Thesis

The Sleepy Brain The Past, Present and Future of Central Disorders of Hypersomnolence Jari Kylian Gool

Publication of this thesis was financially supported by the Amsterdam UMC Location VUmc, Bioprojet Benelux NV, the Graduate School Neurosciences Amsterdam Rotterdam (ONWAR) and the Nederlandse Vereniging voor Slaap- en Waakonderzoek (NSWO). Printing: Ipskamp Printing | ipskampprinting.nl Cover art: Alexander Fasel & Jari Gool Layout and design: Jacolijn de Krom | persoonlijkproefschrift.nl ISBN: 978-94-6473-669-4 © Jari Kylian Gool, Amsterdam 2025, The Netherlands. All rights reserved. No part of this thesis may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval, without prior written permission of the author.

VRIJE UNIVERSITEIT THE SLEEPY BRAIN The Past, Present and Future of Central Disorders of Hypersomnolence ACADEMISCH PROEFSCHRIFT ter verkrijging van de graad Doctor of Philosophy aan de Vrije Universiteit Amsterdam, op gezag van de rector magnificus prof.dr. J.J.G. Geurts, in het openbaar te verdedigen ten overstaan van de promotiecommissie van de Faculteit der Geneeskunde op vrijdag 24 januari 2025 om 11.45 uur in een bijeenkomst van de universiteit, De Boelelaan 1105 door Jari Kylian Gool geboren te Leiderdorp

promotoren: prof.dr. Y.D. van der Werf prof.dr. G.J. Lammers copromotoren: dr. R. Fronczek prof.dr. T.T. Dang-Vu promotiecommissie: prof.dr. O.A. van den Heuvel prof.dr. J.T. van Dissel prof.dr. J.J.G.M. Verschuuren dr. S. Knudsen-Heier dr. F. Siclari dr. M.G.M. Huijbers prof.dr. E.J.W. van Someren

Table of contents Chapter 1 General Introduction 9 Section A – The Past: Identifying Potential Immunological Triggers for Central Disorders of Hypersomnolence Chapter 2 New 2013 Incidence Peak in Childhood Narcolepsy: More Than Vaccination? Sleep. 2021;44(2):zsaa172. 43 Chapter 3 2010 and 2013 Incidence Peaks in Narcolepsy and Idiopathic Hypersomnia Linked to Type A H1N1 and Type B Victoria Influenza Strains Submitted to: Proceedings of the National Academy of Sciences. 63 Chapter 4 Potential Immunological Triggers for Narcolepsy and Idiopathic Hypersomnia: Real-World Insights on Infections and Influenza Vaccinations Sleep Medicine. 2024;116:105–14. 91 Section B – The Present: Brain Structure and Functioning in Narcolepsy Type 1 Chapter 5 Widespread White Matter Connectivity Abnormalities in Narcolepsy Type 1: A Diffusion Tensor Imaging Study Neuroimage Clinical. 2019;24:101963. 131 Chapter 6 Widespread White Matter Axonal Loss in Narcolepsy Type 1 Submitted to: Sleep. 159 Chapter 7 The Sustained Attention to Response Task Shows Lower Cingulo-Opercular and Frontoparietal Activity in People with Narcolepsy Type 1: An fMRI Study on the Neural Regulation of Attention Brain Sciences. 2020;10(7):419. 195

Chapter 8 Enhanced Visual Cortex Activation in People With Narcolepsy Type 1 During Active Sleep Resistance: An fMRI-EEG Study Frontiers in Neuroscience. 2022;16:904820. 225 Section C – The Future: Improving Diagnostic Classification and Identifying New Treatment Options Chapter 9 Data-Driven Phenotyping of Central Disorders of Hypersomnolence With Unsupervised Clustering Neurology. 2022;98(23):e2387-e2400. 241 Chapter 10 The Therapeutic Potential of Opioids in Narcolepsy Type 1: A Systematic Literature Review and Questionnaire Study Sleep Medicine. 2023;109:118-27. 299 Chapter 11 Discussion 327 Chapter 12 References 359 Nederlandse Samenvatting 390 Curriculum Vitae 396 Publication List 398 Training Portfolio 400 Dankwoord 401

General Introduction Based on: Gool JK, Cross N, Fronczek R, Lammers GJ, Van der Werf YD, Dang-Vu TT. Neuroimaging in Narcolepsy and Idiopathic Hypersomnia: from Neural Correlates to Clinical Practice. Curr Sleep Medicine Rep. 2020;6:251–66. Gool JK, Schinkelshoek MS, Fronczek R. What triggered narcolepsy: H1N1 vaccination, virus, or both? Important lessons learned from China. Sleep. 2023;46(3):zsad005. Gool JK, Dang-Vu TT, van der Werf YD. White matter integrity in narcolepsy: the structural blueprint for functional complaints? Sleep. 2024; zsae020. Schinkelshoek MS, Gool JK, Fronczek R, Lammers GJ. Oorzaken en behandeling van hypersomnolentie. Nervus. 2019;4(1). 1

10 Chapter 1 The narcolepsy phenotype was initially described in 1877 by Dr. Carl Friedrich Otto Westphal and first named narcolepsy in 1880 by Dr. Jean Baptiste Gélineau as a syndrome with frequent daytime sleep attacks. Both doctors detailed periods of muscle atonia with preserved consciousness triggered by strong emotions [1-3], a phenomenon termed cataplexy by Dr. Richard Henneberg in 1916 [4, 5]. Excessive daytime sleepiness typically manifests as a diurnal inability to maintain wakefulness, and generally co-occurs with daytime vigilance complaints [6]. The combination of excessive daytime sleepiness and cataplexy has since been pathognomonic for the narcolepsy diagnosis (type 1). Narcolepsy without cataplexy (type 2) and idiopathic hypersomnia are later defined diagnostic entities within the central disorders of hypersomnolence domain and present with excessive daytime sleepiness in absence of cataplexy [7]. This thesis focusses on unravelling the pathophysiologies of the narcolepsies and idiopathic hypersomnia (the past), the effects of narcolepsy on brain structure and functioning (the present), and improving classification and disease management of central disorders of hypersomnolence (the future). Diagnostic differentiation and clinical presentation The American Academy of Sleep Medicine (AASM) currently identifies narcolepsy type 1, narcolepsy type 2 and idiopathic hypersomnia as the three chronic central disorders of Hypersomnolence [8, 9]. Detailed diagnostic criteria of the third edition of the International Classification of Sleep Disorders (ICSD-3) are described in Box 1 [8]. Diagnostic differentiation is based on the combination of medical history taking and ancillary findings of a nocturnal polysomnography and daytime multiple sleep latency test (MSLT), commonly known as a napping test. Polysomnography is primarily used to rule out potential sleep disturbances and evaluate sleep duration, whereas the MSLT is employed to assess the severity of daytime sleepiness and the presence of sleep-onset rapid eye movement periods (SOREMPS), which are indicative of the narcolepsy diagnoses. Narcolepsy type 1 diagnosis can also be confirmed through hypocretin-1 deficiency via lumbar puncture (further details discussed in the pathophysiology section). The diagnostic criteria for narcolepsy type 1 have recently been revised in the ICSD-3 Text Revision, where a nocturnal SOREMP during the polysomnography in combination with typical cataplexy suffices for the diagnosis narcolepsy type 1 [9]. Within this thesis the original ICSD-3 criteria were used [8].

