Two of the top killers of the human race are cardiovascular disease and type 2 diabetes. These diseases, which have shared etiology, collectively form the domain of cardiometabolic health. Multiple functions of cardiometabolic health are perturbed by impaired sleep, and conversely, restored by increased sleep quantity and/or improved sleep quality. That is, impairments in sleep are closely related to impairments in cardiometabolic health. This leads to the proposition that specific features of sleep act as biomarkers for specific features of cardiometabolic health. This thesis aims to determine the impact of sleep on three major domains of cardiometabolic health: 1) atherosclerosis (arterial plaque accumulation), 2) dyslipidemia (unhealthy blood lipids), and 3) hyperglycemia (high blood sugar). Accordingly, three key findings emerge that comprise the three chapters of this report. 1) The quality of human sleep, specifically the degree of fragmentation, raises inflammatory-related white blood cells, thereby conferring an increased risk for atherosclerosis. This was true of sleep fragmentation assessed across a week or a single night, which predicted increasingly higher CAC scores through a mediating association with increased neutrophils. 2) Unique features of an individual’s sleep (duration, efficiency, timing, regularity) predict aspects of triglyceride metabolism under fasting (homeostatic) and post-food intake (allostatic) conditions. First, later sleep timing predicts higher fasting triglycerides, emphasizing a unique relation with triglyceride homeostasis (not allostasis). Second, this homeostatic link is mediated by increased systemic inflammation (GlycA levels). Third, and conversely, declining sleep efficiency negatively impacts post-food lipid clearance, reflecting allostatic regulation. Fourth, these effects are replicated in a second independent cohort. 3) The coupling of NREM sleep spindles and slow oscillations the night prior is associated with improved next-day peripheral glucose control. Further, this sleep-associated glucose pathway may influence glycemic status through altered insulin sensitivity, rather than through altered pancreatic beta cell function. Moreover, these associations are replicated in an independent dataset of over 1900 adults. Of therapeutic significance, the coupling between slow oscillations and spindles was the most significant sleep predictor of next-day fasting glucose, even more so than traditional sleep markers, relevant to the possibility of an EEG index of hyperglycemia. Collectively, beyond these scientific insights, these results support a sleep–cardiometabolic homeostasis framework, such that unique features of sleep act as biomarkers for unique aspects of cardiometabolic health. Importantly, these findings help reinform public health sleep-related guidelines to reduce the mortality and economic burden of cardiometabolic disease globally.

Impaired sleep is a frequent symptom affecting patients with Alzheimer’s disease (AD). Recent work has shown that sleep disruption occurs very early in the progression of AD, before the presence of significant cognitive decline. The accumulation of tau neurofibrillary tangles and β-amyloid (Aβ) plaques, the hallmark pathologies of AD, have been shown to impact sleep electrophysiology in rodent models. Furthermore, experimental disruption of sleep accelerates AD progression, suggesting that this relationship is bi-directional. With the recent advent of PET radiotracers that bind to in vivo tau and Aβ in the human brain, it is now possible to study how the presence of these proteins affect sleep physiology in healthy older people. Is disrupted sleep an early marker of the pathogenesis of AD? In this dissertation, I address this question by investigating associations between a diverse set of sleep metrics and PET-measured tau and Aβ burden in cognitively healthy older adults. Chapter 1 demonstrates that greater tau and Aβ are associated with distinct impairments in sleep oscillatory signatures, and that self-reported changes in sleep across the lifespan are predictive of pathological burden in late-life. Chapter 2 utilizes a longitudinal design to establish that impairments in both micro-electrophysiological and macro-architectural measures of sleep predict the subsequent rate of Aβ accumulation. Chapter 3 describes associations between tau and Aβ and real life patterns of sleep behavior, and shows that individuals with greater pathological burden are worse at estimating their own sleep quality. Together, these data illustrate the ways that sleep, measured objectively and subjectively, in the sleep lab and at home, is affected in the earliest stages of AD. These findings indicate the sensitivity of human sleep physiology to AD, and suggest that sleep may provide a non-invasive biomarker of tau and Aβ burden in cognitively normal older adults. Most importantly, while more work is needed to confirm these findings, these relationships offer the possibility that improving sleep quality may slow the progression of AD.

