Doctoral thesis assessing the effects of one-week of escitalopram administration on healthy humans during sequence motor learning training. Published in 3 research articles.

There are robust sex differences in brain anatomy, function, as well as neuropsychiatric and neurodegenerative disease risk (1-6), with women approximately twice as likely to suffer from a depressive illness as well as Alzheimer’s Disease. Disruptions in ovarian hormones likely play a role in such disproportionate disease prevalence, given that ovarian hormones serve as key regulators of brain functional and structural plasticity and undergo major fluctuations across the female lifespan (7-9). From a clinical perspective, there is a wellreported increase in depression susceptibility and initial evidence for cognitive impairment or decline during hormonal transition states, such as the postpartum period and perimenopause (9-14). What remains unknown, however, is the underlying mechanism of how fluctuations in ovarian hormones interact with other biological factors to influence brain structure, function, and chemistry. While this line of research has translational relevance for over half the population, neuroscience is notably guilty of female participant exclusion in research studies, with the male brain implicitly treated as the default model and only a minority of basic and clinical neuroscience studies including a female sample (15-18). Female underrepresentation in neuroscience directly limits opportunities for basic scientific discovery; and without basic knowledge of the biological underpinnings of sex differences, we cannot address critical sexdriven differences in pathology. Thus, my doctoral thesis aims to deliberately investigate the influence of sex and ovarian hormones on brain states in health as well as in vulnerability to depression and cognitive impairment:Table of Contents List of Abbreviations ..................................................................................................................... i List of Figures .............................................................................................................................. ii Acknowledgements .....................................................................................................................iii 1 INTRODUCTION .....................................................................................................................1 1.1 Lifespan approach: Sex, hormones, and metabolic risk factors for cognitive health .......3 1.2 Reproductive years: Healthy models of ovarian hormones, serotonin, and the brain ......4 1.2.1 Ovarian hormones and brain structure across the menstrual cycle ........................4 1.2.2 Serotonergic modulation and brain function in oral contraceptive users .................6 1.3 Neuropsychiatric risk models: Reproductive subtypes of depression ...............................8 1.3.1 Hormonal transition states and brain chemistry measured by PET imaging ...........8 1.3.2 Serotonin transporter binding across the menstrual cycle in PMDD patients .......10 2 PUBLICATIONS ....................................................................................................................12 2.1 Publication 1: Association of estradiol and visceral fat with structural brain networks and memory performance in adults .................................................................................13 2.2 Publication 2: Longitudinal 7T MRI reveals volumetric changes in subregions of human medial temporal lobe to sex hormone fluctuations ..............................................28 2.3 Publication 3: One-week escitalopram intake alters the excitation-inhibition balance in the healthy female brain ...............................................................................................51 2.4 Publication 4: Using positron emission tomography to investigate hormone-mediated neurochemical changes across the female lifespan: implications for depression ..........65 2.5 Publication 5: Increase in serotonin transporter binding across the menstrual cycle in patients with premenstrual dysphoric disorder: a case-control longitudinal neuro- receptor ligand PET imaging study ..................................................................................82 3 SUMMARY ...........................................................................................................................100 References ..............................................................................................................................107 Supplementary Publications ...................................................................................................114 Author Contributions to Publication 1 .....................................................................................184 Author Contributions to Publication 2 .....................................................................................186 Author Contributions to Publication 3 .....................................................................................188 Author Contributions to Publication 4 .....................................................................................190 Author Contributions to Publication 5 .....................................................................................191 Declaration of Authenticity ......................................................................................................193 Curriculum Vitae ......................................................................................................................194 List of Publications ................................................................................................................195 List of Talks and Posters ......................................................................................................196

