Thien Thanh Dang-Vu and Sophie Schwartz - UNIGE

Chapter Functional neuroimaging of narcolepsy

27 Thien Thanh Dang-Vu and Sophie Schwartz

IntroductionFunctional neuroimaging techniques used to study narcolepsyinclude single-photon emission computed tomography (SPECT),positron emission tomography (PET), and functional magneticresonance imaging (fMRI). SPECT shows the distribution ofradioactive isotopes, the decay of which is associated with theemission of detectable single gamma photons. Examples ofSPECT isotopes are 99mtechnetium-hexamethylpropylene amineoxime (99mTc-HMPAO) and 99mtechnetium-ethyl cysteinatedimer (99mTc-ECD), both indirect markers of regional cerebralblood flow (rCBF). PET shows the distribution of compoundslabeledwithpositron-emitting isotopes, suchas [15O]-labeledwater(H2

15O), an indirect marker of rCBF, and 18F-fluorodeoxyglucose(18F-FDG), amarker of glucosemetabolism (CMRglu). FunctionalMRI measures the variations in brain perfusion related to neuralactivity, using a method based on the assessment of the BOLD(bloodoxygen level-dependent) signal. SPECTandPETcan alsobecoupled with synthetic ligands to specific receptors of interest, inorder to investigate neuromodulatory changes associated with acondition.

In this chapter, we will review functional brain imaging studiesconducted in narcoleptic patients and evaluating different patho-physiological aspects of the disorder : (1) neurotransmission stud-ies targeting the cholinergic, serotonergic, and dopaminergicsystems (PET/SPECT); (2) the distribution of brain activity acrossthe sleep-wake cycle (PET/SPECT); (3) the neural circuits involvedin emotional and reward processing (fMRI).

Neurotransmission in narcolepsyThe role of various neurotransmitters in the pathophysiology ofnarcolepsy has been explored using PET or SPECT coupledwith specific ligands: acetylcholine (ACh), serotonin (5-HT),and dopamine (DA). Results are summarized in Table 27.1.

Only one study evaluated ACh function in narcolepsy. PETcoupled with [11C]N-methyl-4-piperidyl-benzilate (11C-NMPB)was used to target muscarinic ACh receptors. No difference inmuscarinic ACh binding was found in the pons, thalamus,striatum, and cerebral cortex of 11 patients compared with 21controls [1].

Likewise there is a single work assessing 5-HT neurotrans-mission in narcolepsy. PET with 2′-methoxyphenyl-(N-2′-pyridinyl)-p-18F-fluoro-benzamidoethylpiperazine (18F-MPPF)

was employed to study 5-HT1A receptors. This study conductedon 14 patients showed an increase of 5-HT1A binding – partic-ularly in the anterior cingulate, temporal, and mesio-temporalcortices – during sleep compared with wakefulness [2]. However,no control group was recruited, which prevents from confirmingthe specificity of this result to narcolepsy.

More interest was devoted to DA neurotransmission in nar-colepsy. Several studies are available with either PET or SPECT.Studies of presynaptic DA transporter binding converge to dem-onstrate no significant modification in narcoleptic patients, eitherwith [123I](N)-(3-iodopropene-2-yl)-2β-carbomethoxy-3β-(4-chlorophenyl) tropane (123I-IPT) SPECT [3] or 11C-2β-carbomethoxy-3β-(4-fluorophenyl) tropane (11C-CFT) PET [4].As for postsynaptic D2-receptor binding, only one study found asignificant change between patients and controls: SPECT using123I-iodo benzamide (123I-IBZM) demonstrated a D2 bindingincrease in the striatum of seven narcoleptic patients [3]. Inaddition, a positive correlation was observed between striatal D2

binding and the incidence of sleep attacks and cataplexy.However,other SPECT studies with IBZM [5, 6], as well as PET studies with[11C]raclopride [7, 8] or N-(3-[l8F]fluoropropyl)-spiperone(FPSP) [9] did not confirm striatal changes in D2 binding.

Altogether, available neuroimaging data do not consistentlysupport the involvement of a neuromodulatory deficit in ACh,5-HT, or DA systems for the pathophysiology of narcolepsy.

Brain perfusion and glucose metabolismin narcolepsySeveral functional neuroimaging studies were conducted toevaluate the distribution of brain activity during wakefulnessand sleep in narcolepsy. Three of them described CMRglu[10, 11] and rCBF [12] patterns during resting wakefulness.Only preliminary reports compared rCBF values between sleepand waking [13, 14].

