Abstract
Insomnia has become fully recognized as one of the most prevalent sleep disorders in society with a profound impact on multiple aspects of daytime functioning and quality of life. Major advances in the non-pharmacological approach to insomnia include the work of Morin and colleagues on the behavioral and cognitive treatment of insomnia and the introduction of the behavioral model published by Spielman and Glovinsky (1991). Other researchers quickly followed resulting in an increasing amount of studies validating this perspective with its separate components. In the last 15 years, the nature of the conditioned arousal as one of the components in this model has been a major topic of interest. In this context, the neurocognitive model, published by Perlis and colleagues in 1997, argues for the extension of the arousal concept with a third component: cortical arousal. The latter is reflected by high frequency EEG activity during sleep, which is thought to mirror the lack of cognitive deactivation, resulting in a disruption of the normal sleep onset and maintenance processes. Some studies have shown the presence of high frequency EEG activity during the sleep onset period, NREM and REM sleep. Furthermore, beta and gamma EEG activity seem to be related to the subjective misperception of sleep, so often seen in insomnia patients. However, other studies revealed no significant differences in the sleep EEG between insomnia patients and controls. In addition to the theoretical overview, this topic includes a study exploring the different arousal components in a group of selected insomnia patients with objective findings.
17 insomnia patients diagnosed according to DSM-IV criteria and 12 healthy controls were included in our study. Next to a general assessment of hyperarousal through the use of cortisol assay and questionnaires, a wake EEG and polysomnography were performed to evaluate the presence of cortical hyperarousal both during wakefulness and sleep.
In comparison to a control group, insomnia patients experienced more cognitive and emotional arousal, but no increase in physiological arousal, both subjectively as well as objectively. Indications of cortical arousal were only present during the sleep onset period, reflected by a stable alpha EEG level and slower increase of delta power, related to longer sleep onset latencies. Furthermore, the cortical arousal variables were correlated significantly with objective sleep disruption, not with sleep perception. Together with previous studies, these results point to a large variability in insomnia patients as to the expression of hyperarousal and its different components.
Introduction
Sleep is a behavior we engage in approximately one third of our lives. Falling asleep and staying asleep is a natural phenomenon for most people. However, approximately 10 to 20% of the population report difficulties initiating or maintaining sleep, accompanied by impairment of one or more aspects of daytime functioning [1]. When facing the fact that sleep initiation is delayed or has to be repeated several times during the night, the process of falling asleep loses its obvious characteristics and becomes a conscious and often frustrating task. Most people encounter this kind of situation at least once in their lives. Often a stressful or disruptive event causes the temporary sleep difficulties, which in turn will disappear when the stressor fades away. However, some people are more sensitive to a disruption of their sleep-wake rhythm and will continue having sleep problems even when the initial cause has disappeared. Often reported characteristics of sleep initiation or reinitiating problems are the presence of ‘racing thoughts, rumination and a state of alertness’ at a time and place when relaxation is necessary [2, 3]. In reaction to the resulting sleep difficulties, behavioral coping strategies are developed, aiming at an increase of sleep time. When this process is repeated for a period of time, it becomes connected or linked to the specific environment within which it occurs; a phenomenon called conditioned arousal.
The development of this subtype of insomnia is well described and widely known as the behavioral model [4]. Sleep is considered a complex behavior, partially dependant on daytime stimuli. The focus lies on the conditioned arousal, manifesting itself on different levels, such as anxiety, muscle tension, destructive and/or obsessive cognitions about sleep, and the consequences of sleep shortage. Within this model, the so-called 3 Ps (predisposing, precipitating and perpetuating factors) describe the 3 most important factors playing a key role in the development of insomnia [5]. First of all, it is suggested that insomnia patients are characterized by predisposing factors, making them more sensitive to sleep disruptive phenomena. These trait or predisposing factors can be related to personality traits, biological and/or psychological factors. At this point, however, a significant sleep disruption should not be present. It is only with the occurrence of a precipitating factor, mostly a stressful life event, that an interference of normal sleep processes takes place. As a reaction to the delayed and/or fragmented sleep period certain behavioral strategies, such as an extended time in bed, are employed as a way of ‘catching up’ on sleep. These strategies or perpetuating factors, however, reinforce the relationship between wakefulness and the bedroom, which in turn will result in conditioned arousal, and finally in a more severe sleeping problem. At this point a negative vicious cycle is installed, maintaining the sleep disruption and its facilitating processes. Within this model it is assumed that the sleep disturbances in chronic insomnia are maintained because of these perpetuating factors, which in turn form the focus of attention for treatment interventions.
The Concept of Arousal: Validation of the Behavioral Model
Arousal mechanisms are essential for surviving and have an adaptive function enabling our most basic behaviors, such as movement, sleep, rest, wakefulness, and danger orientation. As such, arousal is part of our daily life and makes it possible to perform certain physical and/or mental tasks. However, the perception of arousal may vary according to the possible presence of emotions related to the situation [6]. Performing sports after work will often not be recognized as a situation of arousal, although all characteristics of physical arousal are present, such as muscle activation, increased heart rate and ventilation. However, when confronted with a possible dangerous situation, we are more aware of the changes inducing a state of arousal. Together with the physical preparation, our cognitive system reacts as well, by scanning the environment for dangerous cues, and as such selective attentional processes are activated. Chronic insomnia patients tend to perceive sleep-related situations, such as the bedroom and bedtime, as stressors, as such inducing an arousal response, as proposed by the behavioral model. This theoretical perspective has given rise to a growing amount of research concentrating on the assessment of different arousal components in insomnia patients. One of the major difficulties in these studies, however, is the differentiation of the different arousal components involved, because of possible similarities in the underlying etiologies [7].
Physiological Arousal
The process of falling asleep is accompanied by a series of events indicating a deactivation of several bodily systems, as such reflecting a state of physiological de-arousal [8]. During sleep onset, a decrease in muscle activation of the upper airway dilator muscles and respiratory pump muscles result in a fall in ventilation [9], which is also accompanied by a gradual decrease in heart rate [10]. Although EMG changes during sleep onset are not often studied in detail, findings regarding other related topics of sleep onset, such as passive behavioural sleep devices or changes in respiratory activity, suggest a decrease in overall EMG activity during sleep onset. In light of these findings, it is obvious that an interference of physiological deactivation can result in an impairment of sleep onset processes, a phenomenon called physiological arousal. This can be often observed in insomnia patients.
One of the first studies assessing aspects of physiological arousal [11] showed that insomnia patients were characterized by elevated rectal temperature, skin resistance, and phasic vasoconstrictions. Hyperarousal was present half an hour before bedtime, as well as during sleep. This study was the starting point for many other researchers evaluating the possible link between physiological arousal and sleep disturbances. Freedman and Sattler [12] and Freedman [13], for example, revealed that insomnia patients dominantly suffering from sleep onset problems, featured increased facial muscle activity during the sleep onset period, as well as increased beta and decreased alpha EEG activity during wakefulness, phase 1 and REM sleep. Other studies reported increased heart rate in insomnia patients during the night [14, 15], as well as in the morning when confronted with a stressful event [14]. Moreover, there appears to be a correlation between specific alterations in body temperature, related to impairments of the circadian rhythm, and different types of insomnia [16]. Sleep onset insomnia might partially be associated with a delay in temperature rhythm. As such, they try to fall asleep during their "wake maintenance zone" [17] resulting in increased sleep latencies. Early morning awakening insomniacs, on the other hand, appear to be characterized by a phase advanced temperature rhythm, causing an early circadian wake up time. Sleep maintenance insomnia apparently is not associated with a temperature rhythm impairment, but with overall nocturnal elevated body temperature [18]. Finally, it is suggested that insomnia patients suffering from a combination of sleep-onset and maintenance problems are associated with a 24-hour elevated core body temperature [16]. The mediating role of the hypothalamic-pituitary-adrenal (HPA) axis has been another main focus in arousal studies in insomnia patients. Some studies have shown that insomnia patients have increased levels of evening cortisol, which are also correlated with the amount of awakenings during the night [19-21]. Backhaus et al. [22], on the other hand, only found decreased cortisol levels in the morning, negatively correlated with the reported subjective sleep quality. Although these studies report different results, they all mention some impairment of hormones related to the HPA axis. Therefore, it can be proposed that the hypothalamic-pituitary-adrenal (HPA) axis is overactivated, mostly due to stress and anxiety, keeping the cortisol levels high en thus interfering with normal sleep onset and maintenance processes [23, 24]. Despite these impressive results, other studies failed to find significant differences in the mentioned parameters of physiological arousal. Riemann and colleagues [25] for example, did not found elevated levels of evening cortisol, but only decreased levels of melatonin. No increased heart rate was found in the study by Monroe (1967). Varkevisser and colleagues [26] performed a 24-hour sleep deprivation protocol in insomnia patients and evaluated several indicators of physiological arousal (cardiovascular parameters, cortisol, and body temperature) but found no significant differences in comparison to healthy controls. These mixed results suggest a vast heterogeneity in regard to the presence and expression of physiological arousal in insomnia patients.
Cognitive Arousal
As discussed above, falling asleep is accompanied by relaxation and deactivation of several bodily functions, which can be impaired at several levels in insomnia patients. However, a reduction in physiological processes is not the only requirement for a rapid sleep onset. The mind is a powerful and sometimes uncontrollable entity, processing all incoming information from external and internal stimuli during the day, which can also interfere with sleep onset. Increased presleep cognitive activity has been consistently associated with the maintenance of insomnia. Studies have repeatedly shown that insomnia patients experience intrusive thoughts, which are mostly negatively toned, as well as excessive worry, typically about sleep, the lack of sleep and its consequences on daytime functioning [27-29]. Studies evaluating self-reported attributions showed that insomnia patients report more presleep cognitive activity during the sleep onset period in comparison to healthy sleepers, and experience more sleep disturbances from presleep cognitive activity [30]. Furthermore, cognitive arousal appears to be more dominantly present as opposed to physiological arousal [3]. Besides the plain existence or presence of disruptive cognitive activity, it is also very interesting to evaluate its specific content in patients suffering from insomnia, especially in light of therapeutic interventions. Regarding the nature of cognitive activity, it has been reported that problem solving, worries and concerns, and listening to noises are a major focus of attention when trying to fall asleep [30]. Furthermore, thoughts about sleep shortage and re-evaluating the day are often reported. Another characteristic of insomniacs in comparison to healthy sleepers is the perceived control over presleep thoughts and worries. Whereas normal sleepers report that their presleep cognitive activity is intentionally, insomniacs describe this as being uncontrollable. Wicklow and Espie [31] conducted a study were they evaluated the content of intrusive thoughts and their relationship with actigraphic measured sleep and self report. Through the use of Principal Component Analysis, they were able to derive three major factors of intrusive thoughts. The first factor was referred to as ‘active problem solving’, which correlated with objective sleep latency and was relatively unaffected by emotional tone. The second factor was ‘present state monitoring’, reflecting self-awareness and self-monitoring. No connection was found with the objective sleep latency, but an inverse correlation was present with emotional tone. ‘Environmental reactivity’ was the third factor, but no relations were found with any sleep parameter. Regarding the subjective sleep report, none of the factors apparently correlated with perception of sleep. The role of cognition in insomnia has been elegantly described by Harvey [32] in her paper ‘the cognitive model of insomnia’. It is suggested that insomnia patients are overly worried about their sleep and the possible consequences of sleep shortage on daytime functioning, which results in ‘excessively negatively toned cognitive activity’. This in turn leads to an arousal response and emotional distress, triggering selective attention towards all kinds of stimuli that are perceived as threads for a good night sleep. The combination of arousal and distress, as well as specific selective attentional processes cause a distortion in the perception of the sleep complaints and impairments in daytime functioning. This in turn fuels again the negatively toned cognitive activity, and is the starting point for a negative vicious cycle. In line with the behavioral perspective described earlier, this negative cognitive activity gives rise to certain beliefs about sleep and their sleep problem, as well as safety behaviors, comparable with the perpetuating factors of the behavioral model. Other researchers have also demonstrated the presence of a sleep related attentional bias. Taylor et al. [33] compared two groups of cancer patients with insomnia, the first group at 0-3 months and the second at 12-18 months after cancer diagnosis. All patients had been good sleepers before the diagnosis. Results from the Stroop paradigm showed that both groups presented an attentional bias for cancer words, but only the persistent insomnia patients who still experienced sleep disturbances a year after their diagnosis, demonstrated attention bias for sleep-related words. In a recent study, Spiegelhalder and colleagues [34] for example, compared insomnia patients, sleep experts and healthy controls using an emotional Stroop task. They found significant higher attentional bias scores in the insomnia group in comparison to the sleep expert group, suggesting that the attentional bias for sleep-related words is due to specific emotional, cognitive or procedural processing rather than differences in habitual exposure to these concepts.
In summary, these results suggest that insomnia patients are characterized by some form of cognitive arousal, mainly consisting of thoughts concerning problem solving and monitoring. As is shown by Harvey’s cognitive model of insomnia, specific selective attentional processes play an important role in the maintenance of insomnia, which in turn has been repeatedly shown by studies who found an attentional bias in insomnia patients for sleep related words or cues. Espie and colleagues [35] reviewed literature about attentional bias in insomnia patients and introduced an important sleep inhibitory process, namely the attention-intention-effort pathway. They point out that the automaticity of the sleep system in insomnia patients is inhibited by three factors: first, the selective attention to sleep; secondly, the explicit intention to sleep; and finally, a dysregulation by both direct and indirect sleep effort. As such, when considering therapeutic intervention, the cognitive processes and attentional bias should be an important focus of attention.
Clinical Applications: Cognitive and Behavioral Interventions for Insomnia
The growing literature regarding physiological and cognitive arousal in insomnia patients resulted in studies evaluating different interventions aiming at a reduction of these arousal components. Since the 1970′s the impact of different relaxation techniques on physiological arousal and sleep quality were examined. Borkovec and Fowles [36] for example, evaluated three different relaxation trainings as well as a waiting list no-treatment control group. They used progressive relaxation, hypnotic relaxation and self-relaxation as a way to influence physiological arousal in insomnia patients. They hypothesized that only the progressive and hypnotic relaxation groups would demonstrate significant improvement after training. However, results showed an equal improvement in sleep onset latency, number of awakenings and waking up refreshed in all three training groups. Surprisingly, the reduction in physiological arousal reflected by skin conductance, heart rate and respiration was not related to treatment outcome. The authors suggested that the general instruction in all groups to relax and focus on pleasant internal feelings might be the mediating factor responsible for the general improvement in all groups. As such they hypothesized that attention focusing may be enough as treatment intervention for moderate insomnia patients. A similar study was performed by Nicassio and Bootzin [37] using progressive relaxation and autogenic training as active treatment groups, and a self-relaxation and waiting list group as control groups. In contradiction with Borkovec and Fowles [36] they only found significant improvement after progressive relaxation and autogenic training, suggesting that the mere instruction to relax at a scheduled time is not enough to result in significant improvements in sleep. As their sample of insomnia patients had more severe sleep problems, it was also suggested that the self-relaxation instruction may lose its value with increasing severity of the sleep complaints.
In addition to this line of research, much attention has been paid to other behavioral interventions, mostly based on operant or instrumental conditioning. The use of ‘stimulus control’ [38], for example, was one of the first interventions successfully applied in insomnia patients, which is based on the fact that spending to much time in bed is an important perpetuating factor playing a key role in the maintenance of the sleep complaints [4]. It has been shown that avoiding wakefulness in bed during the night as well as daytime napping results in a significant improvement of total sleep time, wake after sleep onset, sleep efficiency and sleep onset. The main objective of this intervention technique is to reassociate the sleep environment with relaxation and fast sleep onset, resulting in a new conditioned response as opposed to the arousal response leading to the reported sleep complaints. The American Academy of Sleep Medicine has recommended stimulus control instructions as a ‘standard’ treatment for primary insomnia [39].
A second behavioral intervention that has received much attention is the use of ