AVII SLEEP, Vol. 26, Abstract Supplement, 2003

This 2003 Annual Meeting abstract supplement containsall of the abstracts to be presented at the APSS 17thAnnual Meeting on June 3-8 in Chicago, Illinois. This yearwe experienced the largest increase in submitted abstractsthat has ever occurred, with 1148 submissions. This repre-sents a 51% increase over that submitted in 2002. Thislikely reflects the combined meeting with the World Feder-ation of Sleep Research Societies, the 50th anniversary ofthe discovery of REM sleep being celebrated at the meet-ing, and the sustained increase in science addressing sleepand its disorders. Thus, more science will be presented inChicago this year than has been presented at any previousmeeting addressing this topic.

Of the 1148 abstracts, 169 will be presented in oral format,51 in poster symposiums, and the remainder as posters.With this number of abstracts being presented as posters,the Program Committee had to make some hard decisionsas to how best to handle this. They elected to:1) Thematically group the posters.2) Posters will be up for only one day during the meeting,

but will be up the entire day.3) Most importantly, there will be a 90 minute block of

time each day (Thursday, Friday, and Saturday, June 5,6,and 7) between 1:30pm and 3:15pm that poster viewingwill be unopposed by any other scientific presentation.

Each poster will have a unique 4 digit number within theappropriate category (listed below) which should allow foreasy identification.

The categories for this year’s science have not changedfrom last year and are listed here:

A. Basic NeuroscienceB. General PhysiologyC. Clinical PharmacologyD. DreamsE. Circadian RhythmsF. PhylogenyG. PediatricsH. AgingI. Sleep DeprivationJ. Sleep Disorders - BreathingK. Sleep Disorders - NarcolepsyL. Sleep Disorders - InsomniaM. Sleep Disorders - ParasomniasN. Sleep Disorders - Movement DisordersO. Sleep Disorders - Neurologic Disorders

P. Sleep in Medical DisordersQ. Sleep in Psychiatric DisordersR. Instrumentation & MethodologyS. Sleep EducationT. Molecular Biology & GeneticsU. Sleep & Behavior

Despite the troubled times in America and abroad, weexpect this will be the largest APSS meeting to date. It isour hope that this year’s Annual Meeting will provide theopportunity for the meaningful exchange of both clinicaland basic science related to sleep and its disorders.

David P. White, M.D.Editor-in-Chief

EDITORIAL

0001.ANeostigmine-Induced REM Sleep Enhancement In C57BL/6J (B6)Mouse Is Concentration-Dependent And Blocked By AtropineBowman GN, Douglas CL, Baghdoyan HA, Lydic RDepartment of Anesthesiology, University of Michigan, Ann Arbor, MI

Introduction: Microinjection of the acetylcholinesterase inhibitorneostigmine bromide (Neo) into the pontine reticular formation (PRF) ofB6 mouse causes a REM sleep-like state (REM-Neo) characterized bymotor atonia, EEG activation, and disordered breathing (Sleep 25: 835,2002). The present study is testing the hypothesis that REM-Neo is con-centration-dependent and can be blocked by the muscarinic antagonistatropine sulfate.Methods: Four mice each received PRF microinjections of saline (50nl), four concentrations of Neo (1.33, 13.3, 133, 1330 ng/50 nl), and 13.3ng Neo + 10 ng atropine/50 nl. Dependent measures included %REMsleep, %Wake, %NREM sleep, REM latency, REM epoch duration, andnumber of REM epochs, which were quantified during 4 hours of EEGand EMG recording.Results: ANOVA revealed a significant (p

both NMDA receptor antagonist and GABAA receptor agonist anesthet-ics.

Research supported by NS 33987, HL 60292, MH55772 and 2 PO1AG 09975-07

0004.ACorticotropin-Releasing Hormone (CRH) Mediates Amygdaloid-Kindling Induced Sleep-Wake Alteration in the RatChang F,1 Yi P,2 Tsai C1

(1) Neuroscience Laboratory, Department of Neurology, China MedicalCollege Hospital, Taichung, Taiwan, (2) Department of Nursing, Jen-TeJunior College of Nursing and Management, Miaoli, Taiwan

Introduction: We have previously hypothesized that corticotropin-releasing hormone (CRH) are involved in the regulation of physiologi-cal sleep-wake activity, based in part on observations that central or sys-temic administration of CRH receptor antagonists reduce spontaneouswaking. Other observations indicate that CRH mRNA expression is up-regulated by kindling, which resembles temporal lobe epilepsy. Sleepdisturbance is common, but is frequently overlooked in patients withepilepsy. The present study was designed to further elucidate the role ofCRH in amygdaloid-kindling induced sleep-wake alteration.Methods: Male Sprague-Dawley rats (250-300 g) were surgicallyimplanted with EEG recording electrodes, a chronic guide cannuladirected into the lateral cerebral ventricle and a bipolar electrode placedin central nucleus of amygdala, and were allowed a minimum of a oneweek recovery period. These rats were kept on a 12:12 h Light:Darkcycle at 23 ± 1 Celsius degree. After recovery, 24-h baseline recordingswere obtained from undisturbed animals. Rats were then habituated,with daily kindling manipulation, by daily handling and ICV injectionsof pyrogen-free saline (PFS) timed to coincide with scheduled experi-mental administrations until kindling seizure (Racine’s stage 5 seizure)was apparent. The kindling stimulus was a train of biphasic pulses (1milliseconds duration each) of 50 Hz for 1 second. The intensity for eachstimulus, ranging from 50 to 200 microAmp, was determined as athreshold for the first appearance of the concurrent afterdischarge. Theafterdischarges should more than 2 Hz spikes and have a voltage at leastthree times higher than the pre-stimulation baseline EEG. After Racine’sstage 5 seizure was apparent, 24-h EEGs were recorded beginning at thelight onset, given PFS injection and kindling stimulation prior to thelight (rest) period. Rats were then received two CRH receptor antago-nists, astressin (0.5, 2.5 or 12.5 microgram) or alpha-hCRH (1, 5 or 25microgram), instead of vehicle. Circulating corticosterone concentra-tions were determined by radioimmunoassay.Results: Our results indicate that amygdaloid-kindling, given prior tothe light (rest) period, decreases slow-wave sleep (SWS) and rapid eyemovement sleep (REMS) and increases wakefulness during light (rest)period. Slow-wave activity (SWA) during SWS is also enhanced byamygdaloid kindling during the light (rest) period. In addition, circulat-ing corticosterone concentrations are increased at 2-h after kindlingstimulation. ICV administration of astressin dose-dependently blockskindling’s effects on SWS and wakefulness. alpha-hCRH ICV adminis-tered into rats also reverses sleep-wake alteration induced after kindling.Both high doses of astressin (12.5 microgram) and alpha-hCRH (25microgram) statistically reduce kindling-induced increase of circulatingcorticosterone concentrations. However, the increase of SWA duringSWS is not altered by either astressin or alpha-hCRH.Conclusions: Our results suggest the involvement of CRH in the sleep-wake alteration induced by amygdaloid kindling when kindling stimula-tion manipulated at the beginning of the light (rest) period. However,kindling-induced decrease of REMS and increase of SWS during SWSare not mediated by CRH.

0005.ARest-Activity Cycles in Childhood and Adolescent DepressionArmitage R, Hoffmann R, Emslie T, Rintelmann J, Moore J, Lewis K,Cole DUniversity of Texas Southwestern Medical Center at Dallas

Introduction: There is clear evidence that circadian and ultradianrhythms are abnormal in adults with major depressive disorder (MDD),but few studies examine these rhythms in children and adolescents withMDD. Moreover, few studies evaluate potential gender differences inbiological rhythms. The present study examined circadian rest-activitycycles, based on actigraphy in 100 children with MDD and in healthynormal controls (NCs), 8-17 yrs of age.Methods: Fifty-nine (28 females, 31 males) symptomatic and unmedi-cated depressed outpatients and 41 (20 females, 21 males) normal con-trols participated in study. All participants underwent full psychiatricand medical evaluations to determine eligibility. Regularized sleepschedules were established prior to study commencement. Actigraphswere worn 24 hours per day and recorded motion and light exposureover a minimum of five consecutive days. Thresholds, sensitivities, scal-ing and epoch length were held constant across all individuals. Sleepdiaries and habits questionnaires were also completed daily throughoutthe study. Circadian amplitude and period length of activity cycles werequantified using power spectral analysis of activity counts across theentire recording period. In addition, total activity count, average lightexposure and time spent in >1000 lux light were computed for each par-ticipant. All data were then coded for group (MDD vs NC), gender andage using a cutpoint of 12 years. ANOVAs contrasted between-groupdifferences testing the 3-way interaction first, followed by multiple com-parisons.Results: Significant group by gender by age interactions were obtainedon nearly all rest-activity measures. Although the period length did notdiffer between patients and NCs, circadian amplitude was significantlyblunted in the MDD group, particularly among girls. Adolescent boyswith MDD also showed blunted circadian amplitude, but not those under13 years of age. Moreover, the MDD groups spent significantly less timein bright light and had lower light exposure overall, compared to NCs(p