11 Introduction Narcolepsy type 1 Besides excessive daytime sleepiness and cataplexy, the often referenced narcolepsy type 1 symptom pentad is completed by disturbed nocturnal sleep with frequent awakenings, hypnopompic or hypnagogic hallucinations and sleep paralysis. Many of these symptoms are hypothesized to originate from inappropriate intrusions of REM sleep phenomena while being or feeling awake, including muscle atonia during cataplexy and sleep paralysis, and the dream-like sensory fabulations of hypnagogic hallucinations [4]. Substantial weight gain close to onset of narcolepsy symptoms is regularly seen, typically starting within six months [10]. This adds to an increased prevalence in obstructive sleep apnoea compared to healthy sleepers [11]. Other nocturnal sleep disturbances are also regularly reported, including parasomnias such as REM sleep behaviour disorder, nocturnal eating disorder, and periodic limb movements [11, 12]. People with narcolepsy type 1 generally describe daytime difficulties. Excessive daytime sleepiness often leads to sleep attacks and an inability to maintain vigilance. Vigilance is vital for effective daytime functioning and people with narcolepsy consequently frequently encounter obstacles in studying or working efficiently [6, 13], and are at a higher risk of being involved in traffic accidents [14, 15]. Excessive daytime sleepiness is frequently accompanied by fatigue complaints in people with a central disorder of hypersomnolence, and results in lack of energy [16]. Narcolepsy type 1 is also associated with increased prevalence of depression and anxiety disorders. There is ongoing debate whether there is an association with attention deficit hyperactivity disorder (ADHD) or if attention problems in almost all cases can be explained by the impaired vigilance characteristic of narcolepsy. There are conflicting reports about an increased occurrence of psychotic symptoms in narcolepsy type 1 [17-21]. 1

12 Chapter 1 Box 1 – ICSD-3 diagnostic criteria Narcolepsy type 1 (criteria A and B): A) Daily periods of irrepressible need to sleep or daytime lapses into sleep, present for at least three months. B) E ither one and/or two: (1) Cataplexy and mean sleep latency ≤ 8 minutes and two or more SOREMPs on the MSLT. REM sleep within 15 minutes of sleep onset on the preceding nocturnal polysomnogram may replace one of the SOREMPs. (2) Low CSF hypocretin-1 concentration (< 110 pg/mL or less than one-third of control values). Narcolepsy type 2 (all criteria A-E): A) Daily periods of irrepressible need to sleep or daytime lapses into sleep, present for at least three months. B) M ean sleep latency ≤ 8 minutes and two or more SOREMPs on the MSLT. REM sleep within 15 minutes of sleep onset on the preceding nocturnal polysomnogram may replace one of the SOREMPs. C) No cataplexy. D) CSF hypocretin-1 concentration has not been measured or CSF hypocretin-1 concentration is ≥ 110 pg/mL or greater than one-third of control values. E) T he hypersomnolence and/or MSLT findings are not better explained by other causes. Idiopathic hypersomnia (all criteria A-F): A) Daily periods of irrepressible need to sleep or daytime lapses into sleep, present for at least three months. B) F ewer than two SOREMPs on the MSLT (or fewer than one if nocturnal REM sleep latency was ≤ 15 minutes). C) No cataplexy. D) Either one and/or two: (1) Mean sleep latency ≤ 8 minutes on the MSLT. (2) Total 24-hour sleep time ≥ 660 minutes on 24-hour polysomnographic monitoring or wrist actigraphy (averaged over ≥ 7 days). E) Insufficient sleep syndrome is ruled out. F) T he hypersomnolence and/or MSLT findings are not better explained by other causes.

13 Introduction Narcolepsy type 2 and idiopathic hypersomnia Narcolepsy type 2 and idiopathic hypersomnia are characterized by excessive daytime sleepiness in absence of cataplexy. Many of the symptoms of narcolepsy type 1 may also occur, including hypnagogic hallucinations, sleep paralysis and daytime vigilance and fatigue complaints [22-24]. The diagnosis of narcolepsy type 2 requires multiple SOREMPS during the MSLT and polysomnography (similar to narcolepsy type 1), whereas only up to one SOREMP is allowed for idiopathic hypersomnia [8]. Excessive daily need for sleep (defined as more than 11 hours) is diagnostically used to identify a subgroup of individuals with idiopathic hypersomnia, a phenotype frequently accompanied by sleep drunkenness [23, 24]. Consensus on the definition of sleep drunkenness is lacking but it is generally described as substantial difficulty with awakening that is accompanied by confusional behaviour, typically lasting for more than 30 minutes [25, 26]. Sleep drunkenness in idiopathic hypersomnia is in contrast with narcolepsy type 1 with typically refreshing nature of sleep. Challenges in current diagnostic criteria Despite these internationally accepted criteria, the diagnostic workup often proves difficult to reliably identify different central disorders of hypersomnolence [27-29]. This is partially because of the rarity of these disorders with estimated prevalences between 25-50 per 100,000 individuals for narcolepsy type 1 [30-32], around 25 per 100,000 individuals for narcolepsy type 2 and roughly 10 per 100,000 for idiopathic hypersomnia [33]. Especially narcolepsy type 2, idiopathic hypersomnia, and chronic sleep deprivation may phenotypically overlap at clinical presentation [28]. Individuals with the same central disorder of hypersomnolence diagnosis may also experience different phenotypes. The challenging diagnostic workup is further emphasized by the poor test-retest reliability of the multiple sleep latency test (MSLT) in the absence of cataplexy, where diagnostic crossover of up to 53% was seen for narcolepsy type 2 and 75% for idiopathic hypersomnia [34, 35]. Multiple opinion papers have recently stressed the importance of identifying new objective biomarkers better reflecting the underlying pathophysiological mechanisms of these disorders to resolve the blurred boundaries between hypersomnolence subtypes [27-29]. With this same goal, the European Narcolepsy Network (EU-NN) launched a collaborative prospective clinical database including all central disorders of hypersomnolence (Box 2) [22]. This unique dataset will be used for the analyses presented in Chapters 2 and 9. 1

