| || |
• In insomnia, brain anatomy alterations in the frontoparietal network have been reported but are not consistent across studies, perhaps reflecting the complexity and heterogeneity of insomnia and suggesting that sleep-related symptoms in insomnia are reflected in functional brain alterations rather than structural brain alterations.
| || |
• Insomnia is characterized by a hyperactivation in response to sleep-related and negative emotional stimuli and a hypoactivation in response to other types of stimuli, such as tasks involving executive control.
| || |
• Brain activity and connectivity during resting wakefulness is altered in insomnia, particularly within the default mode and salience networks and between the limbic network and other brain regions. These alterations may contribute to the state of hyperarousal and emotional dysregulation that are present in insomnia.
| || |
• The overall activity decrease in the default mode network in the transition from wakefulness to sleep is less pronounced in insomnia and might reflect heightened arousal.
Definition of insomnia. Insomnia is characterized by difficulties falling asleep and maintaining sleep as well as waking up earlier than desired. It is associated with daytime function impairments, such as fatigue and reported cognitive impairments. The distinguishing features of insomnia have been associated with functional alterations that are, for the most part, in line with the hyperarousal theory of insomnia, which postulates that heightened physiological, cognitive, or emotional arousal observed in insomnia is the result of excessive central nervous system arousal and prevents and disturbs sleep.
Functional brain alterations during cognitive tasks in insomnia. Various cognitive and emotional tasks have been used to evaluate functional brain alterations in insomnia and generally point to a hyperactivation in response to sleep-related and negative emotional stimuli and a hypoactivation in response to other types of stimuli, such as tasks involving executive control (37; 39; 46; 63; 74).
The relationship between emotion and hyperarousal has been investigated in several fMRI studies. In a study, greater activation in response to emotional stimuli was present in the amygdala in individuals with insomnia when trying to reinterpret the meaning of negative pictures (21). However, this hyperarousal might be specific to sleep and insomnia as only emotionally arousing stimuli that were related to sleep elicited greater amygdala activation in an emotional reactivity task in individuals with insomnia (04). In addition, the habituation to emotional images that was observed in good sleepers over successive trials was absent in individuals with insomnia (04).
The impairment of daytime functioning is a key feature of insomnia and mainly affects subjective cognitive functioning. Despite inconsistent evidence for objective cognitive impairments, brain cognitive processing as shown by fMRI seems to be altered. Individuals with insomnia had reduced activation of the left medial and inferior frontal gyri during a verbal fluency task involving the executive control network, without showing poorer performance (03). During a working memory task, lower activation in the frontal gyrus and motor and cingulate cortices was observed in individuals with insomnia; they also failed to deactivate the default mode network (19).
Functional brain alterations during resting wakefulness in insomnia. The growing literature on resting wakefulness in insomnia shows an altered brain activity and connectivity, particularly within the default mode and the salience networks and between the limbic network and other brain regions, which are thought to contribute to hyperarousal and emotional regulation alterations that are present in insomnia. Resting state fMRI functional connectivity was found to be lower in the salience and executive control networks and higher in the sensorimotor network (37). In the salience network, the activation of which is associated with negative internal and external stimuli, functional connectivity was associated with negative effect in individuals with insomnia (13; 63). In the limbic network, functional connectivity was decreased between the thalamus and left amygdala, parahippocampal gyrus, putamen, pallidum, and hippocampus (74) as well as between the amygdala and the insula, striatum, and thalamus in individuals with insomnia (30). The latter study also found increased connectivity between the amygdala and the premotor and sensorimotor cortices, which was associated with decreased sleep quality. Altered functional connectivity within the limbic network may be reflective of altered emotional processing in insomnia, and the increased functional connectivity between the amygdala and sensorimotor cortices may support a model of excessive hyperarousal.
In the default mode network, inconsistent alterations of functional connectivity and regional brain activity in the cingulate cortex, left insula, left cuneus, and left fusiform gyrus were also observed. Indeed, some studies showed increased connectivity and activation in the default mode network, whereas other showed decreased activity (37). Altered functional connectivity in the default mode network was often associated with subjective sleep disturbances. Additionally, decreased functional connectivity between the medial prefrontal cortex and middle temporal lobe and between the left temporal lobe and left inferior parietal cortex was found in individuals with insomnia (51). Because these regions play an important role in cognitive functioning and sleep physiology, these findings suggest that decreased connectivity may underlie the alteration in cognitive functioning, whereas increased activity at rest might reflect a hyperarousal state in individuals with insomnia. The reduction in brain activation during resting wakefulness has been confirmed using FDG-PET, showing an overall reduction in glucose metabolism in individuals with insomnia. More specifically, an altered engagement of the limbic, default mode, and executive control networks might be reflected by lower middle frontal, fusiform, lingual gyri, precuneus/posterior cingulate, and frontoparietal cortex glucose metabolism in insomnia (38).