erogeneous nucleus containing serotonergic, GABAergic, glycinergic,cholinergic and glutamatergic neurons. The NMC receives axonal pro-jections from neurons in the dorsal and intermediolateral column of thespinal cord, as well as the upper central nervous system including cor-tex, thalamus, and the ventral mesopontine junction (VMPJ). Neurons inthe NMC have been shown to project to the dorsal and ventral horns andintermediolateral column of the spinal cord, thalamus and cortex. Phys-iologically, the NMC plays an important role in sensory-motor integra-tion. Somatic and visceral noxious and innocuous afferent pathwaysconverge on the NMC. Activation of the NMC powerfully inhibitsmotoneuron activity. The VMPJ includes the caudal portion of thedopaminergic (DA) retrorubral nucleus (A8), ventral tegmental area(A10), serotonergic supralemniscal area (B9) in the caudoventral mid-brain and rostroventral paralemniscal tegmental field in the pons. Gluta-matergic and GABAergic neurons are also found in the VMPJ. Our pre-vious work demonstrated that 6-OHDA lesions in the VMPJ, whichspecifically damages to the DA neurons, produce periodic muscletwitches during sleep and insomnia in the cat. These results indicate thatDA neurons in the VMPJ play an important role in the regulation ofsleep and muscle activity during sleep. Since anatomical studies haveshown that DA neurons in the VMPJ do not project to the spinal cord andphysiological studies have demonstrated that dopamine release in themotor nuclei is not changed during pontine inhibitory area (PIA) andNMC stimulation-induced muscle tone suppression, we hypothesize thatthe VMPJ dopamine effect on muscle activity is mediated through thecaudal brainstem. In the current study, we tested this hypothesis usingmicroinjection technique. Methods: Experiments were performed in 6 cats decerebrated at thepost-mammillary-pre-collicular level. Decerebration was performedunder haloxane-oxygen anesthesia and anesthesia was discontinuedimmediately after decerebration. Bipolar stainless steel electrodes wereimplanted into the genioglossus and both sides of neck muscle to recordEMG activity. A half microliter of dopamine receptor agonists wasmicroinjected into the PIA and NMC to evaluate muscle response todopamine receptor subtype activation. All agonists were dissolved inartificial CSF and the concentration used was SKF 38393 (a D1 agonist,Sigma): 8 mg/mƒªl; quinpirole (a D2 agonist, Sigma): 4 mƒÝg/mƒÝl.Results: We found that SKF 38393 microinjected into the PIA (102.7%± 7.9% of the baseline) and NMC (97.4% ± 6.8% of the baseline)failed to change muscle tone. In contrast, dopamine D2 receptor agonist,quinpirole, microinjected into the NMC (48.5% ± 10.6% of the base-line), but not the PIA (96.2% ± 8.4% of the baseline) suppressed mus-cle tone in both the genioglossus and neck. The latency and duration ofNMC-quinpirole-induced atonia were 18.2 sec and 4.8 min, respective-ly, which are comparable to those after glutamate injections in thisregion (Lai and Siegel, 1988, latency: 18.1 sec, duration: 4.2 min).Conclusions: Our result indicated that activation of dopamine D2, butnot D1 receptor in the NMC induces muscle tone suppression in bothtongue and nuchal musculature. The VMPJ dopaminergic effect on mus-cle activity during sleep may be mediated through the dopamine D2receptor in the NMC. We hypothesized that 6-OHDA lesions in theVMPJ eliminate dopamine effect on neuronal activity in the NMC,which in term generate muscle hyperactivity in sleep similar to that seenin human RLS/PLMD patients.

This work supported by NIH grants HL41370 and HL60296 and aresearch grant from the Restless Legs Syndrome Foundation.

0007.APontine Phasic Activity is Phase-Locked to Hippocampal ThetaWave During REM Sleep in RatsKarashima A,1 Nakao M,1 Honda K,2 Nakamura K,3 Katayama N,1

Yamamoto M1

(1) Tohoku University, (2) Tokyo Medical and Dental University, (3)Akita Industry Promotion Foundation

Introduction: In cats, ponto-geniculo-occipital wave (PGO wave) isobserved in pons, lateral geniculate nucleus, and occipital cortex duringREM sleep. We showed that the PGO wave was phase-locked to hip-pocampal theta wave which is characterized by its sinusoidal slow fieldpotential (about 5 Hz). In rats, a PGO wave-like phasic field potentialhas been recorded from the pons and the cerebellum, which is referredto as a pontine wave (P wave). In this study, we investigate phase-rela-tionship between P wave and hippocampal theta wave of rats, which iscompared with the results from the cats.Methods: Five Sprague-Dawley rats (400-550g) were used in our exper-iments. Stainless bipolar electrodes (diameter 0.1 mm) were implantedin the pons (AP -0.4 (intraaural zero), ML 1.0, DV 7.0) and in the ante-rior lobe of the cerebellum (AP -2.6 (intraaural zero), ML 0.0, DV 7.0)for recording P wave. Stainless monopolar electrodes (diameter 0.1 mm)were implanted bilaterally in the hippocampus (AP -2.0 (bregma), ML2.5, DV 2.0) for recording the theta wave. The phase relationshipbetween the P wave occurrence and the hippocampal activity was ana-lyzed. This analysis was applied to a single PGO and a first event in acluster PGO (first PGO), respectively. All experimental proceduresinvolving these animals were performed in accordance with protocolsapproved by the Institutional Animal Care and Use Committee.Results: The P waves were observed in all rats and the theta waves wererecorded in four of five. The single P waves tended to occur just after thepositive peak of the hippocampal theta wave in all the rats. To almost thesame degree as the single, the first P waves were phase-locked to the hip-pocampal theta waves.Conclusions: We showed that the P waves were phase-locked to the hip-pocampal theta wave in the rats. This result suggests that the pontine Pwave generator is under modulation by theta-related activities. In cats,the first PGO wave in a cluster was less tightly phase-locked to the thetawave than the single PGO waves. This was attributed to possible differ-ences in contributors to generating single and cluster PGO. In contrast,the result here suggests that the generation mechanism of the cluster Pwave of rats is similar to that of the single.

The authors are grateful of support from the Special CoordinationFunds for Promoting Science and Technology, and a Grant-in-Aidfor Scientific Research nos. 13650481 and 13558095 from Ministryof Education, Culture, Sports, Science and Technology the JapaneseGovernment.

0008.AGABAergic Mechanisms In The REM Sleep Induction Zone Of TheRatMarks GA, Kramer GL, Birabil CGUniversity of Texas Southwestern Medical Center

Introduction: The caudal aspect of the nucleus pontis oralis (PnOc) inthe rat has been referred to as a REM sleep induction zone based on thefindings that unilateral injection of a variety of neurotransmitter agonistsinto this region elicit a long-lasting increase in the expression of REMsleep. Recent studies implicate a role for the GABAergic system in theREM sleep induction zone of the cat inasmuch as injection of GABAareceptor antagonists result in a short-lasting REM sleep induction andGABAa agonists increase wakefulness. In the rat, it has been found thatc-fos expression in GABAergic interneurons of the PnOc is reduced dur-ing the rebound from REM sleep deprivation. This is consistent with areduction in GABA transmission during periods of high propensity forREM sleep. In this preliminary report, we sought to determine the actionof GABA-receptor ligands in the PnOc of the rat as well as to determinelevels of GABA during different states of arousal.Methods: Under anesthesia, Long-Evans Hooded rats were surgicallyprepared for chronic sleep recording and additionally implanted withguide cannulae aimed at medial sites in the PnOc. After recovery, onegroup received multiple unilateral injections (60 nl) of the GABAa

A3 SLEEP, Vol. 26, Abstract Supplement, 2003

antagonist, bicuculline methiodide (1 mM), the agonist, muscimol (0.68mM), and four saline vehicle injections, all one week apart and within 1hr after lights-on. Each injection was followed by 6 hrs of recording. Ina separate group, a pin-style microdialysis probe (BR-2, BAS) was per-fused at 2 or 4 ul/min to obtain dialysate samples in PnOc during iden-tified states of wake, slow wave and REM sleep in freely moving ani-mals. GABA was assayed by derivatization and HPLC-EC.Results: Compared to mean control values, bicuculline induced a short-lasting increase in REM sleep, significant in the third hour after injec-tion. This was followed by a tendency towards decreased REM sleepamounts. Muscimol-injection resulted in continuous wakefulness for thefirst three hours. GABA levels in PnOc were highly related to state. Val-ues for REM sleep were almost half that of wake (REM, SW, AW: 0.11,0.14, 0.20 pmol/ul, respectively).Conclusions: These data are consistent with a role for GABA in thePnOc in the control of REM sleep and wakefulness. High GABAinhibitory tone may maintain the localized, sensory- and motor-relatedexcitation observed in the reticular formation during wakefulness. Lowlevels of GABA during REM sleep may not only permit the more gen-eralized tonic facilitation characteristic of this state, but also may beresponsible for disinhibiting the propagation of excitation throughout thereticular formation. The long-lasting effect to increase REM sleep byvarious agents injected into the PnOc may act commonly to induce along term inhibition of GABA transmission in the REM sleep inductionzone.

Research supported by NIH Grant RO1 57434

0009.ATetrodotoxin Inactivation Of Pontine Regions Suppresses REMSleepXiao J,1 Tang X,1 Ross RJ,2 Morrison AR,2 Sanford LD1

(1) Eastern Virginia Medical School, (2) University of Pennsylvania

Introduction: Studies using various methodologies have implicated n.subcoeruleus (SubC) and n. reticularis pontis oralis (RPO) in the gener-ation of REM. In rats, electrolytic lesions in these regions may give riseto the phenomena of REM without atonia (REM-A) in which the elec-trophysiological features of REM are normal, except for the absence ofmuscle tone. However, electrolytic lesions damage both cell bodies andfibers of passage, and the neural reorganization and adaptation that canoccur post-lesion can complicate interpretation. Tetrodotoxin (TTX) is asodium channel blocker that temporarily inactivates both neurons andfibers of passage, and thus may be functionally equivalent to an elec-trolytic lesion, but without allowing time for neural adaptation. In thisstudy, we examined the influence of microinjections of TTX into SubCand RPO on sleep in freely behaving rats.Methods: Rats (90 day old male Sprague-Dawley) were implanted withelectrodes for recording EEG and EMG. Guide cannulae were implant-ed aimed into SubC (n=6) or RPO (n=2). Each animal received twobilateral microinjections of TTX (TTX1: 2.5 ng/0.1 µl; TTX2: 5.0 ng/0.2µl) and control microinjections of saline alone (SAL). The injectionswere made 2 h following lights on and sleep was recorded for the sub-sequent 22 h. Sleep was scored from computerized records in 10 sepochs. Recordings from the 10 h light period and the 12 h dark periodwere examined separately.Results: Compared to SAL, TTX in SubC increased REM latency (SAL:75±7.7; TTX1: 395±118, p

0011.AInhibition of Rostral Basal Forebrain Neurons Promotes Wakeful-ness and Induces FOS in Orexin NeuronsSatoh S,1 Matsumura H,1 Nakajima T,1 Kanbayashi T,2 Nishino S,3 Yone-da H1

(1) Department of Neuropsychiatry, Osaka Medical College, Osaka,Japan, (2) Department of Neuropsychiatry, Akita University School ofMedicine, Akita, Japan, (3) Stanford University Center for Narcolepsy,C.A., U.S.A.