14 Chapter 1 Pathophysiology The discovery of hypocretin (or orexin) in 1998 led to a revolution in sleepwake research and narcolepsy in particular [36, 37]. Hypocretin and orexin were concurrently identified by two independent research groups and later found to be the same peptide. Two types of hypocretin were identified both stemming from the precursor prepro-hypocretin which is exclusively produced in the lateral and posterior hypothalamus [38]. The hypocretins proved vital for adequate sleep-wake control [39], feeding behaviour [37, 40] and reward processing [41, 42]. One year after the discovery of hypocretin, an inheritable mutation encoding the hypocretin-2 receptor was found to cause canine narcolepsy [43]. Two independent research groups identified a slightly different role for hypocretin in human narcolepsy, as they found selective loss of hypocretin-1 neurons in the lateral and posterior hypothalamus [44, 45]. Rodent models have consistently shown that loss of the wake-promoting and stabilizing role of hypocretin-1 leads to a typical narcolepsy type 1 phenotype [39]. Hypocretin-1 can be reliably measured from the cerebrospinal fluid (CSF) and currently serves as the only reliable objective diagnostic biomarker in diagnosing any of the central disorders of hypersomnolence [46]. The role of hypocretin has extensively been investigated in central disorders of hypersomnolence, yet many studies did not specify whether hypocretin-1 and/ or hypocretin-2 were studied. Since most studies likely focused on hypocretin-1, we will henceforth use “hypocretin” to generally denote hypocretin-1, unless stated otherwise. Narcolepsy type 2 and idiopathic hypersomnia have to date been understudied compared to narcolepsy type 1, and due to the multidimensional nature of these disorders, their pathophysiology remains poorly understood. In these two conditions, it has been suggested that endogenous gammaaminobutyric acid (GABA) might promote daytime sleepiness by enhancing GABA-A receptor signalling but this hypothesis was mainly based on a pilot study of a small and heterogeneous sample [47]. Longitudinal studies have uncovered that a proportion of (typically young) individuals with narcolepsy type 2 develop cataplexy within years after onset of excessive daytime sleepiness [48, 49]. Whether these individuals already were hypocretin deficient when the sleepiness symptoms arose and were hereby misclassified, or whether hypocretin deficiency develops closer to cataplexy onset, remains unknown. Reports of delayed development of cataplexy years after confirmed hypocretin deficiency at least suggests frequent misclassification in narcolepsy type 2 [50]. Lack of understanding of underlying pathophysiologies of narcolepsy type 2

15 Introduction and idiopathic hypersomnia has so far hindered scientific and therapeutic breakthroughs. Environmental triggers The aetiopathogenesis of hypocretin deficiency in narcolepsy type 1 is currently thought to have an autoimmune basis with roots in both genetic and environmental factors. Genetic predisposition of the human leukocyte antigen (HLA) complex HLA-DQB1*06:02 increases susceptibility to develop narcolepsy type 1 by over 200-fold, and up to 98% of patients carry this haplotype [51-53]. HLA-DQB1*06:02 has been considered a genetic factor that is essential but not sufficient to develop narcolepsy type 1 as 20-30% of the general European population also carries this haplotype. Associations with other HLA classes have been reported (such as HLA-DR and HLA-A) albeit not as strong as HLA-DQ, and HLA-DQB1*06:02 in particular [51-53]. The onset of narcolepsy type 1 most commonly manifests during adolescence and young adulthood [54, 55] and has been associated with immunological life events. In the winter of 2009-2010 the influenza A virus subtype H1N1 resurfaced through a pandemic, after which clearly increased incidence rates of narcolepsy type 1 were quickly reported in Scandinavian children [56, 57]. The H1N1 vaccine named Pandemrix® (GlaxoSmithKline Biologicals, Wavre, Belgium) that was frequently administered in these countries, was initially deemed to be the culprit. Not long after, research groups from countries with low vaccination grades (such as China, the United States, Taiwan, and several other European countries) also reported a more modest increase in narcolepsy type 1 incidence [57-61]. A possible role for the H1N1 virus itself was thus emphasized. The discovery that immune system-related triggers such as Pandemrix vaccination and H1N1 infection are associated with narcolepsy type 1 development, in combination with the H1N1 virus not circulating for decades before its reintroduction in 2009-2010 [62], has sparked interest into the possibility of other triggers also causing narcolepsy type 1. Mainly streptococcal infections have been associated with onset of narcolepsy type 1 but direct causal evidence of underlying pathophysiological mechanisms [10, 63-67], and assessment of other potential triggers remains limited. Non-flu potential triggers remain poorly studied, also in the context of narcolepsy type 2 and idiopathic hypersomnia. In Chapter 2 we present narcolepsy incidence rates before and after the 2009-2010 H1N1 pandemic using the EU-NN database. Our analyses in Chapter 3 additionally provide correlations with local severity data of the preceding flu season to identify whether particular influenza strains are attributable to 1