In a first PET study with 18F-FDG PET, CMRglu duringbaseline wakefulness was compared between 24 narcolepticpatients and 24 healthy subjects [10]. Significant decreaseswere observed in the posterior hypothalamus and medio-dorsalthalamus of patients. In this study, three patients had narco-lepsy without cataplexy, and four were treated with medications(stimulants, antidepressants). In addition, no electroencephalo-graphic (EEG) recording was carried out to objectively monitor

Neuroimaging of Sleep and Sleep Disorders, ed. Eric Nofzinger, Pierre Maquet, and Michael J. Thorpy. Published byCambridge University Press. © Cambridge University Press 2013.

223

the subjects’ vigilance state during the study. To address theselimitations, the same group then conducted a 99mTc-ECDSPECT study during wakefulness in 25 patients and the samenumber of controls: this time, all patients had a history ofcataplexy and no history of pharmacological treatment (forsleepiness or cataplexy) [12]. Moreover, EEG was availableduring the study to ensure that the subjects were fully awakeduring the procedure. The study found decreased rCBF in thehypothalamus and thalamus, in line with the 18F-FDG PETstudy. Additional significant decreases were also observed inthe caudate, superior/middle frontal gyri, postcentral gyrus,parahippocampal gyrus, and cingulate cortex (Figure 27.1). Amore recent study by another group used PET with 18F-FDGduring wakefulness in 21 patients with narcolepsy and cata-plexy, and 21 matched controls. In contrast with the two pre-vious studies, no significant CMRglu decrease was found.Instead, increases were observed in the anterior and mid cingu-late cortex, and in the right visual association cortex. As in theearlier 18F-FDG PET study, the two major limitations of theseresults are the inclusion of patients treated with psychostimu-lants and/or anticataplectic drugs (14 out of 21), and theabsence of objective (EEG) monitoring of vigilance state.

In brief, with the exception of one 18F-FDG PET study thatfound increased CMRglu particularly in the cingulate cortex[11], other functional neuroimaging studies during wakefulnessconfirmed a hypothalamic dysfunction in narcoleptic patients,in agreement with a deficit in the hypocretin system [10, 12].Other hypoperfusions (thalamus, caudate, prefrontal, and post-central cortices, limbic areas) might relate to clinical featuresassociated with narcolepsy, such as altered emotional process-ing (see below) and attentional deficits.

As for the study of brain activity during sleep, early studiesusing 133Xe inhalation showed that narcoleptics compared withcontrols had a lower rCBF during wakefulness mostly in thebrainstem-cerebellum region, while at sleep onset compared towakefulness rCBF decreased in controls but increased in narco-leptic patients in the same region [13, 15]. In contrast, a99mTc-HMPAO SPECT study did not find any rCBF differencebetween wakefulness and rapic eye movement (REM) sleep in

narcoleptic patients [14]. As this last study did not include acontrol group, the specificity of this result to narcolepsy couldnot be established. Future studies, using state-of-the-art functionalneuroimaging techniques, should reevaluate brain activity

Table 27.1. SPECT- and PET-ligand studies in narcolepsy

Study Imaging Target Number of pat./ctrl. Treatment Results

Sudo et al. [1] PET 11C-NMPB ACh (muscarinic) 11/21 None No change

Derry et al. [2] PET 18F-MPPF 5HT1A 14/0 12/14 N/A (no control group)

Eisensehr et al. [3] SPECT IPT DA (transporter) 7/7 None No change

Rinne et al. [4] PET 11C-CFT DA (transporter) 10/15 None No change

Eisensehr et al. [3] SPECT 123I-IBZM DA (D2) 7/7 None Increase in striatum

Hublin et al. [5] SPECT 123I-IBZM DA (D2) 6/8 None No change

Staedt et al. [6] SPECT 123I-IBZM DA (D2) 10/10 None No change

Khan et al [7] PET [11C]raclopride DA (D2) 17/32 12/17 No change

Rinne et al. [8] PET [11C]raclopride DA (D2) 7/7 6/7 No change

MacFarlane et al. [9] PET FPSP DA (D2) 6/6 None No change

N/A = not available.