Introduction: The factors regulating the activities of orexin-containingneurons have been examined from the aspect of feeding and metabolism.On the other hand, a circadian variation in the activites of orexin neuronshas been demonstrated by use of FOS immunohistochemistry, whichraises the possibility that the activities of the neuronal elements regulat-ing sleep-wake states might also influence the activities of the orexin-containing neurons. However, this possibility has not been examinedwell. The ventromedial portion of the rostral basal forebrain is thoughtto contain the neuronal elements regulating the sleep-wake states, sincethe application of several neuroactive compounds in this brain regionaltered the sleep/wake states of an animal. Thus, we investigatedwhether the activities of the rostral basal forebrain neurons alter theactivities of the orexin neurons. Methods: Adult male rats of the Sprague-Dawley strain were used. Theyreceived the surgical operations monitoring EEG, EOG, and EMG aswell as those implanting the microdialysis probes in the ventromedialpart of the rostral basal forebrain. After an enough recovery period, weperformed following studies in accordance with the Japanese Neuro-science Society Guide for Animal Care and Use of Laboratory Animals.Firstly, we performed microdialysis-perfusion of the ventromedial por-tion of the rostral basal forebrain of rats with muscimol to make a focalinactivation in this brain region, and then examined its influence on theirbehavioral state. Secondly, we performed double immunohistochemistryto detect FOS and orexin to examine how this focal inactivation of therostral basal forebrain neurons influences the activities of orexin neu-rons. Thirdly, we studied the expression pattern of FOS following theunilateral perfusion of muscimol to examine whether the change in theexpression of FOS is simply a consequence of the change in the behav-ioral state.Results: Bilateral-perfusion with muscimol (5, 15, and 50 µM) produceda dose-dependent decrease in the amount of sleep. This perfusion withmuscimol at 50 µM produced FOS-like immunoreactivity in 37 % of theorexin neurons located in the tuberal part of the hypothalamus, whereasthe FOS-like immunoreactivity was sparse in orexin neurons of thesleeping control rats (p=0.001 by Mann-Whitney U-test). Unilateral-per-fusion with muscimol (50 µM) also suppressed sleep. In this case, FOS-like immunoreactivity was seen in 40 % of the orexin neurons on theside ipsilateral to the perfusion site but only in 10 % of them on the con-tralateral side (p=0.018 by Wilcoxon signed rank test).Conclusions: These functional data suggested that a sleep-generatingelement in the ventromedial part of the rostral basal forebrain provide aninhibitory influence on the activities of the orexin neurons in the tuberalpart of the hypothalamus.

0012.AInteractions Between GABAergic and Cholinergic Pontine Reticu-lar Mechanisms in the Control of Active Sleep and Wakefulness Xi M, Morales FR, Chase MHDepartment of Physiology and the Brain Research Institute, UCLASchool of Medicine, Los Angeles, CA 90095

Introduction: Cholinergic systems within the nucleus pontis oralis(NPO) play a critical role in the generation of active (REM) sleep. Wehave recently reported that microinjections of GABA antagonists into

the NPO produce active sleep, whereas microinjections of GABA ago-nists produce wakefulness (Xi et al., 2001). These findings indicate thata GABAergic system in the NPO plays an important role in the controlof behavioral states of active sleep and wakefulness. We therefore sug-gest that the generation of active sleep depends on an interactionbetween pontine cholinergic and GABAergic systems. Specifically, wehypothesize that the excitatory cholinergic control of NPO neurons isgated by a pontine GABAergic inhibitory system. To test this hypothe-sis, we examined the behavioral responses of chronic, unanesthetizedcats following the injection, into the NPO, of cholinergic and GABAer-gic receptor agonists and antagonists.Methods: Experiments were performed on three adult cats that wereprepared for monitoring behavioral states and for drug administration.The effective NPO region was defined by the stereotaxic coordinates atwhich an injection of carbachol (0.25 ml, 22 mM) induced active sleepwith a latency shorter than 4 min. In experimental sessions, microinjec-tions of carbachol and muscimol, a GABAsAs receptor agonist (0.25 ml,10 mM), or scopolamine, a muscarine receptor antagonist (0.25 ml, 10mM) and bicuculline, a GABAsAs receptor antagonist (0.25 ml, 10mM), were carried out. In all cats, control solutions of saline (0.25 ml)were injected into the same site that received the injection of drugs.Results: Microinjections of carbachol (n=5) or bicuculline (n=3) into theNPO elicited, with a short latency, active sleep and induced an increasein the time spent in this state. However, muscimol, when it was injectedinto the same area of the NPO 20-30 minutes prior to the injection of car-bachol, completely blocked the active sleep-inducing effects of carba-chol; the increase in active sleep induced by carbachol was also abol-ished (active sleep: 3.1±4.3%, n=4, [muscimol+carbachol] vs.60.1±5.4%, n=3, [saline+carbachol], P

Physiol. 1996; 490: 745-758) may potentially contribute to suppressionof GG activity. However, the physiological role of glycinergic inhibitionin mediating suppression of hypoglossal motor output to GG muscle innatural REM sleep has not been determined. We developed an animalmodel to chronically manipulate neurotransmission at the hypoglossalmotor nucleus using microdialysis across natural sleep-wake states inrats (Jelev A, Sood S et al., J Physiol. 2001; 532: 467-481). The presentstudy tests the hypothesis that glycine contributes significantly to thesuppression of GG activity in REM sleep during room air and CO2-stim-ulated breathing.Methods: Ten male Wistar rats were implanted with electroencephalo-gram and neck muscle electrodes to record sleep-wake states, and GGand diaphragm wires to record respiratory muscle activities. Microdial-ysis probes were implanted into the hypoglossal motor nucleus, withsites verified by histology, and were perfused with artificial-cere-brospinal-fluid (ACSF) or 0.1mM strychnine (glycine receptor antago-nist) during room air and CO2-stimulated breathing (6-8% inspiredCO2). We have shown that this dose of strychnine effectively reversesglycine-mediated inhibition of hypoglossal activity (Morrison JL et al. JAppl Physiol. 2002; 93: 1786-1796).Results: Strychnine at the hypoglossal motor nucleus increased respira-tory-related GG activity during room air (P=0.010) but not CO2-stimu-lated breathing (P=0.221) with the effect in room air being independentof sleep-wake states (P=0.625), i.e., a non-specific release from inhibi-tion. Nevertheless, GG activity was minimal in those REM sleep periodswithout GG twitches, with GG suppression from non-REM being >85%regardless of ACSF or strychnine at the hypoglossal motor nucleus androom air or CO2-stimulated breathing. GG activation in response tomicrodialysis perfusion of serotonin at the end of the experiments con-firmed an intact hypoglossal motor nucleus in all rats.Conclusions: The results show that although inhibitory glycinergicmechanisms are present at the hypoglossal motor nucleus, such mecha-nisms make a minimal contribution to the marked suppression of GGactivity and reflex responses in natural REM sleep suggesting differentcontrol of postural and respiratory motoneurons.

Research supported by Canadian Institutes of Health Research,Grant Number 15563.

0014.ASleep Spindles and Slow Waves During Light NREM Sleep ActivateMemoryMiyauchi S., Misaki M, Imaruoka T, Tanaka YRain Function Group, Communications Research Laboratory

Introduction: Recent neurophysiolgy hypothesizes that sleep spindlesand slow waves during NREM sleep are profoundly associated withlearning, especially with memory consolidation during wakefulnessbefore sleep. By recording EEG and fMRI simultaneously during lightNREM sleep and followed by an analysis of fMRI data accompanyingthe occurrence of slow waves and sleep spindles, we succeeded in get-ting robust activation in several cortical and subcortical areas most ofwhich are associated with memory processing.Methods: Six normal volunteers (25-47 years old) participated in theexperiments. Subjects were instructed to fall and remain asleep in an MRmagnet. EEGs (Oz, Pz, Cz, and Fz) and ECG were recorded inside theMR scanner. After removing artifacts caused by magnetic field fromEEG, the power spectra of EEGs during the MR scan interval were com-puted. The power spectrum was divided into the delta, theta, slow alpha,fast alpha and sigma (12-16 Hz) bands, and average powers in each fre-quency band were calculated. The fMRI data were acquired on a 1.5 TMR scanner. Five hundreds volumes — each consisting of 30 slices(voxel size: 4x4x5 mm; TE: 66 ms, FA: 90�‹, TR: 8 s.) — were acquiredcontinuously over a period of 66 minutes. During the scans, the MRmagnet room was kekept absolutely dark and no stimulus was presented

except for MR scan noise. The fMRI data were analyzed using SPM99.We entered the time series of theta-band and sigma-band as user-speci-fied regressors in the design matrix to find activation areas accompany-ing occurrence of theta waves and sleep spindles, respectively.Results: In spite of the unusual sleeping conditions, all the subjectsreached sleep stage 2. Though the subjects did no task, several brainareas were activated accompanying increases in slow waves and occur-rence of sleep spindles. The hippocampus, parahippocampal gyrus, lin-gual gyrus and BA 10 were activated accompanying slow waves duringlight NREM sleep. The thalamus, pulvinar, and retrosprenial cortex wereactivated accompanying occurrence of sleep spindles. As for the poste-rior temporal, posterior cingulate and putamen, these areas were activat-ed accompanying occurrences of either sleep spindles or slow waves, orboth.Conclusions: Neuropsychological and brain imaging studies suggestthat not only the hippocampus and parahippocampal gyrus, but also thelingual gyrus, retrosplenial cortex and BA 10 play an important role inmemory processing. Recent neurophysiology hypothesizes that sleepspindles and slow waves relate to memory reprocessing for consolida-tion during NREM sleep. This hypothesis is based on animal studies inwhich neural structures (place cell in the hippocampus and only a fewneocortical neurons) engaged in the process of learning during wakingcould be re-activated for memory consolidation during subsequentNREM sleep. Taken together, the result of the present study suggests thatmemory reprocessing for consolidation occurs during light NREM sleepinvolving several cortical and subcortical areas in human brain.

