (“first night effect”). Therefore, our participants were allowed to spend the night before the first scan at home, which guaranteed a high degree of “normality” the next morning. Importantly, most of the participants were employees of the Juelich Research Cen-ter (staff of approximately 5,000). For that reason, most of the study participants were familiar with the location, which in turn

Figure 4—Significant correlations between the relative changes of [18F]altanserin binding potentials (BPP) and the time awake after sleep deprivation (left column, n = 14) as well as the relative changes in psychomotor vigilance task (PVT) performance (10% fastest reaction speeds, right column, n = 14), respectively. The graphs depict the superior frontal gyrus atlas region. Correlation coefficients and significance levels for the other atlas regions are given in Table 3.

Relat

ive ch

ange

s in

BPP

Relat

ive ch

ange

s in

BPP

-20%

-10%

0%

10%

20%

30%

40%

-20% -15% -10% -5% 0% 5% 10%

Relative changes in PVT performance

r = -0.64P = 0.014

r = -0.65P = 0.013

-20%

-10%

0%

10%

20%

30%

40%

1750 1800 1850 1900 1950

Time awake (min)


has also contributed to a low stress level. After completing the PET experiment the participants returned to their normal daily working routine. Except for the scanning time and the sleep de-privation period during the night, the daily routine of the partici-pants was therefore mostly identical to the day of the first scan. The current setting assured that sleep deprivation was the main and intended difference between the two scans.

One could also question whether the observed effect size is big enough to be clearly distinguished from normal test-retest variability. It is therefore important to note that on a group level high stability has repeatedly been demonstrated for [18F]altan-serin PET. In a reproducibility study the relative difference be-tween test and retest scans was on average only -0.2% in five cortical regions14 (also calculated based on the data of Tables 3 and 4, of reference14, n = 8). In a more recent test-retest evalua-tion of [18F]altanserin PET15 the average relative difference for 10 cortical regions was comparably low (-0.8%; n = 6). In ad-dition, in a study on the longitudinal stability of [18F]altanserin PET the average relative difference of six cortical regions was -3.2% after a 2-yr period in 12 healthy control participants.32 These reports clearly demonstrate that the observed effect size of 9.6%, although small in absolute terms, is rather big in rela-tion to the reported test-retest variability of [18F]altanserin PET.

Another theoretical constraint derives from our approach to quantify 5-HT2AR with [18F]altanserin BPP. We cannot rule out the possibility that the observed changes might (at least partly) be caused by a change of receptor affinity or by a re-duced competition with endogenous serotonin. However, there is no experimental evidence so far that sleep deprivation affects 5-HT2AR affinities and, as previously mentioned, an increase in serotonin concentration after sleep deprivation can be hypoth-esized. Furthermore, three reports have been published sug-gesting that [18F]altanserin BPP is insensitive to displacement by endogenous serotonin.12,33,34

There was a clear trend of deterioration in PVT performance after sleep deprivation but levels of significance were not reached. As a potential explanation, training effects might have overlaid to some extent the worsening in reaction speed after sleep deprivation. Our study participants trained the task only once, but the first trials are known to improve reaction speed and reaction time most significantly.35

As previously mentioned, the longer a participant was awake, the increase in BPP was less. The decreasing BPP with increasing waking time might follow a common pattern of sleepiness ratings or cognitive measures after sleep depriva-tion: after progressive deterioration during the night period of deprivation alertness improves gradually during the follow-ing daytime period. In line with this pattern we found that the greater the observed increase in [18F]altanserin BPP, the greater the deterioration in cognitive performance (although we did not find a relation between the self-rating of sleepiness and the BPP changes). Changes in PVT are generally considered to reflect the attentional and arousal state of the participant and have been proven as a measure of sleep loss.36 In a functional MRI study, best performance on the PVT (fast reactions) was found to be associated with increases in the BOLD (blood oxygen level de-pendent) signal in the middle and inferior frontal and inferior parietal lobe.37 A study on arterial spin labeling perfusion re-vealed activations or increases of cerebral blood flow during

the PVT in middle and inferior frontal cortex, insula, anterior cingulate cortex, and inferior parietal lobe.38 In a PET study of the cerebral metabolic rate of glucose after 24 h of sleep de-privation, a global decrease (8%) was observed with most pro-nounced cortical changes in prefrontal and parietal cortices.39 These data are in line with the regional pattern of BPP increases and their correlations to PVT, as observed in this study. Most of these areas are considered to be part of the widespread at-tentional network.40

The current findings of increased 5-HT2AR density might also be related to a recent concept of sleep regulation, the synaptic homeostasis theory.41 It proposes a process of synaptic potentia-tion, which implies a general increase in receptor density during prolonged wakefulness. The observed increase in 5-HT2AR avail-ability could therefore reflect an increase in synaptic strength.