Figure 27.1 Brain perfusion decreases during wakefulness innarcolepsy. Brain mapping of the regions where cerebral perfusionis decreased in narcoleptic patients compared to normal subjects(99mTc-ECD SPECT). Results are overlaid on T1 MRI, and aresignificant at FDR-corrected P < 0.05. (A) Significant hypoperfusionwas observed in bilateral anterior hypothalami (arrowhead) and inthe right parahippocampal gyrus (short arrow). Bilateral cingulategyri and white matters in bilateral middle frontal gyri (long arrow)showed decreased cerebral perfusion. (B) Hypoperfusion was evidentin bilateral posterior thalami (arrowhead) and in the white mattersof the bilateral postcentral and supramarginal gyri (short arrow).(C) In the sagittal view of right hemisphere, significant hypoperfusionwas observed in the caudate nucleus (arrowhead), in the subcallosalgyrus (short arrow), the cingulate gyrus extending alongcorpus callosum (long arrow), and in the parahippocampal gyrus (dottedarrow). (D) 3-dimensional rendering view showing decreased cerebralperfusion in bilateral paracentral areas (arrowhead) and superior/middlefrontal gyri (short arrow). The frontal lobe is on the right and theoccipital lobe on the left. (Reprinted from Neuroimage; Vol. 28(2); Yeon JooE, Hong SB, Tae WS, Kim JH, Han SJ, Cho YW, Yoon CH, Lee SI, Lee MH,Lee KH, Kim MH, Kim BT, Kim L. “Cerebral perfusion abnormalityin narcolepsy with cataplexy”; pp 410–16; Copyright 2005, with permissionfrom Elsevier.)

Section 5: Neuroimaging of sleep disorders

224

patterns during the different stages of sleep in narcoleptic patientscompared with controls.

Neural correlates of emotional processingin narcolepsyProcessing of emotional information potentially plays an impor-tant role in narcolepsy–cataplexy. Indeed it is well known thatemotions, particularly those with a positive component (jokes,laughter, etc.), can trigger cataplectic episodes. Functional MRIstudies therefore assessed brain responses to humorous stimuli innarcoleptic patients. In a first study, 12 narcoleptics with cataplexyand 12 controls were scanned during presentation of humorousand neutral pictures [16]. In patients, pharmacological treatmentwas discontinued for at least 14 days before the fMRI session.Humorous pictures (compared to neutral pictures) were associ-ated with an increased amygdala response together with adecreased response in the hypothalamus of patients compared tocontrols (Figure 27.2). In a second fMRI study, a similar paradigmwas used in 10 narcoleptic patients with cataplexy compared to 12healthy controls [17]. Medications were stopped for at least fivedays prior to the experiment. In agreement with the previousstudy, higher brain response to humorous cartoons were observedin several areas including the amygdala (as well as the inferiorfrontal gyrus, superior temporal gyrus, insula, nucleus accum-bens) in narcoleptics compared to controls. However, fMRIresponse in the hypothalamus to humorous stimulation was notfound decreased in this study, but rather increased.

Not only positive emotional stimulation was assessed withfMRI in narcolepsy. Brain responses to unpleasant stimuli werealso investigated in nine unmedicated narcoleptic patients withcataplexy and nine matched controls [18]. The task consisted inan aversive conditioning paradigm (visual conditioning stimuli

and painful electrical stimulation). Results showed increasedneural responses to conditioned stimuli in the amygdala andincreased functional connectivity between the amygdala andmedial prefrontal cortex in the control group but not in thenarcolepsy group.

Altogether these studies suggest a dysfunction of amygdalo-hypothalamic and amygdalo-neocortical interactions duringthe processing of emotional information in narcolepticpatients, possibly underlying central mechanisms of cataplexy.

Neural correlates of reward processing innarcolepsyAnticipation of reward (e.g., when playing games) constitutes aparticular emotional experience prone to trigger cataplexy inhumans [19], which suggests a potential involvement of thehypocretin system in reward brain circuits, and possible alter-ations of these circuits in narcolepsy with cataplexy.Accordingly, data in rodents show a close interplay betweenhypocretin neurons and reward-related brain regions, such asthe nucleus accumbens and the DA ventral tegmental area [20–22]. In order to further evaluate a potential dysfunction ofreward-related neural processes, an fMRI study was conductedin 12 unmedicated narcoleptic patients with clear-cut cataplexyand 12 matched healthy controls, while performing a taskinvolving the mesolimbic and midbrain reward system [23].This task consisted of a modifed version of a monetary incen-tive delay task. Brain responses to high motivational cuesincluded the ventral tegmental area in the control group, butnot in the narcolepsy group. Responses to successful trials(compared to failed trials) revealed increased activity in theventromedial prefrontal cortex and nucleus accumbens in thecontrol group but not in the patients group. Likewise, increased