0015.AInhibitory Control Of REM Sleep By Medial Septal Neurons In RatsMallick HN, Srividya R, Gulia KK, Kumar VMAll India Institute of Medical Sciences

Introduction: A number of basal forebrain structures including septumhave been implicated in the regulation of sleep-wakefulness (S-W). Neu-ronal activity of septum is correlated with S-W. Electrolytic lesion stud-ies of the medial septal (MS) area indicate that hippocampal theta waveis dependent on the integrity of this area. However, cell specific lesionstudies to find out the involvement of the MS in sleep are lacking. Thisstudy was hence undertaken to find the role of the MS neurons in sleep.Methods: The study was carried out in 6 adult male Wistar rats (180-250g) with chronically implanted electrodes for recording sleep param-eters i.e. EEG, EMG and EOG, continuously for 24 hrs. All the record-ing electrodes were connected to an IC socket and the whole assemblywas fixed to the skull with dental cement. After post-operative recovery,all the parameters were recorded for 24 hrs. The cell specific neurotox-in NMDA (5ml in 0.2ml saline) was injected intracerebrally to destroythe MS neurons and all the parameters were recorded again for 24 hrs ondays 7th and 21st. S-W was assessed on the basis of EEG, EOG andEMG recordings taking 30 sec epochs and calculated into stages i.e.wakefulness (W), slow wave sleep (S1, S2), and rapid eye movementsleep (REM). The comparison of these components of the S-W of ratswas made between the prelesion (control) with postlesion values i.e. day7th and 21st of MS lesion, using Friedman’s two way ANOVA. At thecompletion of the experiment, the lesion sites were confirmed histolog-ically using cresyl violet staining.Results: The lesion in the MS produced an increase in wakefulness anda decrease in S1 as well as S2 when compared with the control values.There was an increase in REM sleep when compared with the controlvalues. The percentage of REM sleep (ratio of REM sleep time and totalsleep time) on both postlesion days showed a marked increase (16.5%and 15.7%) when compared with the prelesion values (11.1%). Therewas decrease in percentage of slow wave sleep on day 7th (83.5%) and21st of lesion (84.3%) compared to prelesion control (88.9%). Although,the data did not reach statistical significance but increase in REM sleepand decrease in slow wave sleep were very evident in this study. This is

SLEEP, Vol. 26, Abstract Supplement, 2003 A6

in contrast to the earlier reports on the role of other basal forebrain areason S-W.Conclusions: The present findings suggest that the MS may be involvedin the regulation of SWS and the inhibition of REM sleep.

Research supported by DST, New Delhi

0016.AOrexin A and B Levels Change in the Hypothalamus During theEstrus Cycle and Ageing in Female RatsPorkka-Heiskanen T, Alanko L, Kalinchuk A, Heiskanen S, Stenberg DInstitute of Biomedicine, University of Helsinki, Helsinki, Finland

Introduction: Orexins A and B are novel hypothalamic neuropeptideswhich have been shown to regulate food intake, energy balance, auto-nomic nervous system, vigilance states and secretion of several hor-mones. Infusions of orexin A have been shown to increase wakefulnessand decrease sleep. These observations suggest that orexins may play arole in the regulation of vigilance states. Orexin cells in the lateralhypothalamus send projections to several brain areas, including thenuclei, which regulate vigilance states and the hypothalamus. We havepreviously shown that orexin gene expression and the brain levels ofboth orexin A and B decrease in the course of ageing in male rats. Sleepvaries across the menstrual cycle in women and across the estrus cyclein female rats. In female rats both non-REM sleep and EEG power den-sity in the delta band were lower in proestrus than in other phases of theestrus cycle. We hypothesised that changes in hypothalamic orexin lev-els may at least partially explain the changes in nonREM sleep duringthe estrus cycle.Methods: 40 young (3-4 mo) and 30 middle aged (7-8 mo) female ratswere inspected for vaginal smears for at least 5 cycles. Of the young rats6.6% had irregular cycles, while more than 60% of the one year old ratshad irregular cycles. Rats were decapitated at different phases of theirestrus cycle, and brain tissues and trunk blood were collected. Orexin Aand B were measured from the hypothalamic tissue using radioim-munoassay.Results: Both orexin A and B were higher during the proestrus day thanon other days of the estrus cycle in young and old rats. Both orexin Aand B levels in the hypothalamus were lower in old than young animalsregardless of the cycle day.Conclusions: Since orexin A increases waking and decreases sleep, theincrease in orexin content during proestrus could at least partiallyexplain the decrease in sleep during the proestrus day. The physiologicalsignificance of increased vigilance is obvious, since the females prepareto mate. As in males, also in females the amounts of both orexin A andB peptides decrease during ageing. The decrease in non-REM sleep dur-ing proestrus can at least partially be explained by increased orexin lev-els.

Research supported by The European Union grant QLK6-CT-2000-00499, the Academy of Finland (# 45997) and Finska Läkaresäll-skapet

0017.AEffect Of The Wake-Promoting Agent Modafinil On Sleep-Promot-ing Neurons From The Ventrolateral Preoptic Nucleus: An In VitroPharmacological StudyGallopin T,1 Luppi PH,1 Rambert FA,2 Frydman A,2 Fort P1

(1) CNRS FRE2469, Lyon, France, (2) Laboratoire L Lafon, Maison-Alfort, France

Introduction: A large body of experimental and clinical data suggestthat the pharmacological profile of modafinil, an increasingly popularwake-promoting drug used in the treatment of narcolepsy, differs from

those of classical psychostimulants such as amphetamine. The brain tar-gets on which modafinil acts to induce wakefulness are however still amatter of debate. We made the hypothesis that modafinil could increasewaking by inhibiting the GABAergic sleep-promoting neurons from theventrolateral preoptic area (VLPO). Such action could be direct or indi-rect via the potentiation of an inhibition induced by a waking neuro-transmitter. To test this hypothesis, we studied the effect of modafinilapplication on the membrane potential and firing rate of VLPO neuronsrecorded in rat brain slices. We also determined whether a modafinil pre-treatment modify the effect of noradrenaline, carbachol, serotonin, his-tamine, dopamine, and clonidine application on VLPO neurons activity.Methods: Brain slices containing the VLPO were obtained from 15-20days old rats (OFA) according to standard procedures. Individual 300 or400mm thick slices were transferred to a thermoregulated (32°C) cham-ber, under either a dissecting microscope for sharp intracellular record-ings or a Zeiss Axioscop with an infrared camera for juxtacellularrecordings of identified cells. Slices were maintained immersed and con-tinuously superfused at 3-5 ml/min with oxygenated ACSF. Two param-eters of the drug responses were systematically quantified: 1) the ampli-tude corresponding to the maximal change in firing rate (in Hz or per-centage of the basal firing rate) and 2) the duration of the effect (in min).Amplitude and duration of the effect were compared between control,drug and wash conditions using a paired Student’s t test.Results: Pretreatment with modafinil (20, 60, 100, 200 mM) specifical-ly increased the inhibition of VLPO neurons induced by noradrenaline(0.5, 1, 10, 100mM) but had no effect when applied alone or in combi-nation with serotonine (100 mM), carbachol (5 mM) or histamine (100mM). Furthermore, pretreatment with modafinil did not increase theinhibition induced by dopamine (10, 100 mM) and clonidine (0.05 mM).Pretreatment with nisoxetine (10 mM), a selective noradrenaline reup-take blocker, similarly increased the noradrenaline-induced inhibition ofVLPO cells. Further, the potentiation by modafinil was occluded whenmodafinil and nisoxetine were co-applied.Conclusions: These results suggest that modafinil blocks the reuptake ofnoradrenaline by the noradrenergic terminals on sleep-promoting neu-rons from the VLPO. Such mechanism could at least partially be respon-sible for the waking effect of modafinil, through the potentiation of thetonic noradrenaline inhibition, present during waking, on GABAergicsleep-promoting neurons from the VLPO.

0018.AChanging Photic Stimulation Induces Phase-Locking Between Visu-al Units and Theta Rhythm in Wakefulness and SleepPedemonte M, Gambini JP, Velluti RANeurofisiología, Departamento de Fisiología. Facultad de Medicina,Universidad de la República. Montevideo, Uruguay

Introduction: The hippocampal theta rhythm (q) was shown to be asso-ciated with movements, attention, auditory processing, autonomic func-tions, learning and memory and has been postulated as an associator ofdiscontiguous events. We had reported phase-relationships between qrhythm and visual/auditory unit firing with both spontaneous and evokedactivity during wakefulness, slow wave and paradoxical (REM) sleep. Itis our tenet that the described phase-locking contribute to the informa-tion processing during wakefulness and even during sleep. The dynam-ic of the visual units temporal correlation with q rhythm is our presentresearch target looking for phase-locking induced fluctuations.Methods: Guinea-pigs were implanted with conventional electrodes forsleep-wakefulness control. A craniotomy over the lateral geniculate(LGn) region was made and left open for glass microelectrode access.Two metal bars attached to the skull permitted the stereotaxic positionreproduction. Visual stimuli consisted of flashes of light at different fre-quencies. The visual evoked extracellular unitary activity was analyzedwith peri-stimulus time, interval and autocorrelation histograms. Thehippocampal electrogram (bipolar recording) was studied using wave-

A7 SLEEP, Vol. 26, Abstract Supplement, 2003

form autocorrelation and power spectra. The crosscorrelation betweenhippocampal q field activity and spikes was obtained by spike-triggeredaveraging and was considered significant when it became flat calculatedwith the same data after “shuffling” the spike series. Our aim was tocompare the temporal correlation between the unit and the hippocampalq before and after sudden frequency light flashes shifts. This paradigmwas applied both in wakefulness and sleep.Results: In wakefulness as well as in slow wave sleep q rhythm period-ically appears well organized in the hippocampal electrogram for 6-8sepochs while the probability of phase-locking between the LGn unitaryand the q rhythm increased. This event sequence was observed severaltimes throughout recording sessions with continuous and regular photicstimulation. On the other hand, after a sudden change in flash frequen-cy, q waves amplitude and power increased while the unit firing phase-locking probability was enhanced for about 4-6 seconds. This inducedtemporal correlation was observed in wakefulness and slow wave sleep.Conclusions: The phase-locking between visual units and q rhythm is atime fluctuating phenomenon. This changing temporal correlation maybe present without any known provoking factor and/or during regularphotic stimulation; it could be the result of attention shifts in wakeful-ness. The induced phase-locking -by modifying the stimulus condition-is a phenomenon that appears during waking and also slow wave sleep,thus, assuming that the visual processing may be present in both states.We hypothezise that q rhythm could contribute introducing a temporaldimension to the processing of visual information in sleep and waking.