With regard to the functional consequences of these regu-latory processes we found that sleep deprivation led to an in-crease in the EEG power in the theta and delta bin of the waking EEG, which confirms previous reports indicating an increased sleep propensity in line with the homeostatic model of slow wave sleep enhancement after sleep deprivation. Interestingly, there was no correlation to the BPP of [18F]altanserin. Because increases in the theta power of the waking EEG have been iden-tified as physiologic measures of enhanced sleep propensity,42 the 5-HT2AR may not be primarily involved in this process. It has previously been proposed that serotonergic modulation of the wake-promoting system in the basal forebrain does not in-duce sleep per se. Instead, serotonin might be a trigger of the behavioral state of drowsiness or quietwaking,43 which prepares for the subsequent onset of sleep. In turn, the role of serotonin receptors is to allow modulating the level of serotonin efficacy in a relatively robust manner because varying concentrations of serotonin are highly dynamic and prone to fluctuations.

The flip-flop circuit model of the transition of sleep and wake44 proposed that a rapid switching to one state is ensured by disinhibition of one state and inhibition of the other state, re-spectively. The gamma aminobutyric acid-containing neurons in the ventrolateral preoptic area (VLPO) of the hypothalamus seem to inhibit the arousal systems during sleep, which in turn inhibit (when active) the hypothalamic area. Serotonergic in-puts from the dorsal raphe nuclei, which are part of this arousal system, inhibit the activity of VLPO neurons. It was speculated that the previously mentioned 5-HT2AR antagonists enhance slow wave sleep by antagonizing the inhibitory input to the sleep promoting cells of the VLPO. The spatial resolution of PET does not allow to report alterations of [18F]altanserin bind-ing in the VLPO. If the generally observed increase in 5-HT2AR availability also applies for the VLPO, the proposed switch may be influenced toward an increased inhibition of VLPO.

Because sleep deprivation as well as serotonergic antidepres-sants are efficient strategies in the therapy of depression it can be hypothesized that both treatments might share a common molecular basis. With regard to pathologic conditions such as depression it will be important to investigate whether serotonin receptor densities are responding in the pathologic condition to sleep deprivation.

In conclusion, this study provides in vivo evidence of in-creased levels of 5-HT2AR availability in the human brain after one night of sleep deprivation.


ACKNOWLEDGMENTSMagdalene Vögeling, Angela Weisshaupt, Lutz Tellmann,

Elisabeth Theelen, Suzanne Schaden, Hans Herzog, Johannes Ermert, and Heinz H. Coenen are gratefully acknowledged for excellent technical assistance as well as radioligand supply. The authors thank Hans-Peter Landolt for valuable discussions.

DISCLOSURE STATEMENTThis was not an industry supported study. The authors have

indicated no financial conflicts of interest.

REFERENCES1. Longordo F, Kopp C, Luthi A. Consequences of sleep deprivation on neu-

rotransmitter receptor expression and function. Eur J Neurosci;29:1810-9.2. Elmenhorst D, Meyer PT, Winz OH, et al. Sleep deprivation increases A1

adenosine receptor binding in the human brain: a positron emission tomog-raphy study. J Neurosci;27:2410-5.

3. Volkow ND, Wang GJ, Telang F, et al. Sleep deprivation decreases bind-ing of [11C]raclopride to dopamine D2/D3 receptors in the human brain. J Neurosci;28:8454-61.

4. Zeitzer JM, Maidment NT, Behnke EJ, et al. Ultradian sleep-cycle variation of serotonin in the human lateral ventricle. Neurology;59:1272-4.

5. Landolt HP, Wehrle R. Antagonism of serotonergic 5-HT2A/2C recep-tors: mutual improvement of sleep, cognition and mood? Eur J Neuro-sci;29:1795-809.