Controls > NC patientsA

B

Hypothalamus: 12x, 3y, –18z

Amygdala: 33x, 3y, –21z

humor

fMR

I effe

ct s

ize

neutral

humor

neutral

PatientsControls

PatientsControls

fMR

I effe

ct s

ize 4

3210

–1–2

543210

–1–2–3–4

NC patients > Controls

5

4

3

5

4

3

t–value

t–value

Hypothalamus: 12x, 3y, –18z

Amygdala: 33x, 3y, –21z

fMR

I effe

ct s

ize

PatientsControls

PatientsControls

fMR

I effe

ct s

ize 4

3210

–1–2

543210

–1–2–3–4

NC patients > Controls

5

4

3

5

4

3

–value

t–value

Figure 27.2 Neural correlates of emotionalprocessing in narcolepsy. Functional MRIresponse is decreased in the hypothalamus (A)and increased in the amygdala (B) duringpresentation of humorous pictures comparedto neutral pictures, and more so in narcolepticpatients than in healthy controls (p < 0.001).(Adapted from Brain; Vol. 131(Pt2); S. Schwartzet al., “Abnormal activity in hypothalamus andamygdala during humour processing in humannarcolepsy with cataplexy” ; pp 514–522;Copyright 2008, with permission from OxfordUniversity Press.)

Chapter 27: Functional neuroimaging of narcolepsy

225

brain responses to successful positively cued trials were foundin the nucleus accumbens and lateral prefrontal cortex in con-trols; in narcoleptics increased responses to these trials werefound in the amygdala, consistent with reports of increasedamygdala response to stimuli associated with highly positiveemotions (see above). Finally, in the narcolepsy group, signifi-cant positive correlations were found between disease durationand fMRI responses to high motivational cues in the nucleusaccumbens and ventromedial prefrontal cortex (Figure 27.3).Altogether these findings show evidence for a disruption of

neural circuits involved in reward processing in narcolepsy.Furthermore, these data suggest a progressive functional recov-ery of reward-related brain structures in narcoleptic patientswith longer disease duration.

ConclusionFunctional brain imaging studies in narcolepsy can be sum-marized as follows:

1. Narcolepsy is not associated with a specific alteration of thecentral cholinergic or dopaminergic activity.

2. Functional brain activity patterns of narcoleptic patientsduring resting wakefulness are characterized byabnormalities located in the hypothalamus – in agreementwith a loss of hypocretinergic neurons in this disease – aswell as in various cortical areas, possibly in relation withcognitive and attentional deficits encountered by thesepatients.

3. Altered emotional processing associated with cataplexy alsoinvolves a dysfunction of the hypothalamus, in addition toneural changes within limbic structures, in particular theamygdala.

4. Narcolepsy with cataplexy finally involves a dysfunction ofneural circuits implicated in reward processing andencompassing the nucleus accumbens and the midbrainventral tegmental area.

Further studies should investigate more closely brain activitychanges across the sleep/wake cycle in narcoleptic patients.Indeed narcolepsy not only induces severe daytime symptoms,but is also frequently associated with sleep disruption, includingchanges in sleep microarchitecture [24, 25].

AcknowledgementsThis research was supported by the Fonds National de laRecherche Scientifique (Belgium), the Swiss National ScienceFoundation, the Fonds Léon Frédéricq (Belgium), the BelgianCollege of Neuropsychopharmacology and BiologicalPsychiatry, and the Canadian Institutes of Health Research.

References1. Sudo Y, Suhara T, Honda Y, et al.

Muscarinic cholinergic receptors inhuman narcolepsy: a PET study.Neurology. 1998;51(5):1297–302.

2. Derry C, Benjamin C, Bladin P, et al.Increased serotonin receptoravailability in human sleep: evidencefrom an [18F]MPPF PET study innarcolepsy. Neuroimage. 2006;30(2):341–8.

3. Eisensehr I, Linke R, Tatsch K, et al.Alteration of the striatal dopaminergicsystem in human narcolepsy. Neurology.2003;60(11):1817–19.

4. Rinne JO, Hublin C, Nagren K, HeleniusH, Partinen M. Unchanged striataldopamine transporter availability innarcolepsy: a PET study with[11C]-CFT. Acta Neurol Scand.2004;109(1):52–5.

5. Hublin C, Launes J, Nikkinen P,Partinen M. Dopamine D2-receptors inhuman narcolepsy: a SPECT study with123I-IBZM. Acta Neurol Scand. 1994;90(3):186–9.

6. Staedt J, Stoppe G, Kogler A, et al. [123I]IBZM SPET analysis of dopamine D2receptor occupancy in narcoleptic

patients in the course of treatment. BiolPsychiatry. 1996;39(2):107–11.