Functional brain alterations during sleep and sleep-wake transitions in insomnia. Significant brain activity changes occur during sleep: the overall level of activity decreases in the default mode network, and there is a disconnection (functional connectivity decrease) between its posterior and anterior parts. More specifically, the posterior cingulate cortex and inferior parietal lobe disconnect from the medial prefrontal cortex and anterior cingulate cortex during sleep (28). This disconnection becomes more pronounced as sleep deepens and reaches its lowest level during slow-wave sleep. In individuals with insomnia, this overall activity decrease that happens during sleep seems less pronounced as the decline in regional brain glucose metabolism from wakefulness to sleep is smaller in the medial prefrontal cortex, anterior cingulate, insula, thalamus, hippocampus, amygdala, hypothalamus, and brainstem (52). This was confirmed by another FDG-PET study in a larger sample of individuals showing a smaller decline in brain glucose metabolism during NREM sleep in the default mode network (38). Alterations in other networks are present during sleep in individuals with insomnia as fMRI functional connectivity between the thalamus and left amygdala, parahippocampal gyrus, putamen, pallidum, and hippocampus was decreased across all 3 NREM sleep stages.
Summary findings on insomnia
(A) Summary of findings from fMRI studies in insomnia. Overall, there is an increased activation in response to sleep-related and emotional stimuli in the left temporal (including amygdala) and occipital lobes and right frontal...
Structural brain alterations in insomnia. Brain anatomy alterations are inconsistent in individuals with insomnia, potentially because insomnia is a complex and heterogeneous disorder. The discrepancy of results as well as the lack of large sample data make it difficult to draw a clear picture of structural alterations in insomnia and suggest that sleep-related symptoms in insomnia are reflected in functional brain alterations rather than structural brain alterations. A large sample study including 1053 patients with major depressive disorder from the enigmatic major depressive disorder consortium showed that insomnia severity was specifically related to smaller cortical surface area in the right insula, left inferior frontal gyrus pars triangularis, left frontal pole, right superior parietal cortex, right medial orbitofrontal cortex, and right supramarginal gyrus and was independent of major depressive disorder (43). In addition, cortical thinning was found in the anterior cingulate cortex, precentral cortex, and lateral prefrontal cortex in individuals with insomnia disorder (66).
Previous smaller sample studies had inconsistent findings regarding gray matter volume. Hippocampal gray matter volume was found to be reduced in individuals with insomnia in one study, but this finding was not replicated in other studies. The prefrontal regions (orbitofrontal and dorsolateral prefrontal cortex) were found to be smaller in individuals with insomnia in 2 studies, but this was not replicated in a latter study. Increased anterior cingulate cortex gray matter volume was found to be associated with increased insomnia severity in individuals with insomnia, but this was not replicated in a more recent larger study.
Fewer studies have examined white matter alterations. Overall, white matter integrity was affected in frontostriatal, frontothalamic, and cortico-cortical networks as well as within the limbic network (mostly shown by reductions in fractional anisotropy or less restricted water diffusion) in individuals with insomnia (50). Lower tract strength, as a marker of poorer white matter integrity, was observed within the limbic network and was negatively correlated with sleep quality in individuals with insomnia (12), showing the importance of emotion regulation in insomnia.
Brain metabolite alterations in insomnia. In terms of neurochemistry, the main metabolite that has been investigated in insomnia is GABA, which is involved in sleep-wake regulation as demonstrated by the use of benzodiazepines (positive allosteric modulators of the GABAA receptor) for the short-term treatment of insomnia. Individuals with insomnia may present an overall reduced inhibitory tone as lower GABA levels have been found in the anterior cingulate and in parieto-occipital cortices. However, higher GABA levels have been found in the occipital cortex (37). Only a handful of studies have investigated other metabolite levels, without consistent results. Phosphocreatine was decreased in individuals with insomnia, suggesting altered cortical energetic demands (25). Aspartate, glutamine, and creatine were decreased in the lateral occipital cortex in individuals with insomnia and short sleep duration (48). Glutamate, an excitatory neurotransmitter, was increased in the parieto-occipital cortex and was associated with insomnia symptoms in patients with posttraumatic stress disorder (37).