6. Penalva RG, Lancel M, Flachskamm C, Reul JM, Holsboer F, Linthorst AC. Effect of sleep and sleep deprivation on serotonergic neurotransmis-sion in the hippocampus: a combined in vivo microdialysis/EEG study in rats. Eur J Neurosci;17:1896-906.

7. Lopez-Rodriguez F, Wilson CL, Maidment NT, Poland RE, Engel J. Total sleep deprivation increases extracellular serotonin in the rat hippocampus. Neuroscience;121:523-30.

8. Alfaro-Rodriguez A, Gonzalez-Pina R, Gonzalez-Maciel A, Arch-Tirado E. Serotonin and 5-hydroxy-indole-acetic acid contents in dorsal raphe and suprachiasmatic nuclei in normal, malnourished and rehabilitated rats under 24 h of sleep deprivation. Brain Res;1110:95-101.

9. Geyer MA, Vollenweider FX. Serotonin research: contributions to under-standing psychoses. Trends Pharmacol Sci;29:445-53.

10. Landolt HP, Meier V, Burgess HJ, et al. Serotonin-2 receptors and human sleep: Effect of a selective antagonist on EEG power spectra. Neuropsycho-pharmacology;21:455-66.

11. Popa D, Lena C, Fabre V, et al. Contribution of 5-HT2 receptor subtypes to sleep-wakefulness and respiratory control, and functional adaptations in knock-out mice lacking 5-HT2A receptors. J Neurosci;25:11231-8.

12. Matusch A, Hurlemann R, Rota KE, et al. Acute S-ketamine application does not alter cerebral [18F]altanserin binding: a pilot PET study in hu-mans. J Neural Transm;114:1433-42.

13. Pinborg LH, Adams KH, Svarer C, et al. Quantification of 5-HT2A recep-tors in the human brain using [18F]altanserin-PET and the bolus/infusion approach. J Cereb Blood Flow Metab;23:985-96.

14. Smith GS, Price JC, Lopresti BJ, et al. Test-retest variability of serotonin 5-HT2A receptor binding measured with positron emission tomography and [18F]altanserin in the human brain. Synapse;30:380-92.

15. Haugbol S, Pinborg LH, Arfan HM, et al. Reproducibility of 5-HT2A re-ceptor measurements and sample size estimations with [18F]altanserin PET using a bolus/infusion approach. Eur J Nucl Med Mol Imaging;34:910-5.

16. Aeschbach D, Matthews JR, Postolache TT, Jackson MA, Giesen HA, Wehr TA. Dynamics of the human EEG during prolonged wakefulness: evidence for frequency-specific circadian and homeostatic influences. Neu-rosci Lett;239:121-4.

17. Dinges DF, Powell JW. Mikrocomputer analysis of performance on a por-table, simple visual RT task during sustained operations. Beh Res Meth Inst Comp;17:652-5.

18. Hurlemann R, Matusch A, Kuhn KU, et al. 5-HT2A receptor density is de-creased in the at-risk mental state. Psychopharmacology (Berl);195:579-90.

19. Matusch A, Stahr K, Voegeling M, et al. Determination of the plasma input function of [18F]altanserin by selective liquid-liquid extraction. Validation of a TLC method and comparison to HPLC. Neuroimage;31:T55.

20. Innis RB, Cunningham VJ, Delforge J, et al. Consensus nomenclature for in vivo imaging of reversibly binding radioligands. J Cereb Blood Flow Metab;27:1533-9.

21. Tzourio-Mazoyer N, Landeau B, Papathanassiou D, et al. Automated ana-tomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage;15:273-89.

22. Svarer C, Madsen K, Hasselbalch SG, et al. MR-based automatic delinea-tion of volumes of interest in human brain PET images using probability maps. Neuroimage;24:969-79.

23. Benjamini Y, Hochberg Y. Controlling the False Discovery Rate - A Practi-cal and Powerful Approach to Multiple Testing. J R Stat Soc;57:289-300.

24. Volkow ND, Tomasi D, Wang GJ, et al. Evidence That Sleep Deprivation Downregulates Dopamine D2R in Ventral Striatum in the Human Brain. J Neurosci;32:6711-7.

25. Kalbitzer J, Erritzoe D, Holst KK, et al. Seasonal changes in brain sero-tonin transporter binding in short serotonin transporter linked polymorphic region-allele carriers but not in long-allele homozygotes. Biol Psychia-try;67:1033-9.