0019.ASleep Responses to 5-Hydroxytryptophan (5-HTP) Are Altered inInterleukin (IL)-6 Deficient MiceMorrow JD,1 Imeri L,2 Opp MR1,3

(1) Neuroscience Graduate Program, University of Michigan, AnnArbor, Michigan USA, (2) Institute of Human Physiology II, Universityof Milan, Milan, Italy, (3) Department of Anesthesiology, University ofMichigan, Ann Arbor, Michigan USA

Introduction: Serotonin is a neurotransmitter whose effects on arousalstate have been studied for nearly four decades. In an attempt to synthe-size seemingly paradoxical observations of the effects of serotonin onsleep, it has been proposed that serotonin itself promotes wakefulness,but induces the release of distinct, sleep-promoting substances that initi-ate subsequent sleep. These purported sleep-promoting substancesremain to be identified. Cytokines such as IL-1 and IL-6 areimmunomodulatory molecules known to influence sleep-wake behavior.Serotonin promotes the release of IL-1, which in turn induces the syn-thesis and release of IL-6. IL-6 promotes NREM sleep in rats. To test thehypothesis that IL-6 mediates the increases in NREM sleep observedfollowing serotonergic activation, we determined sleep-wake behaviorfrom IL-6-deficient mice after administration of 5-HTP, a serotonin pre-cursor known to enhance serotonergic activity.Methods: C57BL/6J (B6) and B6.129S6-Il6tm1Kopf (IL-6 KO) mice(Jackson Labs) were surgically implanted with transmitters (DSI ETAF20) to measure core body temperature and the electroencephalogram(EEG). Activity was monitored with infrared motion detectors. The micewere maintained on a 12:12 h L:D cycle at 27°C. On separate days, ani-mals received intraperitoneal injections of vehicle, 50 or 100 mg/kg 5-HTP (Sigma Chemicals) 15 min prior to dark onset. Animals wererecorded for 24 h beginning at dark onset. Polygraphic signals were dig-itized and stored for subsequent determination of arousal state.Results: Temperature: Both strains responded to 50 mg/kg 5-HTP witha hypothermic response that lasted 2 h (B6:-0.70 ± 0.10 ºC; IL-6 KO: -0.62 ± 0.07 ºC). The hypothermic response to 100 mg/kg 5-HTP wasmore pronounced and lasted longer (B6:-0.79 ± 0.08 ºC; IL-6 KO: -1.20± 0.11 ºC). NREM sleep: 50 mg/kg 5-HTP produced a transient reduc-tion in NREM sleep of both strains for ~2 h, but had no significanteffects on sleep-wake behavior thereafter. NREM sleep of B6 mice was

not altered by 100 mg/kg 5-HTP (vehicle: 28 ± 4%; 5-HTP: 32 ± 5%).In contrast, NREM sleep of IL-6 KO mice increased for 6 h after 100mg/kg 5-HTP (vehicle: 22 ± 4%; 5-HTP: 39 ± 3% p < .01).Conclusions: Though the sleep responses of IL-6 KO mice to 5-HTPdiffer from those of control mice, these data do not support the hypoth-esis that IL-6 mediates the sleep-promoting effects of serotonin. Thehypothermic responses to 5-HTP indicate biological activity in this sys-tem. However, the increase in NREM sleep of IL-6 KO mice after 5-HTP suggests that the absence of IL-6 decreases the sensitivity to theNREM sleep-suppressing effects of 5-HTP, thereby unmasking theeffects of a putative sleep-promoting substance induced by 5-HTP. Addi-tional studies to elucidate mechanisms mediating these responses areunderway.

Research supported by NIH MH64843

0020.AAnandamide Enhances Extracellular Levels Of Adenosine AndInduces Sleep In RatsMurillo-Rodriguez E,1 Piomelli P,2 Shiromani PJ1

(1) Molecular Sleep Lab/ Dept Neurology. Harvard Medical School &VAMC, (2) Dept. of Pharmacology, University of California, Irvine,

Introduction: Marijuana has been used for social and religious purpos-es for many years.There is increasing evidence that marijuana hasmedicinal value also. It is given to suppress nausea and vomiting andimprove appetite during chemotherapy treatment. The principal compo-nent of marijuana, delta-9-tetrahydrocannabinol (delta9-THC) increasessleep in humans (3). Endogenous cannabinoids, such as anandamide(ANA), also increase sleep (2). The mechanism by which the cannabi-noids promote sleep is not known but could be linked to an induction ofendogenous sleep factors such as adenosine (AD). To test this hypothe-sis we measured sleep and extracellular levels of AD in the basal fore-brain after injection of ANA and the cannabinoid CB1 receptor antago-nist, SR141716, alone or in combination with the endocannabinoid,ANA.Methods: Three-month-old male F344 rats (n=6) were implanted withsleep recording electrodes and a microdialysis guide cannula (IC guide,BAS) into the basal forebrain cholinergic neurons (MCPO; A= -.35; L=2.0; H= 8.5). Animals were then placed in a cage with unrestrictedmobility with food and water ad lib and housed at constant temperature(21 ± 1oC) and under controlled light-dark cycle (lights on: 07:00 to19:00h). ANA and the CB1 receptor antagonist SR141716 were dis-solved in DMSO (40% in saline). Microdialysis sampling procedure.The microdialysis probe (BAS 1mm of length) was inserted 24h beforesampling of AD levels. The following day at 07:00h, animals receivedDMSO as a vehicle. ANA (10mg/kg) was administered on day 3.SR141716A was administered (1mg/kg) on the fourth day. In the lastday, in order to block the effects of ANA, SR141716A was injected(1mg/kg in 0.5 ml) and 15 min later ANA (10 mg/kg in 0.5 ml). Imme-diately after the injections (ip in 1ml), sleep recordings were obtained(from 7:00-12:00h). After injections, microdialysis samples were col-lected (5ul) every hour (flow rate= 0.25ul/min).Adenosine and Sleepanalysis. AD levels were determined using HPLC. The EEG data record-ings were scored for W, SWS and REM sleep. Results: Levels of AD were increased during the first 3h after adminis-tration of ANA (p

sleep maybe severely attenuated.

Research supported by NIH-NIA grants (RO1-AG15853, PO1-09975) and Merit Award from the Medical Research Service,Department of Veterans Affairs supported this work.

0021.AGABAergic Control Of Orexinergic Neurons In The Regulation OfBehavioral States Thakkar MM, Mello E, Winston S, McCarley RWPsychiatry, Harvard Medical School, Boston VA Healthcare System,Brockton, MA 02401, USA

Introduction: We have shown that local perfusion of antisense oligonu-cleotides against orexin-II receptor in the subcoerulean reticular forma-tion increased REM sleep and induced cataplexy-like episodes in ratswhereas, microdialysis perfusion of orexin-A in the cholinergic basalforebrain increased wakefulness. These studies suggest that increasedorexinergic tone induces wakefulness, whereas reduced orexinergic toneincreases REM sleep. What controls the orexinergic neurons? GABA,with actions mediated via the GABA-A receptor, may be one such neu-rotransmitter. To further evaluate the role of GABAergic control of orex-inergic neurons in the regulation of sleep-wakefulness, microdialysisperfusion of a specific GABA-A agonist, gaboxadol (THIP) in the orex-inergic zone of the hypothalamus was performed and its effect on sleep-wakefulness was monitored in freely behaving rats.Methods: Using standard surgical procedures, adult male Sprague-Daw-ley rats (300-400 g) were implanted with standard sleep recording elec-trodes to record the EEG and the EMG. In addition, guide cannula tar-geted toward the orexinergic zone of the hypothalamus (AP -3.3; ML =1.5; DV = 8.5 from bregma) was also implanted. After 4 days of post-operative recovery and habituation, the microdialysis probe was insert-ed through the guide cannula. After 12 hours of recovery from probeinsertion, the experiment was begun. CMA/11 probes were used with a1 mm length and 0.24 mm O.D. dialyzing membrane (CMA/ Microdial-ysis; Acton, MA). All experiments were conducted during the lightcycle, in a sound-attenuated chamber that had the same temperature andlight conditions (lights on at 0700h and lights off at 1900 hours) as theanimals’ home cages, with food and water available ad libitum. Ratswere connected to a polygraph cable for behavioral state recording andthen to the probe inlet and outlet. Unilateral perfusion of artificial cere-brospinal fluid (ACSF = NaCl 147 mM, KCl 3 mM, CaCl2 1.2 mM,MgCl2 1.0 mM, pH 7.2) and/or THIP was carried out at a flow rate of2.0 ml/min. The details of the protocol are: Time of Day Control DayExperimental Day09.30 – 11.00 hr ACSF ACSF11.00 – 14.00hr ACSF THIP (1, 10 or 100 mM)14.00 – 15.30 hrACSF ACSF After the completion of the experiments, the animalswere sacrificed, brains removed and processed for histology andimmunohistochemistry to localize the zone of perfusion.Results: The result of our study suggests that THIP perfused in the orex-inergic perifornical region of the hypothalamus significantly decreasedwakefulness and increased nonREM sleep (n=7; p

(human: Werth et al., 1997; rat: Schwierin et al., 1999; mouse: Huber etal., 2000). In two studies, local differences in SWA were directly relatedto a use dependent recovery process (human: Kattler et al., 1994; rat:Vyazovskiy et al. 2000). High-density (256 channel) analysis of sleepEEG is a promising method to characterize local, task related EEGchanges with increased spatial resolution.Methods: Sleep in three healthy subjects was recorded during a 2-hnight-time sleep episode using a Geodesics Sensor Net composed of 256electrodes. Each subject completed two sessions separated by one week.During the hour before sleep subjects were engaged with a serial reac-tion time task (containing a pseudorandom or an implicitly learnablesequence), or with a visual control task. Waking EEG recordings (4times 1-min, eyes open) were carried out before and after the tasks. Forboth the waking and sleep EEG a semi-automatic artefact removal basedon power in two frequency bands (1-4.5 and 20-30 Hz) was performed.Power spectra of consecutive 4-s epochs (FFT routine, Hanning win-dow) were calculated. Power maps of EEG in different frequency rangeswere calculated.Results: Inspection of the average sleep EEG power maps within eachof the sessions revealed a predominance of power in the delta band (1-4Hz) over the frontal cortex and minimum values over temporal areas ofthe cortex. Power in the spindle frequency range (12-15 Hz) was maxi-mal along the midline. However, the peak in the fast spindle frequencyrange (14-15 Hz) differed in different subjects varying form 1.5 to 7.5cm posterior to Cz along the midline. On average the area of increasedpower relative to the all-brain mean was 31.8 cm2, corresponding to 20electrodes. Within-subject comparisons showed that the topographicpower distribution between the two experimental nights were extremelysimilar.Conclusions: To our knowledge this is the first time high density EEGrecordings were successfully performed during sleep. Overall, the pre-sent findings are in agreement with an extensive study of sleep EEGtopography by Finelli and co-workers (2001) based on 29-electrode all-night recordings. The results also confirm that the EEG power distribu-tion is highly individual, possibly reflecting anatomical differences.Within each subject the high-density distribution of EEG power duringsleep was highly reproducible in two nights separated by one week. Sin-gle-subject, high density analysis is thus a promising approach to 1) pre-cisely characterize topographical changes in SWA that occur duringsleep 2) determine whether local SWA changes are due to use or to plas-ticity and 3) investigate whether topographical changes in the sleep EEGare correlated with changes in waking EEG.