Recent evidence suggests there is an intimate and causal relationship between sleep and anxiety. Despite such progress, several unknowns remain. For example which factors predispose some individuals to be vulnerable to the anxiogenic impact of sleep loss while others appear to be resilient have yet to be identified. Furthermore, whether trait-anxiety, in turn, predicts the sensitivity of individuals to amplified emotional brain reactivity associated with a lack of sleep is similarly unknown. Moreover, little is currently known about the embodied interplay between peripheral and central nervous system mechanisms leading to such abnormalities of affective processing caused by sleep deprivation. Characterizing these mechanisms is necessary not only to gain a deeper understanding of the pathophysiological pathways underlying the condition of anxiety, but also for developing effective treatments for the amelioration of anxiety as well as guiding public health policy aimed at anxiety disorder prevention. Targeting these unanswered questions, this thesis combines functional and structural MRI techniques, together with high-density EEG recordings, to test the overarching hypothesis that sleep deprivation leads to dysregulation of the extended-limbic system contributing to, and interacting with, the state of anxiety. Three specific predictions emerged from this overarching hypothesis and were tested in separate experiments. First, experiment 1 confirms the hypothesis that structural brain morphology in a network of limbic brain regions predicts vulnerability to the anxiogenic impact of sleep deprivation, and that sex moderates this interaction. Second, experiment 2 provides evidence supporting the hypothesis that trait anxiety determines the degree to which sleep deprivation amplifies limbic brain reactivity during the anticipation of potentially aversive emotional experiences. Finally, results from experiment 3 affirms the prediction that, beyond central limbic brain changes, sleep deprivation additionally dysregulates peripheral, autonomic cardiac signaling in response to affective stimuli as well as decouples the normally inter-related association between central brain and peripheral nervous systems. Collectively, these results help characterize a framework in which sleep deprivation contributes to anxiety symptomology through impairments in affective processing by both the central (limbic) and peripheral (autonomic) nervous system functioning. Moreover, these data suggest that the co-morbid features of sleep disruption and altered limbic as well as autonomic function commonly reported across anxiety disorders may be causally related. Considering the continued decreases in sleep time across society and the high prevalence of anxiety disorders, these findings have significant therapeutic, clinical and public health ramifications.

Sleep is fundamental to life. Multiple biological functions, from peripheral immune and cardiovascular systems, to central brain operations governing cognition and emotion, are perturbed by sleep when deficient and restored by sleep when sufficient. This evidence leads to the proposition that sleep is necessary for coordinating whole-organism adaptation, which requires the active co-regulation of both brain and body systems for optimal physiological stability during stress, termed allostasis. Although causal links within and between the brain and body are recognized as essential, a multi-system examination of these interactions under conditions of insufficient sleep has yet to be reported. Elucidating the interdependence of the multi-system impacts of sleep loss is not only necessary to build a deeper understanding of the function of sleep, but also to guide informed treatments for a number of chronic diseases characterized by both sleep disturbances and multi-system etiology, such as chronic pain and cardiovascular disease. This thesis combines functional magnetic resonance brain imaging with measurements of peripheral physiological function in order to test the overarching hypothesis that the dysfunction observed across multiple brain and body systems under conditions of sleep loss are intimately interrelated, rather than independent. Three specific aims emerge from this hypothesis and are addressed in this thesis: (1) The first aim sets forth the hypothesis that sleep loss impairs the function and architecture of brain networks with consequences for cognition, affect, and physiology. A comprehensive review of neuroimaging studies of sleep deprivation in humans demonstrates that sleep loss alters the architecture of human brain networks, and with it, the processes dependent on these networks. Many of the brain regions and networks most impacted by sleep deprivation are affective and viscerosensory brain regions involved in the master regulation of allostasis. (2) The second aim addresses the hypothesis that, in addition to the central brain, sleep loss impairs the functioning of peripheral adaptive systems and that central and peripheral dysfunction are interrelated, focusing on the immune system and pain. An experimental sleep deprivation study in humans demonstrates that sleep deprivation increases levels of circulating pro-inflammatory cytokines, and that this pro-inflammatory state contributes to increased pain perception assessed both objectively in the brain and subjectively through participant ratings. (3) The third aim tests the hypothesis that interrelated dysfunction in the brain and the body is evident in a different model system involved in allostasis, the cardiovascular system. A second sleep deprivation experiment establishes significant interdependence between sleep loss-increased blood pressure, disruption of central brain viscerosensory networks, and worse mood state. Collectively, these results support an embodied framework of sleep loss wherein sleep loss contributes to disease through a disruption of the brain-body coordination required for allostasis. Considering the continual erosion of sleep across society, these findings hold important scientific, clinical and public health implications.