16 Chapter 1 fluctuations in narcolepsy incidence rates. In Chapter 4 we cross-sectionally investigate the self-reported immunological triggers before onset of narcolepsy type 1 and type 2, and idiopathic hypersomnia. The cellular processes induced by a potential immunological trigger that ultimately lead to hypocretin deficiency in narcolepsy type 1 have been extensively studied. Specific CD4+ T-cells targeting hypocretin have been identified in the blood of individuals with narcolepsy type 1 [68, 69] but these findings have not been consistently replicated [70, 71]. In contrast to CD8+ T-cells, such CD4+ T-cells cannot directly have harmed hypocretin neurons, as they only recognize HLA type II molecules on professional antigen-presenting immune cells instead of the HLA type I molecules expressed by neurons [72]. Additional involvement of HLA type II-mediated effector immune cells seems inevitable but this process remains incompletely understood. Postmortem brain tissue microscopy of a donor with secondary narcolepsy type 1 by Ma2 antibody encephalitis showed hypothalamic CD8+ T-cell infiltration, suggesting a role in disease development [73]. Hypocretin-reactive cytokine-producing CD8+ T-cells have been found in blood sera of children with narcolepsy type 1 (but not in adults) [74]. Increased CD8+ T-cell autoreactivity to narcolepsy type 1-related proteins has additionally been reported in narcolepsy type 1 adults compared to HLA DQB1*06:02 positive healthy controls [75]. CD8+ T-cells have therefore been associated with hypocretin deficiency but a causal relationship has yet to be proven. Taken together, the underlying immunological mechanisms of narcolepsy type 1 remain only partially understood. Neuroimaging The discovery of hypocretin deficiency occurred around the same time as the rising use of magnetic resonance imaging (MRI) [44, 45]. MRI has since been widely implemented to investigate narcolepsy type 1 [76]. An improved understanding of brain structure and functioning in central disorders of hypersomnolence could assist in unravelling disease pathophysiology and provide important new targets for disease management strategies. For a full overview of structural and functional MRI, PET and SPECT studies investigating central disorders of hypersomnolence, please see Appendix A. Structural neuroimaging in narcolepsy type 1 Using volumetric analyses to investigate structural differences led several studies to observe volume reductions in the hypothalamus (Figure 1) [77-81], with one of these studies also reporting correlations between these changes

17 Introduction and disease severity [80]. Several other studies, however, have failed to find any structural differences in the hypothalamus [82-85] and a recent study reported hypothalamus volume increases in individuals that developed narcolepsy type 1 after Pandemrix vaccination [86]. Other structural findings in narcolepsy type 1 include a consistently reduced volume in frontotemporal regions of the cortex observed across cortical morphometry and thickness [7882, 84, 85, 87-90]. These may be related to typical narcolepsy type 1 complaints, including daytime impairments such as deficits in attention [6, 91-94] and subjective memory complaints [94-96]. These cognitive difficulties could be further explained by a lower volume of the (anterior) hippocampus that has been described by some studies [88, 97-100]. Inconsistent differences within the limbic system in narcolepsy type 1 have also been observed, including reduced volumes of medial prefrontal cortex and anterior cingulate in juveniles and young adults [80, 88] in addition to the amygdala in older adults [97, 101]. These changes have been associated with the frequency and severity of cataplexy as disturbed emotion regulation [97] but could also be related to the altered functional control of the basal amygdala on the brainstem nuclei that normally regulate muscle tone as has been described in rodent studies [102, 103]. Figure 1. Structural T1-weighted magnetic resonance imaging studies investigating central disorders of hypersomnolence. Regions were generally included in case there were at least three independent studies indicating changes within this region in people with a central disorder of hypersomnolence compared to controls. Reported differences could reflect any T1-weighted structural outcome measure, such as volume, cortical thickness, gyrification and/or surface area. The numbers between brackets correspond to the reference numbers of the studies. HC = Healthy controls; NT1 = Narcolepsy type 1; IH = Idiopathic hypersomnia. Figure was created with Biorender. As structural volumetric analyses have resulted in mixed results in narcolepsy type 1, more recent studies moved towards diffusion-weighted 1

18 Chapter 1 imaging (DWI) analyses to investigate white matter morphology through water diffusivity (Figure 2) [104]. Tract-based spatial statistics studies have reported consistent brain-wide differences irrespective of disease duration, mainly in the hypothalamus-thalamus-orbitofrontal pathway and brainstem as parts of the sleep-wake regulation system, the reward and limbic system, and the corticospinal tract [84, 105-107]. The fronto-occipital differences were negatively correlated with subjective sleepiness scores and positively correlated with REM sleep latency [108]. Two studies that used voxel-based statistics showed inconsistent DWI differences in hypothalamic, brainstem, and cortical regions [109, 110]. In Chapter 5 we present multimodal investigation of white matter morphology in narcolepsy type 1 using DWI-based tract-based spatial statistics (TBSS), quantitative regions-of-interest (ROI) analyses and hypothalamus seeded-tractography. In analyses described in Chapter 6 we perform postmortem human histopathological analyses in narcolepsy type 1 to further assess axonal density and injury, and myelin integrity. Figure 2. Diffusion-weighted imaging (DWI) studies investigating structural connectivity in central disorders of hypersomnolence. Regions were generally included in case there were at least three independent studies indicating white matter changes within this region in people with narcolepsy compared to controls. Reported differences could reflect any DWI outcome measure that represents white matter integrity, such as fractional anisotropy (FA) or mean diffusivity/apparent diffusion coefficient (MD/ADC). The numbers between brackets correspond to the reference numbers of the studies. The study by Chen et al. [212] did not divide narcolepsy into type 1 and type 2 when comparing structural connectivity to controls. Post hoc comparisons between narcolepsy type 1 and type 2 did not reveal differences between narcolepsy subtypes, and the study by Chen et al. [212] was thus reported for both the “NT1 < HC” and the “NT2 < HC” contrasts. HC = Healthy controls; NT1 = Narcolepsy type 1; NT2 = Narcolepsy type 2; WM = White matter. Figure was created with Biorender and adapted from Gool et al. [104].

19 Introduction Functional neuroimaging in narcolepsy type 1 Functional MRI (fMRI) is most commonly conducted during so-called resting state (to extract information about intrinsic brain networks) or during specifically designed tasks (to extract information related to specific cognitive processes). Both types of fMRI paradigms have been applied to mainly narcolepsy type 1. Using resting state fMRI, multiple studies have reported decreased resting state activity within the default mode network as the main resting network [111114], which also seemed less connected to networks that are normally active during cognitive tasks (salience, dorsal attention and executive network) [115]. The combination of deactivation of the default mode network and upregulation of the cognitive networks is essential to perform well in cognitive tasks. The decreased coupling of these networks in narcolepsy suggests a dysregulation of mental resources in favour of staying awake over actual task performance [116]. Limbic network activity has been thoroughly assessed using fMRI taskbased studies on emotional processing and cataplexy attacks [117-119]. These studies have reported both higher and lower hypothalamus activity in combination with mainly enhanced amygdala and reward system activity during emotional processing and cataplexy attacks. Loss of hypothalamic control over the mesolimbic reward system in NT1 during emotional stimuli seems responsible for triggering cataplexy attacks. Limited fMRI studies have been performed to investigate the neural activation patterns underlying other common symptoms of narcolepsy type 1, and none had yet focussed on vigilance impairments and the occurrence of unwanted sleep attacks. In Chapters 7 and 8 we present results from task-based fMRI studies investigating brain activation patterns during the sustained attention to response task (SART) and an active sleep resistance paradigm in individuals with narcolepsy type 1 compared to healthy controls. 1