7. Khan N, Antonini A, Parkes D, et al.Striatal dopamine D2 receptors inpatients with narcolepsy measured withPET and 11C-raclopride. Neurology.1994;44(11):2102–4.

8. Rinne JO, Hublin C, Partinen M, et al.Positron emission tomography study ofhuman narcolepsy: no increase instriatal dopamine D2 receptors.Neurology. 1995;45(9):1735–8.

9. MacFarlane JG, List SJ, Moldofsky H,et al. Dopamine D2 receptors quantified

Figure 27.3 Neural correlates of reward processing in narcolepsy. FunctionalMRI responses to a monetary incentive delay task in narcoleptic patients. Brainresponses to high motivational cues in the nucleus accumbens (A) andventromedial prefrontal cortex (ventromedial PFC; B) are positively correlatedwith disease duration (r2 = 0.85 and 0.82, respectively; p < 0.001). (Adapted fromAnnals of Neurology; Vol. 67(2); A. Ponz et al., “Abnormal activity in reward braincircuits in human narcolepsy with cataplexy” ; pp 190–200; Copyright 2010, withpermission from John Wiley and Sons.)

Section 5: Neuroimaging of sleep disorders

226

in vivo in human narcolepsy. BiolPsychiatry. 1997;41(3):305–10.

10. Joo EY, TaeWS, Kim JH, Kim BT, HongSB. Glucose hypometabolism ofhypothalamus and thalamus innarcolepsy. Ann Neurol. 2004;56(3):437–40.

11. Dauvilliers Y, Comte F, Bayard S, et al. Abrain PET study in patients withnarcolepsy-cataplexy. J NeurolNeurosurg Psychiatry. 2010;81(3):344–8.

12. Yeon Joo E, Hong SB, Tae WS, et al.Cerebral perfusion abnormality innarcolepsy with cataplexy. Neuroimage.2005;28(2):410–16.

13. Meyer JS, Sakai F, Karacan I, Derman S,Yamamoto M. Sleep apnea, narcolepsy,and dreaming: regional cerebralhemodynamics. Ann Neurol. 1980;7(5):479–85.

14. Asenbaum S, Zeithofer J, Saletu B, et al.Technetium-99m-HMPAO SPECTimaging of cerebral blood flow duringREM sleep in narcoleptics. J Nucl Med.1995;36(7):1150–5.

15. Sakai F, Meyer JS, Karacan I, YamaguchiF, Yamamoto M. Narcolepsy: regionalcerebral blood flow during sleep andwakefulness.Neurology. 1979;29(1):61–7.

16. Schwartz S, Ponz A, Poryazova R, et al.Abnormal activity in hypothalamus andamygdala during humour processing inhuman narcolepsy with cataplexy.Brain. 2008;131(Pt 2):514–22.

17. Reiss AL, Hoeft F, Tenforde AS, et al.Anomalous hypothalamic responses tohumor in cataplexy. PLoS One. 2008;3(5):e2225.

18. Ponz A, Khatami R, Poryazova R, et al.Reduced amygdala activity duringaversive conditioning in humannarcolepsy. Ann Neurol. 2010;67(3):394–8.

19. Anic-Labat S, Guilleminault C, KraemerHC, et. al. Validation of a cataplexyquestionnaire in 983 sleep-disorderspatients. Sleep. 1999;22(1):77–87.

20. Fadel J, Deutch AY. Anatomicalsubstrates of orexin-dopamineinteractions: lateral hypothalamic

projections to the ventral tegmentalarea. Neuroscience. 2002;111(2):379–87.

21. Narita M, Nagumo Y, Hashimoto S,et al. Direct involvement of orexinergicsystems in the activation of themesolimbic dopamine pathway andrelated behaviors induced by morphine.J Neurosci. 2006;26(2):398–405.

22. Korotkova TM, Sergeeva OA, ErikssonKS, Haas HL, Brown RE. Excitation ofventral tegmental area dopaminergic andnondopaminergic neurons by orexins/hypocretins. J Neurosci. 2003;23(1):7–11.

23. Ponz A, Khatami R, Poryazova R, et al.Abnormal activity in reward brain circuitsin human narcolepsy with cataplexy. AnnNeurol. 2010;67(2):190–200.

24. Bove A, Culebras A,Moore JT,WestlakeRE. Relationship between sleep spindlesand hypersomnia. Sleep. 1994;17(5):449–55.

25. Khatami R, Landolt HP, Achermann P,et al. Insufficient non-REM sleepintensity in narcolepsy-cataplexy. Sleep.2007;30(8):980–9.

Chapter 27: Functional neuroimaging of narcolepsy

227