Introduction : La narcolepsie et l'hypersomnie idiopathique sont des troubles d'hypersomnie centrale peu compris, caractérisés par une somnolence diurne excessive causant des perturbations majeures au niveau du fonctionnement diurne et de la qualité de vie. Bien que certains aspects cliniques soient propres à chaque condition, plusieurs caractéristiques se chevauchent et ces dernières demeurent très difficile à diagnostiquer adéquatement. Le manque de distinction entre les troubles d'hypersomnolence centrale est extrêmement problématique, limitant la compréhension des mécanismes physiopathologiques sous- jacents. Objectif : À l'aide de la morphométrie cérébrale, l'objectif de l'étude est d'établir des différences anatomiques (c.-à-d., épaisseur corticale, volume sous-cortical) entre la narcolepsie avec cataplexie (NT1), la narcolepsie sans cataplexie (NT2), l'hypersomnie idiopathique (HI) et des participants en santé contrôles dans diverses régions du cerveau qui sont fonctionnellement liées au sommeil et au maintien de l'éveil. Méthodes : Une séquence d’acquisition d’images IRM anatomiques pondérées en T1 fut acquise sur 15 patients NT1, 15 NT2, 15 HI et 15 participants contrôles en santé (n = 60). Les images anatomiques furent traitées avec la suite logicielle FreeSurfer (FreeSurfer version 6.0.1) afin d'obtenir des mesures d'épaisseur corticale et de volume sous-cortical. Les mesures morphométriques obtenus pour différentes régions furent comparées entre les groupes par ANOVAs, ajustées pour l'âge. Résultats : Les résultats ont démontré une réduction volumétrique de la matière grise dans plusieurs structures sous-corticales associées au sommeil et au maintien de l'éveil dont l'hypothalamus et l'amygdale pour les patients NT1 et les patients NT2, comparés aux participants contrôles en santé. Les patients HI quant à eux n'ont pas démontré de différence volumétrique au niveau de l'hypothalamus comparativement aux participants contrôles, mais plutôt une diminution du volume de l'amygdale et du noyau accumbens, des structures associées à un réseau fonctionnel modulant la vigilance. Aucune différence significative d'épaisseur corticale n'a été retrouvée entre les groupes. Conclusion : Les résultats montrent des changements neuroanatomiques distincts entre les patients NT1 et HI, suggérant des mécanismes physiopathologiques différents et soulignent le phénotype hétérogène des patients NT2. / Introduction : Narcolepsy and idiopathic hypersomnia are poorly understood central disorders of hypersomnolence characterized by excessive daytime sleepiness leading to severe daytime disturbances and poor quality of life. Although some clinical features are specific to each condition, many characteristics overlap, and a reliable diagnosis remains difficult to achieve. The lack of clinical distinction between central disorders of hypersomnolence is extremely problematic and hinders the understanding of their underlying pathophysiological mechanisms. Objective : Using brain morphometry, the objective of this study is to establish anatomical differences (i.e., cortical thickness and subcortical volume) between narcolepsy with cataplexy (NT1), narcolepsy without cataplexy (NT2), idiopathic hypersomnia (HI) and healthy controls in brain regions involved in the modulation of sleep and wakefulness. Methods : T1-weighted MRI sequences were acquired in 15 NT1 patients, 15 NT2, 15 HI and 15 healthy controls (n = 60). Anatomical images were preprocessed using the FreeSurfer software package (FreeSurfer version 6.0.1) to obtain measures of cortical thickness and subcortical volume. Group differences in brain morphometric measurements acquired for different brain regions were analyzed using ANOVAs, adjusted for age. Results : Results displayed reduced gray matter volume in subcortical structures associated with the modulation of sleep and wakefulness, including the hypothalamus and the amygdala in NT1 and NT2 patients, compared to healthy controls. On the other hand, HI patients did not show volume changes in the hypothalamus compared to healthy controls, but instead showed a volume reduction of the amygdala and the nucleus accumbens, both structures associated with a functional network involved in the modulation of alertness. No significant group difference in cortical thickness was found. Conclusion : These results show distinct neuroanatomical changes between NT1 patients and HI patients, suggesting separate pathophysiological mechanisms and underline the heterogeneous phenotype of NT2 patients.

In the brain, neuronal cells located in different functional regions communicate through a dense structural network of axons known as the white matter (WM) tissue. Bundles of axons that share similar pathways characterize the WM anatomy, which can be investigated in-vivo thanks to the recent advances of magnetic resonance (MR) techniques. Diffusion MR imaging combined with tractography pipelines allows for a virtual reconstruction of the whole WM anatomy of in-vivo brains, namely the tractogram. It consists of millions of WM fibers as 3D polylines, each approximating thousands of axons. From the analysis of a tractogram, neuroanatomists can characterize well-known white matter structures and detect anatomically non-plausible fibers, which are artifacts of the tractography and often constitute a large portion of it. The accurate characterization of tractograms is pivotal for several clinical and neuroscientific applications. However, such characterization is a complex and time-consuming process that is difficult to be automatized as it requires properly encoding well-known anatomical priors. In this thesis, we propose to investigate the encoding of anatomical priors with a supervised deep learning framework. The ultimate goal is to reduce the presence of artifactual fibers to enable a more accurate automatic process of WM characterization. We devise the problem by distinguishing between volumetric and non-volumetric representations of white matter structures. In the first case, we learn the segmentation of the WM regions that represent relevant anatomical waypoints not yet classified by WM atlases. We investigate using Convolutional Neural Networks (CNNs) to exploit the volumetric representation of such priors. In the second case, the goal is to learn from the 3D polyline representation of fibers where the typical CNN models are not suitable. We introduce the novelty of using Geometric Deep Learning (GDL) models designed to process data having an irregular representation. The working assumption is that the geometrical properties of fibers are informative for the detection of tractogram artifacts. As a first contribution, we present StemSeg that extends the use of CNNs to detect the WM portion representing the waypoints of all the fibers for a specific bundle. This anatomical landmark, called stem, can be critical for extracting that bundle. We provide the results of an empirical analysis focused on the Inferior Fronto-Occipital Fasciculus (IFOF). The effective segmentation of the stem improves the final segmentation of the IFOF, outperforming with a significant gap the reference state of the art. As a second and major contribution, we present Verifyber, a supervised tractogram filtering approach based on GDL, distinguishing between anatomically plausible and non-plausible fibers. The proposed model is designed to learn anatomical features directly from the fiber represented as a 3D points sequence. The extended empirical analysis on healthy and clinical subjects reveals multiple benefits of Verifyber: high filtering accuracy, low inference time, flexibility to different plausibility definitions, and good generalization. Overall, this thesis constitutes a step toward characterizing white matter using deep learning. It provides effective ways of encoding anatomical priors and an original deep learning model designed for fiber.