20 Chapter 1 Figure 3. Resting state functional magnetic resonance imaging (fMRI) studies investigating functional connectivity in central disorders of hypersomnolence. Regions were generally included in case there were at least two independent studies indicating changes within this region in people with a central disorder of hypersomnolence compared to controls. Reported differences could reflect any fMRI outcome measure, such as within and between network connectivity, seed-based connectivity and fractional low-frequency amplitude of low-frequency fluctuations (fALFF). The numbers between brackets correspond to the reference numbers of the studies. HC = Healthy controls; NT1 = Narcolepsy type 1; IH = Idiopathic hypersomnia. Figure was created with Biorender. Neuroimaging in narcolepsy type 2 and idiopathic hypersomnia Most neuroimaging studies to date have been conducted in small samples (typically 10-30 individuals per group), while narcolepsy type 2 and idiopathic hypersomnia remain relatively understudied. From the limited number of structural neuroimaging studies performed in narcolepsy type 2 and idiopathic hypersomnia, it appears that people with narcolepsy type 2 [99, 110] and idiopathic hypersomnia [120] have a smaller hippocampus and larger precuneus, respectively. Structural brain differences in narcolepsy type 2 and idiopathic hypersomnia exhibit some similar but weaker differences compared to narcolepsy type 1. Even though direct cross-disorder comparisons are lacking, functional resting-state network connectivity differences appear less striking in idiopathic hypersomnia as compared to narcolepsy type 1 [111-115, 120] and no resting-state fMRI studies have been performed in narcolepsy type 2. Given that this field is still relatively new and no single MRI study has thus far included people with narcolepsy type 1, type 2, and idiopathic hypersomnia, it remains difficult to draw robust conclusions on differences between diagnoses.

21 Introduction Treatment options No curative treatments are currently available for any of the central disorders of hypersomnolence. Across the diagnoses, the non-pharmacological options include regular bedtimes, scheduled daytime napping, patient education and psychotherapy. Various wake-promoting agents are effective against EDS and include methylphenidate, dexamphetamine, modafinil/armodafinil, pitolisant and solriamfetol. These compounds act on dopamine, histamine, noradrenaline and other systems [4, 121-123]. Commonly used compounds for treating cataplexy are sodium oxybate or antidepressants like clomipramine or venlafaxine [121, 123]. As a GABAB receptor agonist, sodium oxybate leads to more consolidated nocturnal sleep and is effective against all major narcolepsy symptoms, even though its exact mode of action remains largely unknown [124]. Pitolisant is also effective against cataplectic attacks, but its effect is usually weaker than sodium oxybate. Based on individuals’ symptoms, different combinations of sodium oxybate, wake-promoting agents and antidepressants may be used. The currently available treatments often have adverse side effects and incompletely alleviate excessive daytime sleepiness, which are essential considerations in finding the optimal treatment regimen [123]. A recent study in rodents and postmortem human brain tissue has suggested that opioids interact with the hypocretin system and could hereby possibly relieve narcolepsy type 1 symptoms [125]. Mainly case reports have been reported and no structured assessment has been conducted on this potential new treatment option in humans with narcolepsy type 1 [126-131]. In Chapter 10 we assess the possible effects of opioid use on narcolepsy type 1 symptomatology. Outline of the thesis This thesis assesses the past (Section A), present (Section B) and future (Section C) of central disorders of hypersomnolence. Section A on the origin of central disorders targets the role of immunological triggers in developing central disorders of hypersomnolence. We studied annually fluctuating incidence rates of narcolepsy type 1 (Chapter 2) in the EU-NN database, annual incidence rates of all central disorders of hypersomnolence in relation to preceding flu season severity in multiple complete European centres (Chapter 3) and realworld insights on immunological events that people with a central disorder of hypersomnolence reported prior to disease onset (Chapter 4). The chapters on the present in Section B first focus on brain morphology of people with 1

22 Chapter 1 narcolepsy type 1 with DWI analyses on white matter morphology (Chapter 5) and human postmortem immunohistochemistry analyses on axonal density, orientation and myelin integrity (Chapter 6). This is followed by fMRI studies to investigate the behavioural and neural correlates of sustained attention (Chapter 7) and active sleep resistance (Chapter 8) in narcolepsy type 1. The final two chapters on the future of central disorders of hypersomnolence in Section C compose unsupervised machine learning analyses to identify more reliable subgrouping of people with a central disorder of hypersomnolence to potentially improve diagnostic classification (Chapter 9) and a combined literature review, questionnaire and semi-structured interview study to investigate the potential of opioids as a treatment option for narcolepsy type 1 (Chapter 10). The culmination of the thesis will be presented in Chapter 11, a comprehensive discussion wherein the findings will be summarized and contextualized in scientific discourse with proposed avenues for future research.

23 Introduction Appendix A MRI and nuclear imaging studies in central disorders of hypersomnolence Table 1. Structural MRI findings in narcolepsy and idiopathic hypersomnia Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Kaufmann et al. 2002 VBM (1.5T) NT1 12 (6) 32 (16) 36.9 (15.8) 36.2 (14.7) MSLT, cataplexy, HLA 6:6 No change in hippocampal GM. Reduced volume in the frontotemporal cortex. Draganski et al. 2002 VBM (1.5T) NT1 29 (17) 29 (17) 39.7 (11.3) 38.6 (9.3) NR NR Reduced GM in the hypothalamus, nucleus accumbens and frontotemporal cortex. Overeem et al. 2003 VBM (1.5T) NT1 15 (8) 15 (8) 44.7 (14.3) 44.5 (14.2) PSG, MSLT, cataplexy, CSF 13:2 No change in hippocampal GM. Brenneis et al. 2005 VBM (1.5T) NT1 12 (4) 12 (2) 35.8 (13.2) 35.0 (8.4) PSG, MSLT, cataplexy, HLA 10:2 No change in hippocampal GM. Reduced volume in the frontotemporal cortex. Buskova et al. 2006 VBM (1.5T) NT1 19 (9) 16 (7) 43.4 (13.8) 40.3 (10.9) PSG, MSLT, cataplexy, HLA 9:10 Reduced GM in the hypothalamus. Joo et al. 2009 VBM (1.5T) NT1 29(14) 29(14) 31.2 (NR) 31.2 (NR) PSG, MSLT, cataplexy, HLA 0:29 Reduced GM in the hypothalamus, nucleus accumbens, thalamus and frontotemporal cortex. Kim et al. 2009 VBM (1.5T) NT1 17 (4) 17 (4) 24.6 (4.9) 26.6 (5.2) MSLT, cataplexy, HLA 11:6 Reduced GM in the hypothalamus, thalamus and frontotemporal cortex. 1