26. Hammers A, Asselin MC, Hinz R, et al. Upregulation of opioid receptor binding following spontaneous epileptic seizures. Brain;130:1009-16.

27. Mintun MA, Sheline YI, Moerlein SM, Vlassenko AG, Huang Y, Snyder AZ. Decreased hippocampal 5-HT2A receptor binding in major depressive disorder: in vivo measurement with [18F]altanserin positron emission to-mography. Biol Psychiatry;55:217-24.

28. Hume S, Hirani E, Opacka-Juffry J, et al. Effect of 5-HT on binding of [(11)C] WAY 100635 to 5-HT(IA) receptors in rat brain, assessed using in vivo microdialysis nd PET after fenfluramine. Synapse;41:150-9.

29. Van Oekelen D, Luyten WH, Leysen JE. 5-HT2A and 5-HT2C receptors and their atypical regulation properties. Life Sci;72:2429-49.

30. Bhattacharyya S, Puri S, Miledi R, Panicker MM. Internalization and re-cycling of 5-HT2A receptors activated by serotonin and protein kinase C-mediated mechanisms. Proc Natl Acad Sci U S A;99:14470-5.

31. Akiyoshi J, Hough C, Chuang DM. Paradoxical increase of 5-hydroxytryp-tamine2 receptors and 5-hydroxytryptamine2 receptor mRNA in cerebellar granule cells after persistent 5-hydroxytryptamine2 receptor stimulation. Mol Pharmacol;43:349-55.

32. Marner L, Knudsen GM, Haugbol S, Holm S, Baare W, Hasselbalch SG. Longitudinal assessment of cerebral 5-HT2A receptors in healthy elderly volunteers: an [18F]-altanserin PET study. Eur J Nucl Med Mol Imag-ing;36:287-93.

33. Paterson LM, Tyacke RJ, Nutt DJ, Knudsen GM. Measuring endogenous 5-HT release by emission tomography: promises and pitfalls. J Cereb Blood Flow Metab;30:1682-706.

34. Pinborg LH, Adams KH, Yndgaard S, et al. [18F]altanserin binding to hu-man 5HT2A receptors is unaltered after citalopram and pindolol challenge. J Cereb Blood Flow Metab;24:1037-45.

35. Elmenhorst D, Elmenhorst EM, Luks N, et al. Performance impairment during four days partial sleep deprivation compared with the acute effects of alcohol and hypoxia. Sleep Med;10:189-97.

36. Van Dongen HP, Maislin G, Mullington JM, Dinges DF. The cumulative cost of additional wakefulness: dose-response effects on neurobehavioral functions and sleep physiology from chronic sleep restriction and total sleep deprivation. Sleep;26:117-26.

37. Drummond SPA, Bischoff-Grethe A, Dinges DF, Ayalon L, Mednick SC, Meloy MJ. The neural basis of the psychomotor vigilance task. Sleep;28:1059-68.

38. Lim J, Wu WC, Wang J, Detre JA, Dinges DF, Rao H. Imaging brain fatigue from sustained mental workload: an ASL perfusion study of the time-on-task effect. Neuroimage;49:3426-35.

39. Thomas M, Sing H, Belenky G, et al. Neural basis of alertness and cogni-tive performance impairments during sleepiness. I. Effects of 24 h of sleep deprivation on waking human regional brain activity. J Sleep Res;9:335-52.

40. Posner MI. Measuring alertness. Ann N Y Acad Sci;1129:193-9.41. Tononi G, Cirelli C. Sleep function and synaptic homeostasis. Sleep Med

Rev;10:49-62.42. Landolt HP, Retey JV, Tonz K, et al. Caffeine attenuates waking and sleep

electroencephalographic markers of sleep homeostasis in humans. Neuro-psychopharmacology;29:1933-9.

43. Cape EG, Jones BE. Differential modulation of high-frequency gamma-electroencephalogram activity and sleep-wake state by noradrenaline and serotonin microinjections into the region of cholinergic basalis neurons. J Neurosci;18:2653-66.

44. Saper CB, Scammell TE, Lu J. Hypothalamic regulation of sleep and circa-dian rhythms. Nature;437:1257-63.

sleep deprivation increases cerebral serotonin 2a … deprivation... · sleep, vol. 35, no. 12,...

Documents