Narcolepsy and idiopathic hypersomnia
| || |
• In narcolepsy with cataplexy, structural alterations, such as smaller gray matter volume and altered white matter integrity, are widespread and encompass the hypothalamus and its projections, which is in line with the evidence for loss of hypocretin neurons in the hypothalamus.
| || |
• Functional alterations in narcolepsy with cataplexy are observed in the hypothalamus, which shows altered activation and regional blood flow during both resting wakefulness and while performing cognitive tasks, such as emotion regulation during positive emotions or cataplexy attacks.
| || |
• Idiopathic hypersomnia does not show similar structural alterations as compared to narcolepsy with cataplexy but seems to present a disruption of the default mode network, both at the structural level and at the level of functional connections within regions of the default mode network.
Definition of narcolepsy and idiopathic hypersomnia. Narcolepsy is characterized by excessive daytime sleepiness and abnormalities in REM sleep. Narcolepsy is categorized into 2 types: narcolepsy with cataplexy (type 1) and narcolepsy without cataplexy (type 2). Narcolepsy with cataplexy is associated with a deficiency in the hypothalamic neuropeptide orexin-A (hypocretin-1), which is involved in the abnormal sleep-wake patterns and cataplexy that are characteristic of narcolepsy. Idiopathic hypersomnia is primarily characterized by excessive daytime sleepiness, together with an often-prolonged total sleep time and difficulties waking up (sleep drunkenness), for which the pathophysiological mechanisms remain unclear. Individuals suffering from idiopathic hypersomnia do not present cataplexy, rapid REM sleep onset (at the multiple sleep latency test), or any consistent hypocretin-1 deficiency.
Functional brain alterations during cognitive tasks in narcolepsy. Based on clinical observations that cataplectic episodes in narcolepsy can be triggered by positive emotions, brain processing involved in emotion and reward have been studied (14). Two fMRI studies used humorous images to study emotional processing and showed alterations in the amygdala and hypocretin system. The first study found reduced hypothalamic activity and increased amygdala activity (64). The second study found enhanced activity in the hypothalamus and emotional network (58). Reward processing was altered in individuals with narcolepsy as they failed to show functional connectivity between the amygdala and medial prefrontal cortex (56). The investigation of other tasks, such as attention or working memory, suggest that narcolepsy is characterized by an imbalance of cognitive resources allocation in favor of monitoring and maintaining attention (23).
Functional brain alterations during resting wakefulness in narcolepsy. Investigation of resting wakefulness in narcolepsy suggests decreased activity in regions that are important targets of hypocretinergic projections (14). An example is decreased regional cerebral blood flow in the brainstem in individuals with narcolepsy (47). Another study found decreased perfusion in the hypothalamus, caudate nuclei, thalamus, cingulate gyrus, and frontoparietal cortex in individuals with narcolepsy with cataplexy (34). However, there have been inconsistent findings regarding glucose metabolism: a first study found hypometabolism in the bilateral hypothalami, thalamus, and frontal-parietal cortex (36) in individuals with narcolepsy, whereas a second study only found hypermetabolism in the cingulate and visual association cortices (17).
Functional brain alterations during sleep in narcolepsy. Findings during REM sleep are inconsistent: some studies report increased activity or an absence of decreased activity that is supposed to occur during sleep (14). A later study found decreased activity in the hypothalamus, caudate nuclei, thalamus, cingulate gyrus, and frontoparietal cortices in narcolepsy with cataplexy (34).
Summary findings on narcolepsy
(A) Summary of findings from fMRI studies in narcolepsy. During winning on a gambling task, activity is reduced in the bilateral nucleus accumbens and the ventromedial prefrontal cortex, and activity is increased in the putamen...
Functional brain alterations during cataplexy in narcolepsy. Because it is difficult to capture cataplectic episodes, only case studies have reported such results, and they show inconsistent findings. Hyperperfusion has been found in limbic areas (including the amygdala and orbitofrontal and anterior cingulate cortices), the thalamus, basal ganglia, brainstem, and parietal cortices (27), whereas hypoperfusion has been found in the prefrontal and occipital cortices (27) and precentral and primary somatosensory cortex (17). The latter observation was supported by an fMRI study that showed marked hypoactivation of the hypothalamus during a cataplectic attack (58).