24 Chapter 1 Table 1. Continued. Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Tondelli et al. 2018 VBM + Automated segmentation (3T) NT1 20 (13) 19 (NR) 12.2 (3.1) NR PSG, MSLT, cataplexy, CSF, HLA 0:20 Reduced GM in cerebellum and medial prefrontal cortex and increased volume in right hippocampus. Cortical thickness in frontal lobe was also reduced. Joo et al. 2011 Automated segmentation (1.5T) NT1 28 (18) 33 (18) 26.9 (7.9) 30.1 (11.1) PSG, MSLT, cataplexy, HLA 0:28 Reduced cortical thickness in the cingulate, frontotemporal, and inferior parietal cortices. Schaer et al. 2012 Automated segmentation (3T) NT1 12 (7) 12 (7) 28.8 (6.8) 31.5 (6.2) cataplexy, CSF, HLA 7**:5 Increased cortical thickness in the lateral prefrontal cortex, and decreased thickness in the paracentral lobule. Kim et al. 2016 Automated segmentation (3T) NT1 33 (12) 31 (12) 27 (5.9) 27 (5.7) PSG, MSLT, cataplexy, HLA NR Reduced bilateral hippocampal and amygdalar centromedial volume. These were associated with longer duration of daytime sleepiness and shorter REM sleep latency. Nemcova et al. 2015 Automated segmentation + manual tracing (1.5T) NT1 + NT2 NT1: 53 (25) NT2: 23 (9) 37 (21) NT1: 39.6 (16.9) NT2: 40.4 (14.8) 36.3 (8.8) PSG, MSLT, cataplexy, HLA NT1 - 53:0** NT2 - 11:12 Reduced hippocampus volume in NT1 (10%, manual tracing only). Reduced nonsignificant hippocampus volume in NT2 compared to controls. Brabec et al. 2011 Manual tracing (1.5T) NT1 11 (6) 11 (6) 41.7 (17.7) NR PSG, MSLT, cataplexy, HLA 9:2 Reduced amygdala volume in NT1. Joo et al. 2012 Manual tracing (1.5T) NT1 36 (11) 36 (11) 29 29 PSG, MSLT, cataplexy, HLA 0:36 Reduced bilateral hippocampus volume in NT1; negative correlation between hippocampus volume and mean sleep latency and REM sleep latency.

25 Introduction Table 1. Continued. Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Kreckova et al. 2019 Manual tracing (1.5T) NT1 48 (22) 37 (21) 40 (17) 36.3 (8.8) MSLT, cataplexy, HLA NR Reduced anterior hippocampus volume (10.5%); negative correlation between anterior hippocampus volume and disease duration. Pomares et al. 2019 Automated segmentation (3T) IH 12 (9) 15 (9) 33.4 (10.1) 31.2 (9.8) PSG, MSLT 12**:0 Increased volume and cortical thickness in the precuneus in IH. Greater positive correlation between left medial prefrontal cortex and precuneus thickness in IH. Jeon et al. 2020 Automated segmentation (3T) NT1 17 (4) 83 (20) 30.5 (6.4) 32.8 (8.6) PSG, MSLT, cataplexy, HLA 17:0 Longitudinal progressive cortical thinning across bilateral dorsolateral frontal and fusiform cortices, right anterior cingulate. Faster progressive cortical thinning and worse disease severity over time in the people with NT1 with early-onset. Xiao et al. 2021 Automated segmentation (3T) NT1 51 (13) 60 (18) 20.1 (7.7) 19.4 (7.0) PSG, MSLT, cataplexy 0:51 Reduced cortical thickness in the bilateral frontal cortex and left precuneus in paediatric NT1. Mixed increased and reduced gyrification in NT1 in the frontal, occipital, parietal and temporal lobes, which correlated with sleepiness severity and hypnagogic hallucinations. Kim et al. 2022 Automated segmentation (3T) NT1 + NT2 NT1: 9 NT2: 6 (8) † 19 (10) 33.2 (15.5) † 36.7 (8.3) PSG, MSLT, cataplexy 0:15 No change in overall hypothalamus or subunits. Reduced path lengths of hypothalamic subunits in narcolepsy. Juvodden et al. 2023 Automated segmentation (3T) NT1 54 (39) 114 (77) 21.8 (11.0) 23.2 (9.0) PSG, MSLT, cataplexy, CSF, HLA 49**:5 Increased volume in the bilateral overall hypothalamus and inferior tubular-inferior subregions in NT1. 1

26 Chapter 1 Table 1. Continued. Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Juvodden et al. 2024 Automated segmentation (3T) NT1 54 (39) 114 (77) 21.8 (11.0) 23.2 (9.0) PSG, MSLT, cataplexy, CSF, HLA 49**:5 Reduced cortical thickness in temporal regions in NT1. Scherfler et al. 2012 VBM + DWI (1.5T) NT1 16 (4) 12 (5) 56.8 (10.1) 59.8 (4.4) PSG, MSLT, cataplexy, HLA 10:6 No change in hippocampal GM. Reduced volume in the frontotemporal cortex. Impaired white matter tracts within the hypothalamus, frontotemporal and anterior cingulate cortices. Menzler et al. 2012 DWI (1.5T) NT1 8 (7) 12 (9) 49.5 (12.7) 56.8 (10.6) MSLT, cataplexy 8:0 Impaired WM within the hypothalamus, brainstem, caudate, frontotemporal and cingulate areas. Nakamura et al. 2013 DWI (1.5T) NT1 + NT2 NT1: 12 (3) NT2: 12 (6) 12 (6) NT1: 29.4 (4.9) NT2: 26.0 (5.2) 29.8 (2.2) PSG, MSLT, cataplexy 0:24 Impaired WM within the amygdala and frontoparietal cortex in NT1 and NT2. Park et al. 2016 DWI (3T) NT1 22 (12) 26 (11) 26.9 (7.9) 30.1 (11.1) PSG, MSLT, cataplexy, HLA 0:22 Impaired WM within the anterior cingulate, orbitofrontal cortex, frontal lobe, internal capsule, corpus callosum genu and thalamus. Tezer et al. 2018 DWI (3T) NT1 + NT2 NT1: 11 (7) NT2: 12 (5) 16 (9) NT1: 32.9 (11.2) NT2: 34.6 (12.1) 36.4 (11.3) PSG, MSLT, cataplexy NT1: 2:9 NT2: 2:10 Impaired WM in both NT1 and NT2 within the thalamus, corpus callosum body, midbrain, cerebellum, temporal and parahippocampal cortex and anterior internal capsule. Juvodden et al. 2018 DWI (3T) NT1 57 (39) 109 (67) 21.8 (11.0) 19.1 (8.3) PSG, MSLT, cataplexy, CSF, HLA 57**:0 Brain-wide impaired WM except for cerebellum.

