Neuroimaging Insights into Insomnia Part 1

Abstract

Insomnia is a frequent symptom or syndrome defined by complaints of trouble in initiating or maintaining sleep or of nonrestorative sleep. This causes significant impairments in several areas of daytime functioning including mood, motivation, attention and vigilance.

Significant advances in our neurobiological knowledge of insomnia have been brought by electrophysiological data (e.g. electroencephalography (EEG) and by functional neuroimaging data (e.g. single photon emission computed tomography (SPECT), positron emission tomography (PET) acquired during wakefulness, transition from waking to non rapid-eye-movement (NREM) sleep and REM sleep itself.

Indeed it has been shown that idiopathic insomnia is characterized by a specific pattern of regional brain activity: (i) during the transition from waking to NREM sleep: failure to decrease brain activity in the ascending reticular activating system, medial prefrontal cortex, limbic/paralimbic areas (including insular cortex, amygdala, hippocampus, anterior cingulate), thalamus and hypothalamus, (ii) during NREM sleep: deactivation of the parietal and occipital cortices, and basal ganglia, and (iii) during wakefulness: deactivation in brainstem reticular formation, thalamus, hypothalamus, prefrontal, left superior temporal, parietal and occipital cortices.

This specific distribution of brain activity might relate to (i) specific impairments in daytime functioning (e.g. hypoactivity in prefrontal cortex during wakefulness is consistent with reduced attentional abilities), (ii) hyperarousal hypothesis as a common pathway in the pathophysiology of insomnia (e.g. overall cortical hyperarousal characterized by an increase in EEG beta/gamma activity (14-35 / 35-45 Hz) at sleep onset and during NREM sleep) and (iii) the potentially overlapping pathophysiology with major depressive disorder as this illness has shown similarly altered cortical patterns (e.g. both illnesses have impairments in limbic/paralimbic areas as well as in basal ganglia).

The goal of this topic is to show that combining recent neurophysiological and neuroimaging data on human sleep offers new insights into the pathophysiological mechanisms of insomnia and potentially opens new therapeutic perspectives.

Keywords: insomnia, sleep, REM, NREM, functional neuroimaging, cognitive neuroscience, hyperarousal, major depressive disorder, brain.

Introduction

In healthy humans, functional and structural neuroimaging has been successfully used to characterize normal stage and pathological conditions of sleep, as reviewed elsewere (Dang-Vu, Desseilles et al. 2007; Desseilles, Dang-Vu et al. 2008). Here we focus on the brain imaging studies devoted to insomnia.

Insomnia is characterized by complaints of repeated difficulty in initiating or maintaining sleep or of nonrestorative sleep, which cause clinically significant distress or impairment in cognitive, social, and occupational, or other important areas of functioning (Cortoos, Verstraeten et al. 2006). Insomnia therefore presents with subjective symptoms. Insomnia can arise directly from sleep/wake regulatory dysfunction or indirectly from comorbid behavioral, psychiatric, neurological, immune, or endocrine disorders, including disturbances secondary to the use of drugs. In this respect, insomnia appears to be a 24-h disorder because it is not restricted to sleep complaints alone but can affect several aspects of daytime functioning as well. Importantly, insomnia is a common disorder in our society, with 10% to 20% of the general population reporting insomnia complaints and related impairment of daytime functioning (Ohayon 2002). We should note that prevalence of insomnia increases with several factors such as: age (increase in older), gender (more frequent in women), occupational status (increase in people undergoing particular private or professional pressure), and medical condition (substance users, neurological or psychiatric comorbid condition). Insomnia can be either acute or chronic and either idiopathic or secondary to several conditions including physical disease or mental illnesses. Several manuals (American Psychiatric Association 1994; American Academy of Sleep Medicine 2005) propose a classification of different subtypes of insomnia but full description of these subtypes goes beyond the aim of our topic.

The hyperarousal hypothesis of insomnia has gained growing attention as an integrative approach to the mechanism of insomnia (Perlis, Giles et al. 1997). This hypothesis presupposes interplay between psychological and physiological factors in the onset and maintenance of insomnia. It postulates that subjects who tend to focus cognitively on the insomnia and start to ruminate about their sleep complaint are prone to perpetuate the disorder, especially when it is associated to maladaptive behaviors such as prolongation of bedtime or daytime napping.

We will first review the structural and functional imaging findings in insomnia. Then we will describe successively the functional imaging of drug response, daytime functioning impairment and the hyperarousal hypothesis. Because depression is often associated with insomnia (Tsuno, Besset et al. 2005) we review hereafter the data pointing to some common underlying neurophysiological mechanisms for both sleep and mood regulation.