0024.AOxidative Stress and Cellular Damage After Sleep DeprivationGopalakrishnan A,1 Ji L,2 Cirelli C1

(1) Department of Psychiatry, University of Wisconsin - Madison, (2) Department of Kinesiology, University of Wisconsin - Madison

Introduction: Oxidative stress occurs whenever there is an imbalancebetween oxidant production and antioxidant defenses, either because theformer is increased and/or because the latter are decreased. At the cellu-lar level, such imbalance can result in structural damage due to oxidativemodifications of proteins, lipids and nucleic acids. It is currently unclearwhether any oxidative damage occurs during sleep deprivation, nor it isknown, should this be the case, whether sleep could play a protectiverole. The aims of this study were four: 1) measure antioxidant enzymeactivity and cellular oxidant production to determine whether sleepdeprivation causes, and recovery sleep relieves from, oxidative stress; 2)measure lipid peroxidation and protein oxidation to search for direct evi-dence of cellular damage after sleep deprivation; 3) determine whethersleep deprivation has differential effects in the brain (cerebral cortex)compared to peripheral tissues (liver, skeletal muscle); 4) compare theeffects of short sleep deprivation (8 hours; gentle handling) with those ofprolonged sleep loss (6-14 days; disk-over-water method).

Methods: Polygraphically recorded male WKY rats (300-450 g) weresacrificed when they met the criteria for the following 5 experimentalconditions: 1) spontaneous sleep for 8 h (n=6), 2) short-term (8 h) totalsleep deprivation (n=7), 3) long-term (6-14 d) total sleep deprivation (n=11 pairs), 4) recovery sleep after long-term (7 d) total sleep deprivation(n=5 pairs) and 5) recovery sleep after long-term (7d) REM-selectivesleep deprivation (n=6 pairs). Cellular oxidant production was estimatedin homogenates by monitoring the rate of oxidation from DCFH to flu-orescent 2’-7’ dichlorofluorescein (DCF). Peroxidative damage to cellu-lar lipids was determined by measuring malondialdehyde formation.Protein oxidation was estimated by measuring carbonyl formation using2,4- dinitrophenylhydrazine. Superoxide dismutase activity (CuZn andMn SOD) was determined spectrophotometrically.Results: Long term sleep deprived rats lost 63 ± 4% (mean ± SD) oftheir non-REM sleep and 95 ± 4% of their REM sleep, while the yokedcontrols lost 28 ± 14% and 42 ± 25% of their non-REM and REM sleep,respectively. In the cerebral cortex, oxidant production increased by~25% in short-term sleep deprived rats relative to their sleeping controlsbut did not change in long term sleep deprived rats relative to their yokedcontrols. SOD enzymatic activities, lipid peroxidation levels and proteinoxidation levels in brain, liver and muscle were similar in all experi-mental groups.Conclusions: This study measured the effect of total sleep loss on mark-ers of oxidative stress (oxidant production and antioxidant enzyme activ-ities) as well as on markers of cellular oxidative damage (lipid peroxi-dation and protein oxidation). The results show no evidence of oxidativedamage, nor any indication of oxidative stress as a result of sleep loss,in any of the tissues examined. Therefore, our data do not support thehypothesis that prolonged waking may represent an oxidative challenge.

0025.AWhole-cell Patch-clamp Recordings From Rat Subcoerulean Neu-rons In Vitro: Effects Of Carbachol And OrexinsBrown RE, Thakkar MM, McCarley RWHarvard Medical School/VA Medical Center Brockton, MA

Introduction: The subcoeruleus region of the pontine reticular forma-tion is an important brainstem site involved in the generation of muscleatonia during rapid-eye-movement (REM) sleep. A subpopulation ofneurons in this region enhance their firing during REM, presumably inresponse to excitatory input from cholinergic REM-on neurons in the lat-erodorsal and pedunculopontine tegmentum - injection of cholinergicagonists into the subcoeruleus region elicits signs of REM sleep, includ-ing muscle atonia. One characteristic of the sleep disorder narcolepsy,which is caused by disruption of the orexin/hypocretin system, is peri-ods of muscle weakness triggered by emotional arousal (cataplexy). Wehave recently shown that perfusion of antisense nucleotides directedagainst the type II orexin/hypocretin receptor in the subcoeruleus regionof the rat leads to attacks of cataplexy. At present no data exist regardingthe intrinsic membrane properties of neurons in this region or the mod-ulation of membrane properties of these neurons by cholinergic agonistsor orexins/hypocretins.Methods: Coronal brainstem slices were prepared from 8-16 day-oldrats using a vibratome. Preliminary experiments established that theextensive network of myelinated fibres made it very difficult to slice thebrain tissue or obtain high-quality recordings in rats older than 16 days,whereas in rats younger than 8 days the lack of obvious landmarks madeit difficult to precisely localize the subcoeruleus area. Two brain slicesbetween -9.16 and -9.8 from bregma rostrocaudally (according to theAtlas of Paxinos and Watson) were cut from each animal and placed(after incubation for 1 hr in a prechamber) in a modified Haas-typerecording chamber, where they were perfused with oxygenated (95%Os2s/5 % COs2s) artificial cerebrospinal fluid. Patch electrodes (3-6MO, when filled with a potassium gluconate based intracellular solution)were positioned in the area of the subcoeruleus (subCA, subCD and

SLEEP, Vol. 26, Abstract Supplement, 2003 A10

subCV) using a dissecting microscope. Whole-cell current-clamprecordings were obtained using the ‘blind’ method and drugs applied viathe perfusion line.Results: The effects of the cholinergic agonist carbachol (2 or 10 )and/or the orexins (100, 200 or 500 nM) were tested in 12 neurons in thepresence of the blocker of voltage-dependent sodium channels,tetrodotoxin (0.5 ). Carbachol caused a depolarization (5/9 neurons),hyperpolarization (3/9) or no effect (1/9). All three neurons which werehyperpolarized were located in the more dorsal portion of the sub-coeruleus region. 7 of 11 neurons were excited by orexins, the other 4neurons showed no effect. Orexins depolarized both neurons which werehyperpolarized by carbachol and those which were depolarized.Conclusions: The lack of orexin excitation of the subcoerulean neurons(which we have shown here) in narcolepsy may be responsible for thepreviously reported upregulation and supersensitivity of another maininput to these neurons - the cholinergic projection from the laterodorsaland pedunculopontine tegmentum.

0026.AChronic Hypoxia Leads to Oxidative Stress in Rat CerebellumRamanathan L,1 Gozal D,2 Siegel JM1

(1) Dept. Psychiatry and Biobehavioral Sciences, University of Califor-nia, Los Angeles and VAGLAHS-Sepulveda, North Hills, CA 91343, (2)Kosair Children’s Hospital Research Institute. Dept. Pediatrics, Pharma-cology and Toxicology, University of Louisville School of Medicine,Louisville, KY, 40202

Introduction: It has recently been reported that patients with obstructivesleep apnea (OSA) showed significant gray matter loss in several brainregions including the cortex, hippocampus and cerebellum, suggestingthat sleep apnea results in neuronal damage. Hypoxia is a common fea-ture of OSA. Rats subjected to chronic hypoxia showed increased apop-tosis in the CA1 hippocampal region and cortex. It is proposed thathypoxia induced damage results from oxidative stress. We thereforeinvestigated the effects of chronic hypoxia on several markers of oxida-tive stress. Using biochemical assays, we measured the activities of theantioxidative enzymes superoxide dismutase (SOD), glutathione perox-idase (GPx) and glutathione reductase (GR) as well as the levels of theendogenous antioxidant, total glutathione (GSHt) and the levels of thelipid oxidation product, thiobarbituric acid reactive substances(TBARS).Methods: Young adult (45-50d) male Sprague Dawley rats wereexposed to either chronic intermittent hypoxia (CIH) or chronic sus-tained hypoxia (CSH) for 6h, 1d, 3d, 7d, 14d or 30d. The IH profile con-sisted of alternating room air with 10% oxygen every 90s during thelight phase (6:00AM-6:00PM), while the SH profile consisted of 10%oxygen throughout the light phase. CIH and CSH rats were exposed toroom air throughout the dark phase (6:00PM- 6:00AM), whereas controlrats were placed in room air throughout the 24h cycle. Analysis of vari-ance and linear regression was used to determine statistical significanceat a level of p

tion during active sleep contributes to the reduction of sensory through-put to higher brain centers conveyed via the dorsal spinocerebellar tract.Glycine and GABA may be (co-)released by local segmental inhibitoryinterneurons and/or long projection neurons located in higher brainareas. The neurons releasing glycine and GABA onto Clarke’s columnDSCT neurons may be part of the same neural systems that result inglycine-mediated inhibition of motor outflow.

Research supported by NIH HL41370, MH64109 (JMS) andNS34716, NS32306 (PJS).

0028.AThe Use of Rat Models to Investigate the Behavioral and Neurobio-logical Consequences of Sleep ApneaStrecker RE, Thakkar MM, McKenna JT, McGuire M, Mulkern KJ,Dauphin LJ, Stronge AM, Mello E, Ling L McCarley RWDepts Psychiatry & Medicine, Harvard Med. Sch. & VA Boston Health-care System, Brockton, MA 02301, USA

Introduction: The primary characteristics of obstructive sleep apnea(OSA) are intermittent hypoxia (IH) and sleep interruption (SI) (frag-mentation). The IH and SI are thought to produce the symptoms/signsassociated with OSA which include daytime sleepiness and various cog-nitive/attentional impairments. However, it is not known which symp-toms are attributable to SI and which to IH because these events are dif-ficult to separate in humans. Here we describe our preliminary workwith two rat models that allow the separation of the effects of SI and IH.We hypothesize that the SI associated with OSA, but not the IH, is pri-marily responsible for the symptom of daytime sleepiness seen in OSA,and that this is mediated by adenosine (AD) effects (Strecker et al. Beh.Brain Res. 2000:115 p.183 ). In contrast, we predict that IH treated ratswill sleep normally after the first 2 days of IH exposure (Gozal et al. J.Neurosci. 2001:21 p.2442). We also plan to test the hypothesis that manyof the learning and memory impairments observed in patients with OSAare the result of chronic IH, acting via cell injury/apoptosis in corticalareas involved in cognitive functions.Methods: A) SI paradigm. A rat model of the sleep fragmentation thatoccurs in OSA was produced using an automated rat treadmill operatingon a cycle of 30s on and 90s off, 24 h/day, for 7 days (i.e., 30 brief awak-enings per h). Control groups included an exercise control with equalamounts of time the treadmill is off and on, but with 10 min on and 30min off, allowing deep sleep. B) IH paradigm. Rats were exposed to aschedule of IH to model the hypoxemia of OSA. Daily exposure (12h;07:00 to 19:00; £ 7 days) to continuous sinusoidal cycling of inspiredoxygen levels (10% O2 for 120 s : 21 % O2 for 120 s) was used. We havebegun to investigate the effects of SI and IH at multiple levels, from themolecular to the behavioral.Results: A) SI paradigm. Rats began to sleep during the 90s undisturbedperiods within several h of SI onset, and they continued to awaken dur-ing the 30s treadmill operation for the 7 days tested. During the first 6 hof SI, extracellular AD levels in the basal forebrain rose steadily (~2fold; N=2) resembling the rise previously described with total sleepdeprivation, and a measure that may reflect increased sleep pressure. B)IH paradigm. The IH schedule produced a desaturation to 74 % on day1, resembling the desaturation seen in OSA. Although severe hypoxiahas been reported to increase extracellular AD levels, we found that 70min exposure to the IH regimen produced only a modest rise in basalforebrain AD levels (140 % of baseline; N=3), which quickly returned tobasal levels, even before the IH was terminated.Conclusions: The availability of rat models of OSA will enhance ourunderstanding of the neurobiological mechanisms underlying the symp-toms of OSA, an important step toward future work on the targeteddevelopment of therapies to treat and/or prevent the symptoms of OSA.