27 Introduction Table 1. Continued. Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Gool et al. 2019 DWI (3T) NT1 12 (8) 11 (7) 33.3 (10.5) 31.8 (13.4) PSG, MSLT, cataplexy, CSF, HLA 5**:7 Brain-wide impaired WM except for cerebellum, localized differences in ventral diencephalon and midbrain and in hypothalamic tracts in connection with the midbrain. Park et al. 2020 DWI (3T) NT1 40 (15) 42 (0) 26.9 (7.2) 36.8 (10.7) PSG, MSLT, cataplexy, HLA 0:40 Impaired WM in bilateral inferior frontooccipital fasciculus. Ni et al. 2022 DWI (3T) NT1 36 (15) 33 (14) 22.8 (6.9) 23.1 (3.9) PSG, MSLT, cataplexy 0:36 Reduced local efficiency, global efficiency, and small-world WM network organisation in NT1. Negative correlation between cognitive performance and global network efficiency in NT1. Hovi et al. 2024 DWI (3T) NT1 19 (10) 19 (7) 15.8 (13.917.2)*** 13.0 (12.919.9)*** PSG, MSLT, cataplexy, CSF 14:5 Brain-wide impaired WM in NT1, which positively correlated with anxiety and depression symptoms and social and behavioural problems. Chen et al. 2024 DWI (3T) NT1 + NT2 NT1: 17 NT2: 13 (12) † 30 (12) 29.5 (13.5) † 29.0 (12.8) PSG, MSLT, cataplexy 0:30 Reduced WM network degree and global efficiency in narcolepsy. Increased connectivity of the cingulate gyrus in narcolepsy, which positively correlated with presence of REM sleep behaviour disorder. 1

28 Chapter 1 Table 1. Continued. Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Gumeler et al. 2023 DTI-ALPS (3T) NT1 + NT2 NT1: 14 (9) NT2: 11 (6) 11 (6) NT1: 33.5 (27.538.5)*** NT2: 40.0 (30.047.0)*** 36.0 (31.058.0)*** PSG, MSLT, cataplexy NR Negative correlation between DTI-ALPS index and WASO and percentage of wakefulness in NT1. Negative correlation between DTI-ALPS index and percentage of N1 sleep and positive correlation with REM percentage in NT2. Drissi et al. 2019 qMRI (3T) NT1 14 (10) 14 (NR) 16.4 (2.2) 16.7 (2.1) CSF 12:2 Reduced R2 values (neuromelanin) in NT1 in the reticular formation of the brainstem. Kim et al. 2008 MRS (3T) NT1 17 (3) 17 (5) 25.1 (4.6) 26.8 (4.8) MSLT, cataplexy, HLA 11:6 Greater concentration of GABA in NT1 in the medial prefrontal cortex. Witt et al. 2018 MRS (3T) NT1 17 (11) 20 (12) 16.5 (1.9) 17.4 (2.6) NR 15:2 No differences in GABA or Glutamate in the medial prefrontal cortex.

29 Introduction Table 1. Continued. Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Gao et al. 2024 EEG + fMRS (3T) NT1 26 (12) 29 (16) 23.0 (1.9) 23.8 (1.5) PSG, MSLT, cataplexy 0:26 Increased GABA/Glutamate-Glutamine ratio and GABA concentration in the prefrontal lobe in NT1. VBM highlights structures in the brain that have different volume by transforming the brain scan of each patient into the same standard template space. Manual tracing involves human experts to outline structures in the brain on MRI scans. Automated segmentation is similar to manual tracing except the tracing is performed by a computer algorithm. DWI is a specific scanning protocol designed to extract information about the orientation and integrity of white matter tracts in the brain. * Tests that were used to verify the hypersomnolence diagnosis of the included patients; ** Subjects withdrew from medication for a period prior to the start of the study. *** median and interquartile range. † NT1 and NT2 subjects were grouped together. DWI = Diffusion-weighted imaging; DTI-ALPS = Diffusion tensor image analysis along the perivascular space; EEG = Electroencephalography; fMRS = Functional magnetic resonance spectroscopy; GM = Grey matter; HLA = Human leukocyte antigen; IH = Idiopathic hypersomnia; MRS = Magnetic resonance spectroscopy; NR = Not reported; NT1 = Narcolepsy type 1; NT2 = Narcolepsy type 2; T = Tesla (strength of MRI magnet); VBM = Voxel Based Morphometry; WM = White matter. 1

30 Chapter 1 Table 2. Nuclear imaging and functional MRI findings in hypersomnolence disorders Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Sakai et al. 1979 133Xe Inhalation NT1 12 (2) 8 (2) 43 (17) 33 (13) PSG, MLST, cataplexy 0:12** Greater blood flow in NT1 at sleep onset, greatest in the brainstem-cerebellar regions. Joo et al. 2005 SPECT NT1 25 (8) 25 (8) 31 (NR) 31 (NR) PSG, MSLT, cataplexy 0:25 Hypoperfusion in NT1 in the anterior hypothalami, caudate nuclei, pulvinar nuclei of thalami, dorsolateral/ ventromedial prefrontal cortices, parahippocampal gyri, and cingulate gyri Boucetta et al. 2017 SPECT IH 13 (10) 16 (10) 33.1 (9.7) 31.0 (9.5) PSG, MSLT 0:13** Hypoperfusion in IH in medial prefrontal cortex, posterior cingulate cortex and putamen. Hyperperfusion in amygdala and temporo-occipital cortices. Joo et al. 2004 FDG-PET NT1 + NT2 NT1: 21 (NR) NT2: 3 (NR) 24 (8) 32 (NR) † 32 (NR) PSG, MSLT, cataplexy 0:24** Hypometabolism in narcolepsy in bilateral rectal and subcallosal gyri, medial convexity of right superior frontal gyrus, bilateral precuneus, right inferior parietal lobule, left supramarginal gyrus, bilateral posterior hypothalami and mediodorsal thalamic nuclei Dauvilliers et al. 2010 FDG-PET NT1 21 (11) 21 (11) 41.8 (17.5) 38.1 (19.5) PSG, MSLT, cataplexy, HLA 7:14 Hypermetabolism in NT1 in the limbic cortex specifically in the anterior and mid cingulate cortex, in the right cuneus and lingual gyrus.