No description available.

No description available.

Introduction : Les analyses de connectivité cérébrale dans les trois troubles majeurs d'hypersomnolence centrale, la narcolepsie de type I (NT1), de type II (NT2) et de l’hypersomnie idiopathique (HI) sont rares dans la littérature. Les études de connectivité structurelle indiquent comment les réseaux de fibres de matière blanche sont organisés et affectés dans le cas de patients atteints de troubles du système nerveux central. L’objectif principal de cette étude est d’observer si la connectivité structurelle est affectée dans la NT1, NT2 et l’HI comparativement à des sujets sains. Une diminution dans les mesures de connectivité globale dans les troubles d’hypersomnolence centrale et leur association à la sévérité des symptômes comme la somnolence sont supposées. Méthodes : Des sujets sains (HC, n=8) et des patients diagnostiqués avec soit l’HI (n=14), la NT1 (n=15) ou la NT2 (n=14) ont suivi une séance d'IRM avec l’imagerie pondérée en diffusion. Les mesures d’intégration (efficacité globale) et de ségrégation (efficacité locale et transitivité), l’index de petit monde, et le degré nodal ont été utilisés. Les variables cliniques d’inertie de sommeil, la somnolence, la dépression, l’anxiété, la sévérité d’hypersomnolence et la durée symptomatique ont été évaluées. Résultats et conclusion : Seul le groupe HI a démontré une efficacité globale significativement plus élevée que le groupe HC, qui pourrait être expliquée par une augmentation de la transmission GABAergique. Le groupe NT2 a démontré une association entre l’efficacité globale et les scores au ESS et HSI. Aucune autre différence significative n’a été démontrée pour les autres mesures globales ni pour les mesures régionales entre les groupes. / Introduction: Structural connectivity studies comparing all three central primary hypersomnolence disorder types, i.e., narcolepsy type I (NT1), type II (NT2) and idiopathic hypersomnia (IH) have been scarce. Differences in structural connectivity may indicate how circuits in the brain are affected in hypersomnolence disorders compared to healthy controls. The main goal of this study is to assess if differences exist between NT1, NT2 and IH patients compared to controls using structural imaging analyses. A decrease in global connectivity measures in all three central hypersomnolence disorders compared to healthy controls and an association to symptom severity like sleepiness is hypothesized. Methods: Patients diagnosed with either NT1 (n=15), NT2 (n=13) or IH (n=14), as well as healthy participants (HC, n=8) were recruited and underwent a DWI session. Data were preprocessed and analyzed using a model based on constrained spherical deconvolution with multi-shell and multi-tissue parameters. Integration (global efficiency) and segregation measures, small-world index and nodal degree were used to evaluate structural connectivity. Clinical variables of sleep inertia, sleepiness, hypersomnia severity, depression, anxiety, and disease duration were evaluated. Results and conclusion: Only IH patients showed a significantly higher global efficiency than HC. An increase in GABAergic transmission is a potential mechanism, but how it relates to IH symptoms is unclear. Global efficiency was also demonstrated to be associated with ESS and HSI scores in the NT2 group. There were no differences between groups for any of the other global or local measures.

Biomedical optics is a rapidly expanding field of research focusing on the development of methods to detect, diagnose, and treat disease using light. While there are a myriad of optical systems that have been developed for biological tissue imaging, optical diagnostics, and optical therapeutics, all of these methods suffer severely limited penetration depths due to attenuation of light by tissue constituent chromophores, including cells, water, blood, and protein structures. Tissue optical clearing is a recent area of study within biomedical optics and photonics, where chemical agents have been used to alter tissue optical properties, reducing optical absorption and scattering and enabling light delivery to and collection from deeper tissue regions. However, there are concerns as to the safety and efficacy of these chemical clearing agents in vivo, especially in the skin, where the projective barrier function of the stratum corneum must be removed. Mechanical optical clearing is a recently developed technology which utilizes mechanical loading to reversibly modify light transport through soft tissues, and much of the work published on this technique has focused on applications in skin tissue. This clearing technique enables deeper light delivery into soft tissues but does not require use of exogenous chemicals, nor does it compromise the skin barrier function. While this clearing effect is thought to be resultant from interstitial water and blood transport, the underlying mechanism has not been concretely identified nor characterized. The hypothesis of this body of work was that interstitial transport of tissue chromophores (e.g. water and blood) causes intrinsic optical property changes, reduces tissue optical absorption and scattering, and improves light delivery in diagnostic applications. To test this hypothesis, we first developed a mathematical framework to simulate mechanical optical clearing, using both mechanical finite element models and optical Monte Carlo simulations. By directly simulating interstitial water transport in response to loading, data from mechanical simulations was combined with optical Monte Carlo simulations, which enabled prediction of light transmission measurements made during mechanical indentation experiments. We also investigated changes in optical properties during mechanical indentation using diffuse reflectance spectroscopy. These studies used controlled flat indentation by a fiberoptic probe to dynamically measure intrinsic optical properties as they changed over time. Finally, we apply mechanical optical clearing principles to functional near-infrared spectroscopy for neuroimaging. By building a prototypical mechanical optical clearing device for measuring cerebral hemodynamics, we demonstrated that mechanical optical clearing devices modify measured cerebral hemodynamic signals in human subjects, improving signal quality. / Ph. D.

Patients with MDD demonstrate impairments in various components of affective processing, which are believed to persist in the remitted phase of the illness and are believed to underlie the vulnerability for future relapse. Despite advances in neuropsychiatry, the neuroanatomical site of action of various treatment modalities remains unclear, leaving clinicians without an algorithm to guide optimal treatment selection for individual patients. This thesis sought to characterize differences in brain activation during affective processing between MDD treatment responders (RS) and non-responders (NR) by combining clinical and neuroimaging variables in a repeat-measure functional magnetic resonance imaging (fMRI) investigation. We induced increases in positive and negative affect using visual stimuli under fMRI conditions in 21 MDD subjects and 18 healthy controls (HC). Based on previous neuroimaging investigations and preclinical animal data, we hypothesized that increased activation of the amygdala and the pregenual cingulate during negative affect induction (NAI), and decreased activity of the ventral striatum during positive affect induction (PAI), would differentiate ultimate NR from RS. Following the first scan, treatment with fluoxetine and olanzapine was initiated in the MDD group, with follow-up scans at one- and six-weeks thereafter. We hypothesized that decreases in depressive symptoms would be associated with decreased activation of the ventromedial prefrontal cortex (PFC) and amygdala during NAI and increased activation of the hippocampus during PAI. Eleven MDD subjects met criteria for clinical remission at study endpoint. Based on trait differences between MDD and HC, we hypothesized that differences observed during NAI would be limited to brain regions involved in regulation of the affective state, including the dorsolateral PFC and the anterior midcingulate cortex. The results of the analyses confirmed the a-prior hypotheses and additionally demonstrated differential activation of the insular, medial temporal, and premotor cortex during repeat PAI and NAI between HC, RS, and NR. These findings provide: i) a neuroanatomical target of successful antidepressant therapy during PAI/NAI; ii) a differential effect of depressive symptoms and dispositional affect on brain activation during PAI/NAI; and iii) an a-prior method to differentiate RS from NR, and iv) demonstrate the need for additional treatment to prevent relapse in the remitted state.

This dissertation is comprised of four studies investigating the hypothesis that adults and children who stutter differ from their same-age fluent peers in the neuroanatomy and neurophysiology underlying auditory speech processing. It has been consistently reported that adults who stutter demonstrate unique functional neural activation patterns during speech production, including reduced auditory activation, relative to nonstutterers. The extent to which these functional differences are accompanied by abnormal morphology of the brain in stutterers is unclear. The first study in this dissertation examined the neuroanatomical differences in speech-related cortex between adults who do and do not stutter using magnetic resonance imaging and voxel-based morphometry analyses. Adults who stutter were found to have localized grey matter volume increases in auditory and motor speech related cortex. The second study extended this line of research to children who stutter, who were found to have localized grey matter volume decreases in motor speech related cortex. Together, these studies suggest an abnormal trajectory of regional grey matter development in motor speech cortex of people who stutter. The last two studies investigated the mechanism underlying the repeated findings of reduced auditory activation during speech in people who stutter in more detail. Magnetoencephalography was used to investigate the hypothesis that people who stutter have increased speech induced suppression of early evoked auditory responses. Adults and children who stutter demonstrated typical levels of speech induced suppression relative to fluent peers. However, adults and children who stutter showed differences from peers in the timing of cortical auditory responses. Taken together, the studies demonstrate structural and functional abnormalities in brain regions related to auditory processing and point to the possibility that people who stutter have difficulty forming the neural representations of speech sounds necessary for fluent speech production.