Functional brain alterations in idiopathic hypersomnia. Only 2 neuroimaging studies have investigated functional alterations in idiopathic hypersomnia. Although these findings require replication, they suggest functional alterations in brain areas involved in the modulation of vigilance states rather than a loss of hypothalamic orexin neurons. One study found that glucose metabolism was increased in the insula, anterior and middle cingulate cortex, and caudate nucleus (18). Another study showed that regional cerebral blood flow was lower in the medial prefrontal cortex as well as in the posterior cingulate cortex, putamen, and cerebellum in individuals with idiopathic hypersomnia (08). Using fMRI, the same group showed lower functional connectivity within the anterior part of the default mode network (ie, medial prefrontal cortex), which was also associated with sleepiness (55).
Structural brain alterations in narcolepsy. Early structural studies of narcolepsy focused on the pontine tegmentum region in the brainstem as this structure controls the transition between vigilance states, but they found inconsistent results that may be due to heterogeneity in patient characteristics or neuroimaging techniques. Later, in line with evidence suggestive of a hypocretin deficiency, several studies investigated gray matter volume of the hypothalamus. Meta-analyses of those studies have shown inconsistent alterations of thalamic and hypothalamic structures in narcolepsy (24). However, consistent gray matter volume decreases in frontotemporal regions, which might be associated with cognitive complaints of memory and attention, have been observed in narcolepsy with cataplexy. Additionally, alterations in white matter integrity have been found in the hypothalamus-thalamus-orbitofrontal pathway and in the brainstem as well as in the inferior fronto-occipital fasciculus. However, no studies have specifically investigated structural brain alterations in narcolepsy without cataplexy.
Structural brain alterations in idiopathic hypersomnia. Fewer studies have investigated structural brain alterations in idiopathic hypersomnia, and those few reports show no global gray or white matter volume alterations. However, regional brain alterations have been found. Cortical thickness and gray matter volume are larger in the posterior default mode network and in the middle occipital gyrus in individuals with idiopathic hypersomnia. More specifically, the volume of the precuneus was positively associated with greater subjective daytime sleepiness (55).
Brain metabolite alterations in narcolepsy. The investigation of brain metabolite levels in narcolepsy suggests decreased N-acetylaspartate/creatine levels in the hypothalamic region, which is consistent with evidence that hypocretinergic neurons are selectively damaged in narcolepsy (67). However, the thalamus and parieto-occipital cortex, which are also innervated by those neurons, did not reveal any alterations in N-acetylaspartate/creatine levels (68). Finally, absolute GABA concentrations were found to be higher in the medial prefrontal cortex in individuals with narcolepsy, and this was positively associated with nocturnal sleep quality (41).
REM sleep behavior disorder
| || |
• Because of the link between REM sleep behavior disorder and Parkinson disease, structural alterations were primarily investigated in the pons and midbrain as well as in the basal ganglia, all of which involve dopamine transmission. However, structural alterations in REM sleep behavior disorder also encompass temporal, frontal, and parietal cortices.
| || |
• Functional alterations are widespread across the cortex and involve the limbic network, sensorimotor network, pons, and thalamus. More specifically, alteration of functional connectivity within the basal ganglia may be an important marker of basal ganglia network dysfunction.
Definition of REM sleep behavior disorder. REM sleep behavior disorder is a disorder characterized by the loss of muscle atonia during REM sleep, which leads to dream enactment that often results in dramatic movements, such as punching, flailing, or jumping out of the bed, as well as injuries and problems with the bed partner. There is increasing evidence that REM sleep behavior disorder is associated with the alpha-synucleinopathies, such as Parkinson disease, dementia with Lewy bodies, and multiple system atrophy; REM sleep behavior disorder is considered a risk factor for alpha-synucleinopathies (33).
Functional brain alterations in REM sleep behavior disorder. Decreased perfusion has been found in the frontal and temporoparietal cortices of individuals with REM sleep behavior disorder; decreased cerebral blood flow has been found in the parietal and occipital cortices as well as in the cerebellum and limbic network (14). Increased perfusion has been observed in the pons, putamen, and hippocampus. The hyperperfusion within the hippocampus predicted the development of dementia with Lewy bodies or Parkinson disease 3 years later, and this occurred in half of the patients (16). REM sleep behavior disorder, a risk factor for alpha-synucleinopathies (33), suggests a dysfunction in dopaminergic transmission, which might play a role in the process leading to neurodegeneration. A few neuroimaging studies investigated the dopaminergic system and showed greater dopamine transporter density in individuals with REM sleep behavior disorder, whereas other studies showed no difference or lower dopamine transporter density (14). Longitudinal studies have shown that individuals who developed a neurodegenerative disorder had lower dopamine transporter at baseline (31).