Research supported by NIMH 39683 & 01798; Dept Vet. Aff.

0029.ASpontaneous Diurnal And Sleep Deprivation-Induced FluctuationsIn Adenosine Levels In The Rat Basal ForebrainMcKenna JT,1,2 Strecker RE,1,2 Dauphin LJ,1,2 Mulkern KJ,1,2 SintonCM,1,2 Stronge AM,1,2 Tao R,1,2 McCarley RW1,2

(1) Harvard Medical School, Dept. of Psychiatry and Division of SleepMedicine, Boston, MA, (2) Boston VA Healthcare System, Brockton,MA

Introduction: The purine nucleoside adenosine (AD) has been proposedto be an endogenous sleep factor, facilitating sleepiness by inhibitingneurons of the magnocellular basal forebrain (BF). This region isthought to play an important role in the regulation of cortical arousal andthe maintenance of wakefulness. Work in the cat has demonstrated that6h of sleep deprivation (SD) led to a rise in extracellular AD in the BFand cortex. The proposed research extends these findings to the rat, anocturnal species with a diurnal pattern of sleep-wakefulness that moreclosely resembles that of man. We predict that BF AD levels would risespontaneously during the later half of the rat’s active period, correlatingwith the homeostatic sleep drive.Methods: In vivo samples were collected from microdialysis probes inthe BF of Sprague-Dawley rats, and analyzed with a microbore high per-formance liquid chromatography system coupled to UV detection. Diur-nal fluctuation: Rats were kept on a lights on 7 AM/off 7 PM schedule.Beginning at least 14h after probe insertion, 3h samples were collected.Sleep Deprivation: After evaluation of diurnal fluctuation of AD levels,hourly samples were collected, beginning at 10 AM on the day of SD. At1 PM, experimenters kept animals awake for 6h by means of gentle han-dling, followed by 3h of recovery sleep.Results: During the last 6h of the light cycle/sleep period spontaneouslevels of BF AD reached a nadir (1 to 7 PM). As well, BF AD levelspeaked during the second half of the active period (1 to 7 AM). Six h ofSD produced significant increases in BF AD levels, which did not returnto baseline with 3h of recovery sleep.Conclusions: The homeostatic sleep drive occurs daily in the rat with adiurnal pattern of sleep-wakefulness, being maximal at the end of theactive period and minimal at the end of the rest/inactive period. Theseexperiments support a role for AD as an inhibitory sleep factor acting onselect brain regions and arousal-related neuronal subtypes. Additionalwork is needed to confirm that the clear SD-induced increase in BF ADdepends primarily on different sleep and activity levels, and, to a lesserextent, on a circadian influence. Understanding the site-specific mecha-nisms by which AD influences arousal and sleep will benefit researchrelated to various sleep disorders including sleep apnea and insomnia.

Research supported by NIMH 39683 & 01798; Dept Vet. Aff.; NIHFellowship to JTM (T32 HL07901)

0030.AAMPA-Induced Intracellular Calcium Responses are Inhibited byInterferon g in Primary Cultures of Fetal Hippocampal NeuronsDe A, Simasko SM, Krueger JMWashington State University

Introduction: Interferon gamma (IFN g) is a cytokine that enhances nonrapid eye movement sleep (NREMS) in rabbits (Kubota et. al. 2001, J.Neuroimmunol. 119, 192-198). Although IFN g is primarily thought toact as an immune system modulator, IFN g is produced in the centralnervous system by astrocytes and microglia, and IFN g receptors arefound in neurons, astrocytes and oligodendrocytes. Further, IFN g-likeimmunoreactivity is present in some neurons. These observations sug-gest that IFN g may participate in the regulation of sleep. The mecha-nism by which IFN g affects neurons to produce brain mediated actionsis not understood. In the present study we tested the hypothesis that INF

SLEEP, Vol. 26, Abstract Supplement, 2003 A12

g alters the function of AMPA-type ionotropic receptors, an importantmediator of synaptic activity.Methods: Hippocampal neurons were cultured from day 18-20 fetal rats.Isolated hippocampi were washed in Hank’s Balanced Salt Solution andincubated for 15-20 min at 370C. After incubation, cells were mechani-cally dissociated by several gentle aspirations into a 20 cc syringethrough a 22 G needle. Dissociated cells were then centrifuged andresuspended into Dulbecco’s Modified Eagle’s Medium (DMEM) with10% fetal calf serum. The cells were plated onto 100 mg/ml poly-l-ornithine coated 22 mm square glass coverslip in a 6-well plate. Thecells (1-2 x 10cells/coverslip) were grown in DMEM supplemented with10% fetal bovine serum for 2 days and then in DMEM with serum sup-plement (1 mM apotransferrin, 5 mg insulin, 60 nM sodium selenite and100 mM putrescine) in a humidified atmosphere of CO2:air (5:95) at37C. After 9-12 days of culture intracellular Ca++ levels were deter-mined by ratiometric imaging techniques in single neurons by the use offura-2.Results: Acute exposure (2-5 min) to IFN g (10-20 ng/ml) did not alterintracellular Ca++ levels, however, calcium increases induced by AMPA(10 mM) were attenuated by 40 % in the presence of IFN g, an effect thatwas reversible upon IFN g washout. Chronic exposure (30 min to 24 h)to IFN g did not alter subsequent AMPA-induced calcium responsesmeasured in the absence of IFN g. IFN g had no effect on calciumresponses induced by depolarization with high K+.Conclusions: These data suggest that IFN g may alter neuronal activityby an acute action to decrease AMPA receptor function. The relationshipbetween this cellular action of IFN g and the NREMS inducing effectsof IFN g remains to be determined.

This work is supported by NS 25378 (awarded to J.M. Krueger) andNIAAA grant # RO1 AA 13248 (awarded to S.M. Simasko).

0031.AMaintenance Of Wakefulness In Orexin Knockout MiceMochizuki T,1 Crocker A,1 McCormack S,1 Yanagisawa M,2 Sakurai T,3

Scammell TE1

(1) Department of Neurology, Beth Israel Deaconess Medical Center,Boston, MA, (2) Department of Molecular Genetics, University ofTexas, Southwestern Medical Center, Dallas, TX, (3) Dept of Pharma-cology, Institute of Basic Medical Sciences, University of Tsukuba,Tsukuba, Japan

Introduction: The cause of sleepiness in narcolepsy is unknown butmay be due to increased sleep drive, decreased wake drive, or behavioralstate instability, with low thresholds to transition between states. Orexinknockout (KO) mice have a narcolepsy-like phenotype with short boutsof wakefulness, sleep fragmentation, and behavioral arrests resemblingcataplexy (Cell, 1999, 98: 437). We studied orexin-KO mice under avariety of conditions to determine which of these processes is likely tounderlie the sleepiness of narcolepsy.Methods: Male, 3-4 month old, orexin-KO mice (backcrossed withC57BL/6J for 6-8 generations) and wild-type (WT) littermates wereimplanted with EEG and EMG electrodes and temperature/activitytransmitters (TA-F20, DSI). After 2 weeks of recovery, recordings wereperformed on a 12:12 LD cycle (lights on 7am). Behavioral states werescored in 10 second epochs using Sleep Sign software (Kissei Comtec)followed by manual correction.Results: Both groups of mice had similar amounts of wake and NREMsleep, but the orexin-KO mice had more REM sleep during the night,with little diurnal variation. Orexin-KO had more than twice as manybouts of wakefulness, but these were much shorter than in WT. Duringthe dark period, a typical wake bout in WT lasted about 20 minuteswhereas in orexin-KO, wake bouts usually lasted only about 5 minutes.During this period, 60% of wakefulness in WT mice occurred in bouts>21 minutes, but only 20% of wake in orexin-KO mice occurred in bouts

of this length. To determine whether these short wake bouts might be dueto decreased waking drive, we transferred mice to new cages at 1pm, andboth groups remained awake for about 45 minutes. To determinewhether orexin-KO mice might have altered sleep homeostasis, wedeprived orexin-KO and WT mice of sleep for 2-8 hours using gentlehandling. Both groups had normal amounts of recovery NREM sleepwith a clear increase in the number and duration of NREM bouts, thoughorexin-KO mice still had some sleep fragmentation.Conclusions: These experiments help define the mechanisms resultingin poor maintenance of wakefulness in orexin-KO mice. In contrast tomice lacking histamine (Jour Neurosci, 2002, 22:7695), the ability oforexin-KO mice to fully maintain wakefulness in a new environmentsuggests that their wake-promoting systems are relatively intact. Anincrease in homeostatic sleep drive also appears unlikely because orex-in-KO mice have normal NREM rebound after sleep deprivation. Theirnormal amounts of wake and NREM sleep also makes these processesunlikely. Orexin may play an essential role in stabilizing wakefulness,and the behavior of these mice may be a consequence of behavioral stateinstability, with low thresholds to transition between wake and sleep.

Research supported by NIH grants MH62589, MH068785,MH01507, and HL60292.