Structural and Functional Imaging in Insomnia

Structural imaging make possible to detect small differences in brain morphology associated with a particular condition. In particular, voxel-based morphometry (VBM) is based on high-resolution magnetic resonance imaging (MRI) scans and allows comparisons of grey and white matter across the brain and between groups. Proton magnetic resonance spectroscopy (1H-MRS) allows to assess the regional brain content in different compounds such as gamma-aminobutyric acid (GABA).

Only one study has assessed the structural anatomy of idiopathic (or primary) insomnia by using VBM (Riemann, Voderholzer et al. 2007). Riemann and collaborators used MRI (1.5 Tesla) in 8 unmedicated patients suffering from chronic idiopathic insomnia (3 males; mean age (standard deviation) of 48.4 (16.3) years) and 8 good sleepers matched for age, sex, body mass index, and education level. They found that patients have a significant reduction of hippocampal volumes bilaterally (see Figure 1), as compared to the good sleepers (Riemann, Voderholzer et al. 2007). Because of the size of the study sample, the results should be interpretated with caution. However, findings are congruent with the empirical data on (i) sleep-dependent encoding capacity of the hippocampus (Walker 2009) and (ii) impaired sleep-related memory consolidation in idiopathic insomnia (Nissen, Kloepfer et al. 2006).

The first study on neurochemical differences in patients with insomnia was recently conducted using 1H-MRS in 16 non-medicated individuals (8 women) with idiopathic insomnia (mean age (SD) = 37.3 (8.1) years) and 16 (7 women) normal sleepers (37.6 (4.5) years) (Winkelman, Buxton et al. 2008). Average brain GABA levels were nearly 30% lower in patients as compared to controls and were negatively correlated with wake after sleep onset (WASO). Since GABA is a major inhibitory neurotransmitter, this result may be consistent with the increase of brain glucose metabolism in several areas covered by this 1H-MRS study, such as the thalamus (Nofzinger, Buysse et al. 2004). Nevertheless an important methodological limitation of this study is the lack of anatomical specificity that limits further interpretations.

Several functional imaging techniques make possible to assess regional brain activity at rest, between two distinct conditions during a task or in association with any physiological process. While single photon emission computed tomography (SPECT) and positron emission tomography (PET) show the distribution of radioisotope emitting single gamma photons or compounds labeled with positron-emitting isotopes, functional magnetic resonance imaging (fMRI) measures the variation in brain perfusion related to neural activity by using a method based on the assessment of the blood-oxygen-level-dependent (BOLD) signal. The latter reflects the relative decrease in deoxyhemoglobin concentration that follows the local increase in cerebral blood flow in an activated brain region.

To our knowledge, only a few studies have assessed the functional neuroanatomy of idiopathic insomnia disorder by recording brain activity during NREM sleep. Nofzinger et al. used 18fluorodeoxyglucose (18FDG) PET to measure regional brain metabolism (indexed by cerebral metabolic regional glucose consumption, CMRglu) in 7 patients with idiopathic insomnia and 20 healthy age-matched and gender-matched subjects during waking and NREM sleep (Nofzinger, Buysse et al. 2004). Insomnia patients showed a global CMRglu increase during the transition from waking to sleep onset as compared to healthy subjects, suggesting that there is an overall cortical hyperarousal in insomnia. Moreover, insomniac patients exhibited less reduction of relative CMRglu from waking to NREM sleep in the ascending reticular activating system, hypothalamus, insular cortex, amygdala, hippocampus, anterior cingulate, and medial prefrontal cortices, as illustrated in Figure 1. An increased metabolism was also observed in the thalamus, which might reflect persistent sensory processing and information processing as well as subsequent shallower sleep. In contrast, during wakefulness, decreased metabolism was found in subcortical (thalamus, hypothalamus, and brainstem reticular formation) as well as in cortical regions (prefrontal cortex bilaterally, left superior temporal, parietal, and occipital cortices). These findings suggest that insomnia might involve abnormally high regional brain activity during sleep, associated with reduced brain metabolism during waking. The observed reduction in prefrontal cortex activity during wakefulness is consistent with reduced attentional abilities and impaired cognitive flexibility resulting from inefficient sleep and is consistent with a chronic state of sleep deprivation (Thomas, Sing et al. 2000; Drummond and Brown 2001; Durmer and Dinges 2005).

Another early study by Smith et al. (Smith, Perlis et al. 2002), which compared 5 insomniacs with 4 normal sleepers using SPECT, found no significant regional increase during NREM sleep but reduced regional cerebral blood flow (rCBF) in frontal medial, occipital, and parietal cortices, as well as in the basal ganglia during this period (see Figure 1). Interestingly, in Nofzinger’s study, decreases in activity in these same regions were also found in insomniacs, but during wakefulness. However, some methodological limitations in the Smith’s study need to be considered. Firstly, the blood flow was only sampled during the first NREM cycle. Therefore, the observed decreased metabolism in insomniacs might reflect cortical hypoarousal during the initial phases of NREM sleep following sleep onset, while it remains possible that the patients were more aroused over later NREM sleep cycles, which would be more consistent with higher beta activity later at night (Perlis, Merica et al. 2001). Secondly, the blood flow was measured after a longer duration of NREM sleep in insomnia patients than in healthy subjects, leading to a possible underestimation of activity in the patients because blood flow decreases over long NREM periods. Because of such methodological limitations, these preliminary results cannot rule out the hyperarousal hypothesis of idiopathic insomnia.

Four of the insomnia patients from the Smith’s study were rescanned after they had been treated by cognitive behavioral therapy (which included sleep restriction and stimulus control (Smith, Perlis et al. 2005). After treatment, sleep latency was reduced by at least 43%, and there was a global 24% increase in CBF with significant increases in the basal ganglia. The authors proposed that such increase in brain activity might reflect the normalization of sleep homeostatic processes.

Idiopathic Insomnia

Figure 1.Structural and functional abnormalities in insomnia. Regional cerebral metabolism during NREM sleep in idiopathic insomnia. 

Similarly, a recent fMRI study showed that 21 old patients suffering from chronic insomnia, compared to 12 matched controls, displayed a hypoactivation of the medial and inferior prefrontal cortical areas (BA9, 44-45) (Altena, Van Der Werf et al. 2008). The prefrontal abnormalities were revealed by using a category and a letter fluency task during a waking fMRI acquisition before and after a 6 weeks period of nonpharmacological sleep therapy. This therapy included cognitive behavioral therapy, body temperature and bright light interventions, sleep hygiene and physical activity counseling. There were no significant behavioral differences between groups, allowing thus the interpretation in term of differential recruitment of brain areas. Interestingly, abnormalities recovered after a nonpharmacological sleep therapy (n = 10) but not after a wait list period (n = 10).

Theses results should be refined by using larger samples of well diagnosed patients and matched controls in protocols combining structural, neuropsychological, neuroendocrine, neurochemical, functional imaging and polysomnographic studies. Hopefully, these interesting initial results will inspire further investigation on the effects of psychotherapy on brain functioning in insomnia.

Functional Imaging of Hypnotic Drugs Response in Healthy Individuals

To our knowledge, there is no neuroimaging study of hyponotic drugs response in insomniacs but only in healthy subjects. In addition, most of the studies studied the effect of acute and not chronic administration.

Functional neuroimaging allows some insights into the mechanisms of several sedative drugs, although the neuroimaging data are still sparse. Most of the studies concern the class of benzodiazepines (see Table 1). For instance, lorazepam administration markedly decreases regional brain glucose metabolism in thalamus and occipital cortex during wakefulness (Volkow, Wang et al. 1995; Schreckenberger, Lange-Asschenfeldt et al. 2004). In the former study, changes in metabolic activity in thalamus were significantly related to lorazepam-induced sleepiness and were partially reversed by flumazenil, a benzodiazepine antagonist (Volkow, Wang et al. 1995). It was suggested that benzodiazepine-induced changes in thalamic activity may account for their sedative properties. This is reinforced by a study in normal subjects that finds a close relationship between bilateral thalamic activity and alpha rhythm, generally considered to be the marker of restful wakefulness. Both glucose metabolism and alpha rhythms are reduced under lorazepam in bilateral thalamus (Schreckenberger, Lange-Asschenfeldt et al. 2004).

Another type of short-acting benzodiazepine, triazolam, is correlated with a decrease in blood flow in the basal forebrain and amygdaloid complexes during NREM sleep (Kajimura, Nishikawa et al. 2004). These results suggest that hypnotic effect of the benzodiazepines may be mediated mainly by deactivation of the forebrain control system for wakefulness and also by the anxiolytic effect induced by deactivation of the amygdaloid complexes (Kajimura, Nishikawa et al. 2004). The impairment of episodic memory encoding by this drug, on the other hand, is associated with dose-related deactivation in prefrontal cortex, medial temporal lobe and left anterior cingulate cortex (Mintzer, Kuwabara et al. 2006).

Midazolam, another drug of this class with short half-life and known to cause anterograde amnesia in healthy subjects, diminishes functional connectivity in posterior cingulate cortex (Greicius, Kiviniemi et al. 2008) in resting state analysis of fMRI data. PET studies found decrease of cerebral blood flow in prefrontal cortex, insula, temporal lobe, and associative areas comparing before and after infusion of midazolam (V eselis, Reinsel et al. 1997; Bagary, Fluck et al. 2000; Reinsel, Veselis et al. 2000).