31 Introduction Table 2. Continued. Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Dauvilliers et al. 2017 FDG-PET NT1 + IH NT1: 16 (4) IH: 9 (7) 19 (3) NT1: 30 (NR) IH: 27 (NR) 36 (NR) PSG, MSLT, cataplexy, CSF NT1: 0:16 IH: 0:9 Hypermetabolism in both patient groups in the anterior and middle cingulate, and insular cortex. No differences between NT1 and IH. Huang et al. 2016 FDG-PET NT1 71 (30) 20 (8) 16.2 (4.2) 15.1 (5.3) PSG, MSLT, cataplexy, HLA, CSF 0:71 Hypometabolism in NT1 in superior frontal, medial frontal, orbitofrontal, posterior cingulate and angular gyrus. Hypermetabolism in NT1 in the olfactory, hippocampus, parahippocampus, amygdala, fusiform, left inferior parietal lobe, left superior temporal lobe, basal ganglia, thalamus, cerebellum, right hypothalamus, midbrain and pons. Huang et al. 2018 FDG-PET NT1 + NT2 NT1:104 (37) NT2: 29 (10) 26 (10) NT1: 20.1 (9.1) NT2: 19.3 (5.6) 19.1 (5.3) PSG, MSLT, cataplexy, HLA 0:133 Hypermetabolism in NT1 compared to NT2 in the fusiform gyrus, striatum, hippocampus, thalamus, basal ganglia, and cerebellum. More hypometabolism in NT1 compared to NT2 in frontal lobe, posterior cingulate cortex, angular gyrus and parietal lobe and less hypometabolism in Heschl’s gyrus and paracentral lobule. 1

32 Chapter 1 Table 2. Continued. Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Trotti et al. 2021 FDG-PET NT1 + IH NT1: 14 (10) IH: 16 (14) 9 (6) NT1: 30.0 (8.3) IH: 38.1 (8.2) 32.2 (15.5) PSG, MSLT, cataplexy, CSF NR Hypermetabolism in NT1 compared to controls in the fusiform gyrus, middle occipital gyrus, superior and middle temporal gyri, insula, cuneus, precuneus, pre- and post-central gyri, and culmen. Hypermetabolism in IH compared to controls in the precuneus, inferior parietal lobule, superior and middle temporal gyri, and culmen. Chin et al. 2024 FDG-PET NT1 + NT2 NT1: 224 (117) NT2: 90 (56) 26 (10) NT1: 23.3 (9.5) NT2: 24.0 (9.5) 19.1 (NR) PSG, MSLT, cataplexy, HLA 0:314 Using metabolism rates from the left basal ganglia, Heschl and striatum, machine learning models achieved diagnostic classification accuracy of 99%. Barateau et al. 2024 TSPO-PET NT1 41 (20) 35 (16) 21.3 (9.9) 34.9 (18.5) PSG, MSLT, cataplexy, HLA, CSF 0:41 No change in [18F]DPA-714 binding (SUV/SUVr) in the hypothalamus and thalamus in NT1. Lower whole brain SUVr in NT1, which suggests lower microglia density. Drissi et al. 2016 EEG + fMRI (resting state) NT1 + NT2 NT1: 15 (10) NT2: 1 (1) 16 (10) NT1: (16.7) NT2: 17.9 NR (1320) PSG, MSLT, cataplexy, HLA, CSF 12:2 (2 unknown) Less time spent in EEG microstate in narcolepsy, which was associated with the default mode network. Narcolepsy, had altered resting state brain dynamics.

33 Introduction Table 2. Continued. Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Fulong et al. 2018 fMRI (resting state) NT1 51 (13) 60 (18) 25.8 (6.6) / 14.0 (2.7) 25.4 (4.3) / 13.3 (2.3) PSG, MSLT, cataplexy 0:51 Lower low-frequency fluctuations in bilateral medial superior frontal gyrus, bilateral inferior parietal lobule and supramarginal gyrus and greater low-frequency fluctuations in the bilateral sensorimotor cortex and middle temporal gyrus in both paediatric and adult NT1. Paediatric NT1 had greater lowfrequency fluctuations in the right putamen and thalamus. Xiao et al. 2019 fMRI (resting state) NT1 26 (8) 30 (12) 25.8 (6.6) 25.4 (4.3) PSG, MSLT, cataplexy 0:26 Lower functional connectivity in NT1 within the executive and salience networks, and increased functional connectivity in the bilateral frontal lobes. These were correlated with severity of daytime sleepiness. Pomares et al. 2019 fMRI (resting state) IH 12 (9) 15 (9) 33.4 (10.1) 31.2 (9.8) PSG, MSLT 0:12 Lower functional connectivity in IH within the medial prefrontal cortex. 1

34 Chapter 1 Table 2. Continued. Study Neuroimaging technique Disorder Sample size (number of females) Mean age, years (SD) Diagnosis * Medicated: unmedicated Main findings in patients (compared to controls) Patients Controls Patients Controls Järvelä et al. 2020 fMRI (resting state) NT1 21 (12) 21 (12) 28.1 (9.2) 28.3 (9.2) Cataplexy 19:2 Different functional network connectivity in NT1, mainly between the default mode network and other resting state networks, suggesting delayed and monotonic inter-network information flow. Xiao et al. 2020 fMRI (resting state) NT1 26 (5) 30 (6) 13.9 (2.7) 13.3 (2.3) PSG, MSLT, cataplexy 0:26 Lower overall clustering coefficient and small-worldness in paediatric NT1, and lower functional connectivity between the limbic system and the default mode network, and increased connectivity in the visual network. Findings were positively associated with subjective sleepiness, impulsiveness and depressive symptoms. Zhu et al. 2021 EEG + fMRI (resting state) NT1 25 (9) 25 (10) 22.4 (6.9) 22.5 (3.8) PSG, MSLT, cataplexy NR Whole brain lower global efficiency and small-world properties during stage N2 sleep in NT1. These measures negatively correlated with longer disease duration and worse cognitive testing.

RkJQdWJsaXNoZXIy MjY0ODMw