In addition, functional connectivity is altered in the basal ganglia network (59) and in the motor cortex and sublaterodorsal tegmental nucleus (22) in individuals with REM sleep behavior disorder. Lower within-network functional connectivity has been identified between the striatal and prefrontal regions, within the executive control network, between the midbrain and pallidum, within the basal ganglia network, and between motor and somatosensory regions as well as within the sensorimotor network (71). Finally, global basal ganglia resting state connectivity in REM sleep behavior disorder was similar to individuals with Parkinson disease, suggesting that activity within the basal ganglia network may be a marker of basal ganglia dysfunction (59).
Summary findings on REM sleep behavior disorder
Summary of findings from PET and SPECT studies in REM sleep behavior disorder. During episodes of REM behavior disorder, perfusion is increased within the supplementary motor cortex. At wakefulness, perfusion is decreased in th...
Structural brain alterations in REM sleep behavior disorder. REM sleep behavior disorder is characterized by smaller gray matter volume in the thalamus; pontomesencephalic tegmentum; and temporal, frontal, and parietal cortices (22). More severe reductions in volume have been found in Parkinson disease with associated REM sleep behavior disorder (42). Only a couple of studies found increased volume in the caudate nucleus (26). This lower volume in the pontomesencephalic tegmentum is of importance because it contains cholinergic, GABAergic, and glutamatergic neurons implicated in the promotion of REM sleep and muscle atonia (09). Idiopathic REM sleep behavior disorder has also been associated with higher perivascular space burden in the centrum semiovale, basal ganglia, substantia nigra, and brainstem compared to individuals with Parkinson disease, which correlates with REM sleep behavior disorder symptom severity (65). White matter alterations such as higher fractional anisotropy, a marker of more restricted water diffusion, have been found in the pons (70) and other regions known to be involved in REM sleep regulation, such as the pontine region, (62), cerebellar peduncles, and brainstem (26).
Brain metabolite alterations in REM sleep behavior disorder. REM sleep behavior disorder does not seem to involve detectable metabolite alterations in mesopontine regions. Only one case study revealed an increased choline/creatine ratio in the brainstem (49); others failed to replicate these findings (32).
Obstructive sleep apnea
| || |
• Cognitive deficits observed in obstructive sleep apnea might be explained by lower resting state activity and functional connectivity in the frontoparietal, sensorimotor, and salience networks, which are involved in numerous cognitive processes.
| || |
• In contrast, the pattern of structural alterations in obstructive sleep apnea is less consistent. Gray matter atrophy has been observed in the frontal and parietal cortex, temporal lobe, anterior cingulate, thalamus, hippocampus, and cerebellum. White matter alterations encompass the corpus callosum, sensorimotor cingulate cortex, corticospinal tract, insular cortex, basal ganglia, and limbic network.
| || |
• Functional and structural changes may be partially reversible with continuous positive airway pressure treatment.
Definition of obstructive sleep apnea. Obstructive sleep apnea is characterized by recurrent episodes of partial or complete occlusion of the upper airway during sleep, which impedes breathing despite continued inspiratory efforts. The repetitive obstructive events that occur during sleep in obstructive sleep apnea result in variable changes in blood gas concentrations, which lead to hypoxemia (low blood oxygen) and hypercapnia (high blood carbon dioxide). Due to the body’s attempts to regain airflow, the apneic events are strongly correlated with arousals from sleep, leading to significant sleep fragmentation and a reduction in deeper sleep stages. It is well established that obstructive sleep apnea is associated with impairments in cognitive function, particularly in vigilance, attention, memory, and executive function domains. Importantly, obstructive sleep apnea can be controlled with treatments such as continuous positive airway pressure, which serves to maintain airway patency and prevent obstructions.