Table 1. Functional imaging of hypnotic drugs response in healthy individuals.

Study	Imaging	State	Treatment	Number of subjects	Placebo	Paradigm /Task	Results
Volkow et al. (1995)		Wakefulness, sedation	Lorazepam	21	Yes	No task	Decrease in thalamus, occipital cortex
			Flumazenil	9	No	After lorazepam	Partial reverse of the decrease
Schreckenberger et al. (2004)		Wakefulness, sedation	Lorazepam	10	Yes	No task	Decrease in thalamus, occipital cortex, temporo-insular areas Decrease of alpha power
Kajimura et al. (2004)		Wakefulness, NREM (stage 2, 3 and 4)	Triazolam	15	Yes	No task	Decrease in basal forebrain and amygdaloid complexes
Mintzer et al. (2006)		Wakefulness, sedation	Triazolam	12	Yes	Episodic memory encoding	Decrease in right PFC, left parahippocampus, left ACC
Veselis et al. (1997)		Wakefulness, sedation	Midazolam, one or two infusions	14	No	Low vs high midazolam EEG effect No task	Decrease in insula, cingulate gyrus, PFC, thalamus, parietal and temporal
Bagary et al. (2000)		Wakefulness, sedation	Midazolam	15	Yes	7 with drug, 8 with placebo explicit memory encoding	Decrease in PFC, superior temporal, parieto-occipital at baseline Increase in left PFC during task in both groups
Reinsel et al. (2000)		Wakefulness, sedation, stage 2 sleep	Midazolam	14	No	Low vs high midazolam EEG effect No task	Decrease m left DLPFC, bilateral OFC, left middle temporal, right hippocampus
Greicius et al. (2008)		Resting state, sedation	Midazolam	9	No	No task	Decreased functional connectivity in posterior Cingulate Cortex
Gillin et al. (1996)		NREM	Zolpidem	12	Yes	No task	Decrease in cingulate gyrus, medial frontal cortex, putamen, thalamus, hippocampus
Finelli et al. (2000)		Wakefulness, stage 2, stage 4 and REM sleep	Zolpidem	8	Yes	Prior sleep deprivation No task	During sleep (all stage): decrease in basal ganglia and insula; increase in parietal cortex During REM sleep: decrease in anterior cingulate; increase in occipital, parietal, parahippocampal cortex and cerebellum
Schlaepfer et al. (1998)		Wakefulness	Opiates : hydromorphone butorphanol	9	Yes	No task	Hydromorphone: increase in ACC, both amygdala, thalamus. Butorphanol: increase in both temporal lobes.

Hydromorphone: increase in ACC, both amygdala, thalamus. Butorphanol: increase in both temporal lobes.

From top to bottom: sorted by treatment and then by date of publiscation. From left to right: reference of the study, modality of brain imaging, state in which the study is conductied and task paradigm used, number of studied subjects, placebo-controlled study, and main result of the study (PFC = prefrontal cortex; ACC = Anterior cingulated cortex; OFC = orbitofrontal cortex; DLPFC = dorsolateral prefrontal cortex).

During sleep induced by zolpidem (an imidazopyridine hypnotic relatively selective for alpha-1 subunit of the omega-1 (BZ1) receptor of the gamma-aminobutyric acid type A) in healthy subjects, rCBF decreases compared to pacebo in the anterior cingulate cortex during REM sleep while it decreases in the prefrontal cortex and the insula during NREM sleep (Finelli, Landolt et al. 2000). Another study finds metabolic decreases in the metbolism in the cingulate, the thalamus and the putamen during NREM sleep (Gillin, Buchsbaum et al. 1996). Data on the sedative effect of opiates are rare. One study aimed at comparing the general effect of hydromorphone (^-receptor agonist) and butorphanol (agonist/antagonist with k component of activity) found different pattern of activation in a SPECT study: hydromorphone compared to plabebo elicited activation of the anterior cingulate cortex, thalamus and amygdala bilaterally, while butorphanol produced a more diffuse pattern (Schlaepfer, Strain et al. 1998).

Despite these interesting results, none of these reports has studied the placebo-controlled effects of sedative agents in a large sample of well diagnosed insomniac patients compared to healthy subjects matched. It is also remarquable that while antidepressant are widely used for the treatement of insomnia, outside of depression context, at our knowledge no brain imaging studies of antidepressants use in idiopathic insomnia patients exists. The same observation applies for over the counter drugs such as melatonine or diphenhydramine. Additionaly, another striking feature is the absence, for a large part of the studies, of EEG control for the state of vigilance that could inform us if the subjects were actually sleeping or not.