0032.ASleep Dependence Of The High Frequency (600Hz) EvokedResponse In Rat Sensory CortexRector DM, Carter KM, Dansereau RPWashington State University

Introduction: During the early phase of cortical evoked responses, abrief burst of low-amplitude, high frequency (300-1000Hz) spike-likewavelets are superimposed on the post-synaptic response. Thesewavelets most likely arise from thalamo-cortical circuits involvinginhibitory feedback mechanisms within the cortical layers. Only a fewinvestigators have looked for the high frequency component of theevoked response, and this component has not been adequately assessedin rodents across sleep states. Additionally, local neural groups mayexhibit sleep like states semi-independent of each other and organismsleep state, and changes in the burst patterns could be used to measure aparticular neural group’s state. We expect to see state dependent changesin the burst patterns, which may be independently observable across dif-ferent cortical columns.Methods: The sleep dependence of the high frequency responses wereinvestigated in Sprauge-Dawley rats with cortical screw electrodes overthe somatosensory cortex. Stimulation pulse, EKG, respiration, neckEMG, and four cortical EEG signals were continuously recorded (5ksps)during 6h periods in the middle of the light cycle. Animals were contin-uously presented with 0.2ms auditory clicks at random intervals (1-2sISI). Each evoked response was filtered (300-1000Hz) and grouped intostate. Averages were performed on blocks of 60 or more seconds of eachstate to provide a quantitative measure of the burst pattern across state.Results: High frequency bursts were observed during the early part ofall evoked responses, regardless of sleep state. However, the peak topeak amplitude of the burst during quiet sleep was 30% larger and last-ed twice as long when compared to the bursts generated during waking.Bursts during REM sleep were of similar amplitude as that during wak-ing, but with a longer duration as seen during quiet sleep.Conclusions: Since our electrodes were placed over the somatosensorycortex with the reference screw located caudal to the auditory cortex, wewould expect any responses to result from the entire volume of the audi-tory cortical tissue. Thus, our results are consistent with the findings ofother investigators who record evoked potentials 1 to 2mm deep to thesurface. The larger burst amplitude during quiet sleep probably coin-cides with the larger post-synaptic potentials also occurring during thisstate, and may reflect reduced inhibition of the thalamic input. Similar-

A13 SLEEP, Vol. 26, Abstract Supplement, 2003

ly, the longer lasting burst duration during both quiet sleep and REMsleep may result from an oscillatory circuit that is not as strongly inhib-ited during sleep. We propose that the circuitry responsible for generat-ing both the post-synaptic and high frequency burst response can beinfluenced at the local level (e.g. cortical columns) by sleep promotingsubstances, and thus the particular high frequency burst patternsobserved may provide a marker for the state of the individual neuralgroups.

This work was generously supported by the 2002 Sleep ResearchSociety J. Christian Gillin Junior Faculty Award and by startupfunds from the WSU VCAPP Department.

0033.ACardiac Cycle Influence On The Sensory Evoked Response In RatDansereau RP, Carter KM, Rector DMWashington State University

Introduction: Few investigators have examined the relationship of thecardiac cycle and sensory acquisition, and those studies have createdsomewhat of a controversy on the behavioral significance of cardiaccycle and observed relationships. Early studies may suggest that EEGactivity is related to events within the cardiac cycle, and that there maybe a baroreceptor influence on the processing of sensory information.However, details of the observed relationships were not examined indepth, and it is difficult to conclude significance from these earlyreports. A cardiac influence on the evoked response has profound impli-cations on the interpretation of trial based stimulation experiments, espe-cially if the trials occur at regular intervals. We examined the temporalrelationship between the cardiac cycle and evoked activity within thesensory cortex.Methods: The relationship between the evoked response and cardiovas-cular activity was examined by recording whisker and auditory stimu-lated evoked potentials from the sensory cortex of the rat, under bothketamine/xylazine anesthesia and freely behaving conditions. An ampli-fier and computer data acquisition system continuously recorded eightchannels (Stimulation pulse, EKG, Respiration, Neck EMG, and fourcortical EEG signals at 5ksps). During the recordings, animals were con-tinuously presented with 0.2 ms auditory clicks or contralateral whiskerstimulation at random intervals (1-2s ISI). Each evoked response wasgrouped into 2ms time bins relative to the peak of the EKG R-wave. Inchronic studies, evoked responses were also separated by sleep state.Results: Under anesthesia, we observed a 20% reduction in the averagedevoked response to whisker stimulation, when the peak of the evokedresponse occurred approximately half way between EKG R-waves. Inthe chronic experiments, whisker stimulation is currently not possibleduring natural sleep states, and the chronic auditory evoked responsewas small (10%) compared to the whisker stimulation; thus, significantvariability in the evoked potential precluded conclusive results. Howev-er, there is some suggestion that the 20% reduction in evoked potentialis preserved during quiet waking, but reversed in direction during quietsleep.Conclusions: The observed relationship between stimulus deliverywithin specific time periods of the cardiac cycle, and patterns related tostate could have a large impact on how evoked potential data should bemost efficiently acquired in future experiments. Phase coupling of EKGwith a regular ISI, or phase locking of the cardiac cycle with a regularstimulus would affect the amplitude of evoked responses. Additionally,trial by trial analysis of individual evoked potentials show a significantamount of variability, which at least in part, could be due to the phase ofthe cardiac-respiratory cycle during which a stimulus was presented.Previous publications have demonstrated the significance of the cardiaccycle within behavioral studies; thus, further investigation is prudent tofurther examine these relationships.

This work was generously supported by the 2002 Sleep ResearchSociety J. Christian Gillin Junior Faculty Award and by startupfunds provided by the VCAPP Department at WSU.

0034.ASleep Dependence Of Auditory Evoked Potentials In Rat SensoryCortexCarter KM, Dansereau RP, Rector DMWashington State University

Introduction: Studies of cortical evoked responses in rat show a com-plex series of waveforms from population post-synaptic potentials andpopulation spikes. Many studies have shown that the amplitude of theevoked potential is state dependent. However, the state related differ-ences vary significantly between studies presumably based on animalmodel differences and electrode placement. Few studies have adequate-ly addressed sleep related evoked potential changes in the rat model.Additionally, within a single experiment there are significant differencesin the amplitude of the evoked response across stimuli. We address atleast some of the variability by examining spontaneous fluctuations inthe auditory evoked potential continuously across long periods ofsleep/wake states and plotting the responses across time.Methods: Continuous 3 to 6 hour recordings were acquired from freelybehaving rats in their home cage with recording screws placed over thesensory cortex. Four channels of cortical EEG and one channel each ofEKG, respiration and neck EMG were recorded while providing 1-2 sec-ond random interval click stimuli. Sleep state was determined by post-hoc analysis and time triggered evoked potentials were divided into stateaverages. Individual trials were also plotted against each other in con-tinuous time series plots.Results: We observed a 36% increase in the amplitude of the largestpeak of the evoked potential during slow wave sleep when compared towaking. REM was not significantly different but showed an additionalcomponent at 35 ms which was not present during slow wave sleep orwaking. Changes in the evoked potential across states were easily rec-ognized by following the shape of the evoked response across continu-ous plots of time triggered evoked potentials. By plotting the peakamplitude of the evoked potential, we also observed spontaneous fluctu-ations in the auditory evoked responses of local field potentials in indi-vidual brain regions during sleep states with a period of 60 seconds.These fluctuations were different across corresponding contralateralregions and organism sleep state suggesting an independent and local-ized origin.Conclusions: The differences in amplitude of the evoked potentialacross sleep states could be a result of relaxed thalamic control of theinput to the cortex during slow wave sleep. There were significant dif-ferences in the temporal pattern of the periodic fluctuations across brainregions suggesting that the signals were unique to a particular brain areaand not a product of a brain wide phenomenon. We believe that thesefluctuations could represent the state changes of individual neuralgroups that are independent from the organism as a whole.

This work is supported by the 2002 Sleep Research Society, J. Chris-tian Gillin Junior Faculty Award and by startup funds provided bythe VCAPP Department at WSU.

0035.AGlucocorticoids Mediate The Effects Of Sleep Deprivation On BrainGlycogen and Glucose LevelsRuby NF, Gip P, Hagiwara G, Sapolsky R, Cao V, Franken P, Heller HCStanford University

Introduction: Glycogen is the largest energy reserve in the brain andlevels decrease in some brain regions after prolonged sleep deprivation

SLEEP, Vol. 26, Abstract Supplement, 2003 A14

(SD). The variation in glycogen levels among brain regions after SDmay be due in part to secretion of glucocorticoids during SD. The glu-cocorticoid, corticosterone (CORT), is markedly elevated during SD andglucocorticoid activity varies greatly by brain region. Because CORTinhibits glucose uptake in central and peripheral tissues, we hypothe-sized that elevations in CORT during SD would facilitate utilization ofbrain glycogen stores and that eliminating endogenous CORT wouldprevent reductions in brain glycogen after SD.Methods: Adrenalectomized (ADX, n=9) and intact (n=8) 34 day oldrats were subjected to a 6-h SD by gentle handling beginning at the dailytime of lights-on. Rats were immediately sacrificed after SD bymicrowave irradiation (3.5 kW/1.7 sec). Tissue samples from the cortex,cerebellum, brainstem, hippocampus, and liver were assayed for glyco-gen and glucose content. Plasma was obtained to assay for CORT andglucose levels.Results: SD reduced glycogen levels in the cerebellum and hippocam-pus by 23% (P

intracerebroventricular (icv) infusion into the third cerebral ventricle ofnewly developed OX2R selective agonist, [Ala11]orexin-B on the sleep-waking cycle in rats.Methods: Male Sprague-Dawley rats were chronically implanted withcortical EEG and neck EMG electrodes. A stainless steel cannula for icvinfusion was chronically inserted into the third ventricle. EEG and EMGwere monitored for three consecutive days, during continuous icv salineinfusion at a rate of 10 ml/h. For 5-h diurnal period, [Ala11]orexin-B (1,10, 40 nmol/50ul saline) replaced the icv infusion of saline.Results: [Ala11]orexin-B at the dose of 10 nmol markedly increased theamount of wakefulness by 153.7 % (n=4, p

Research supported by The Swiss National Science Foundationgrant 100-053005.97 and the Human Frontiers Science Programgrant RG-0131/2000-BR102

0040.AEffect Of Bicuculline Perfused In DRN On 5-HT Levels In The DRNAnd FCFiske E, Grønli J, Hamre F, Bjørkum AA, Portas CM, Ursin RUniversity of Bergen, Norway

Introduction: The suppression of dorsal raphe neuron (DRN) seroton-ergic cell firing has been proposed to be an initiator of REM sleep.GABAergic modulation of the serotonergic neurons may be important insuch generation of REM sleep and perfusion of a GABA agonist intoDRN increase REM sleep. The present study