Functional brain alterations during cognitive tasks in obstructive sleep apnea. Because obstructive sleep apnea is associated with cognitive deficits, functional alterations are more easily detected while performing a cognitive task. A multimodal meta-analysis showed lower activation during various cognitive tasks in individuals with obstructive sleep apnea, specifically in the dorsolateral prefrontal and orbitofrontal cortices; higher activation was found in the insula, anterior cingulate cortex, amygdala, and hippocampus (29). This finding was based on small sample studies investigating different cognitive domains in which working memory, executive functioning, and attention performance were poorer in individuals with obstructive sleep apnea. Lower resting state functional connectivity in the sensorimotor network, salience (insula) network, and frontoparietal network and altered within-network coherence might explain the impaired cognitive and motor performance observed in individuals with obstructive sleep apnea (39). Continuous positive airway pressure treatment can reduce hypoxia and normalize blood oxygen saturation, thereby reducing sleep fragmentation and improving cognitive function. After continuous positive airway pressure treatment, restoration in abnormal baseline activation in the left inferior frontal gyrus, anterior cingulate cortex, and hippocampus (10) and an increase in task-related activation (57) were associated with a significant improvement in working memory performance.
Functional brain alterations during resting wakefulness in obstructive sleep apnea. Overall, studies show that there is a decrease in brain activity as indicated by decreased perfusion in the parietal cortex (20), bilateral parahippocampal gyri, and right lingual gyrus (35) in individuals with obstructive sleep apnea during resting wakefulness. There is also decreased metabolism in the precuneus, middle and posterior cingulate gyrus, parieto-occipital cortex, and prefrontal cortex (73). These decreases might be due to compensatory mechanisms in response to blood flow alterations or blood gas changes in obstructive sleep apnea and are partly reversible with continuous positive airway pressure treatment (40; 45).
Summary findings on obstructive sleep apnea
(A) Summary of findings from fMRI studies in obstructive sleep apnea. During tasks, activation is reduced in the dorsolateral prefrontal cortex, anterior cingulate cortex, middle frontal gyrus, inferior frontal gyrus, cingulate...
Structural brain alterations in obstructive sleep apnea. Structural alterations in obstructive sleep apnea might correspond to neuronal damage resulting from intermittent hypoxia. Reduced cortical thickness has been found in individuals with obstructive sleep apnea in the precentral and postcentral gyri, ie, sensorimotor areas, corresponding to the upper airway musculature areas as well as temporal cortex and insula (44). Cortical thinning in the hippocampus and entorhinal cortex has been found to be related to disease severity and can be slowed down by continuous positive airway pressure treatment (54). In contrast, other studies reported greater cortical thickness in frontal regions (06; 15), which were associated with obstructive sleep apnea severity and could, therefore, be reflective of edema and reactive cellular processes (06).
Gray matter volume decreases seem to be more widespread and encompass the prefrontal, postcentral gyrus, temporal, and parietal cortices as well as the anterior cingulate cortex, posterior hippocampus (CA3/dentate), thalamus, and cerebellum. Most of these changes have been associated with cognitive deficits or disease severity, and smaller hippocampal volume has been associated with excessive daytime sleepiness (72; 29; 05). In contrast, increases in gray matter volume seem to be restricted to the CA1, subiculum, and uncus parts of the hippocampus (72).
White matter integrity is also affected in individuals with obstructive sleep apnea. Mean diffusivity is lower (ie, water diffusion is more restricted) in the corpus callosum, sensorimotor cingulate cortex, corticospinal tract, insular cortex, basal ganglia, and limbic network (11; 60) and might be reversible after treatment with continuous positive airway pressure (54). In some individuals with obstructive sleep apnea, white matter hyperintensities can be observed on MRI scans and might reflect vascular damage that is associated with aging or comorbid hypertension (60).
Brain metabolite alterations in obstructive sleep apnea. The cognitive deficits in obstructive sleep apnea may also be related to the metabolite alterations in specific cerebral regions. The level of N-acetylaspartate/creatine has been found to be decreased in periventricular and frontal white matter (14), possibly reflecting poorer neuronal integrity and glial dysfunction. In addition, choline/creatine levels were increased in the thalamus (01) and temporal cortex (61), which might indicate reactive gliosis or maladaptive changes in membrane metabolism in response to hypoxia. In contrast, other studies have found increased N-acetylaspartate/creatine and choline/creatine ratios in the hippocampus, which correlated with greater obstructive sleep apnea severity (07; 02). Finally, some metabolite ratio alterations were reversible with continuous positive airway pressure treatment (53), whereas others were not (69).