Sleep is highly conserved across phylogeny, yet marked inter-individual differences in sleep physiology are expressed. In humans, such variability is especially evident for non-rapid eye movement (NREM) sleep spindle and slow wave oscillations, representing trait-like physiological "fingerprints". Nevertheless, the mechanisms accounting for these differences, and the functional consequences of this variability, remain largely unknown. This thesis combines structural MRI, sleep EEG, and cognitive memory paradigms to test the hypothesis that brain morphology represents one candidate factor explaining inter-individual differences in NREM oscillations, which, in turn, consequently predict functional learning and memory. From this central hypothesis, three inter-related, yet distinct, studies are presented. Study 1 demonstrates that inter-individual variability in NREM physiology is accounted by grey matter morphology in functionally relevant brain regions. Specifically, grey matter in the sleep-regulating center of the basal forebrain/hypothalamus, together with the medial prefrontal cortex, accounted for individual differences in NREM slow waves. In contrast, grey matter in interoceptive and exteroceptive cortices predicted slower NREM sleep spindles, whereas grey matter volume in bilateral hippocampus was associated with faster sleep spindles. Study 2 establishes that the substructure of the hippocampus--specifically, the CA3/DG subfield--determines inter-individual levels of learning impairment following sleep deprivation, yet conversely interacts with post-deprivation recovery NREM sleep in predicting learning restoration. Finally, Study 3 describes a beneficial influence of fast NREM sleep spindles on differentially gating the selective offline consolidation of memories, based on prior waking instructions, leading to the selective remembering and forgetting of discreet item information. Together, these data indicate that macroscopic differences in structural brain morphology represent one mechanism accounting for individual differences in NREM sleep physiology, as well as associated consequences for learning and memory. More generally, these findings offer translational relevance to the structural brain abnormalities present in conditions such as dementia, sleep apnea, attention-deficit hyperactivity disorder, and post-traumatic stress disorder, where sleep disruption and memory impairment are highly co-morbid.

Despite an emerging link between alterations in motivated behavior and a lack of sleep, the impact of sleep deprivation as well as the potential benefit of sleep, when it is achieved, on human brain mechanisms of reward and punishment remain largely unknown. Targeting these unanswered questions, this thesis aims to determine the impact sleep loss on human brain incentive processing in topics of 1) food desire and 2) monetary gains and losses. From these investigations four main findings have emerged that make up the first four chapters of this report: 1) In the context of food choices, sleep deprivation leads to significantly decreases activity in appetitive evaluation regions within the human frontal cortex and insular cortex during food desirability choices, combined with a converse amplification of activity within the amygdala ultimately leading to increased high calorie food choices. 2) In the context of monetary rewards and losses, sleep deprivation did not lead to measurable changes in activation to the anticipation of rewards and losses in the nucleus accumbens or insula respectively. However, sleep deprivation did lead to diminished medial prefrontal cortex responses to gain outcomes and increased anterior insula response to loss outcomes. 3) Although there were no observable sleep deprivation group effects on anticipation of monetary gains and losses, there was a significant interaction of sleep deprivation with a trait dopamine transporter genetic polymorphism that determined the impact of sleep deprivation on anticipatory response to gain and loss in the nucleus accumbens and anterior insula respectively. 4) In the context of learning from monetary incentives, sleep deprivation led to a specific deficit in the ability to learn from monetary gains but no change in the learning profile for monetary losses. Finally, in addition to these four findings on the impacts of sleep deprivation on human incentive processing, the final chapter (5) focuses on the relationship between sleep, when it is achieved, and next day reward responsivity. Here there is a significant relationship between individual differences in REM sleep beta power and next day nucleus accumbens response to anticipation of rewards. Beyond these basic scientific insights, such results offer potential clinical relevance given that sleep disruption is highly co-morbid with numerous psychiatric and neurological conditions associated with dysfunctional dopaminergic reward processing (e.g., Parkinson's disease, substance abuse, and obesity). Such findings further indicate that sleep intervention may represent an under appreciated and novel therapeutic target particularly for disorders of the reward system.

Folk wisdom has long implied a critical role for sleep in emotional processing, highlighted by sayings as `Sleep on it and you'll feel better in the morning'. However, until very recently there was a lack of objective research investigating this claim. Now rapidly emerging evidence describes a powerful and causal relationship between sleep and affective brain regulation. These findings are mirrored by long-standing clinical observations demonstrating that nearly all mood and anxiety disorders co-occur with one or more abnormalities of sleep. This dissertation aims to (1) set forth a rapid-eye movement (REM) sleep hypothesis of affective brain homeostasis, optimally preparing the organism for next-day social and emotional functioning, (2) provide behavioral evidence of detrimental impairments in emotional functioning following sleep deprivation, and (3) outline affective brain and behavioral benefits of sleep when it is obtained. Finally, the discussion will describe how this model may explain the prevalent relationships observed between sleep and affective disorders, including relevant treatment mechanisms, with a particular focus on post-traumatic stress disorder (PTSD).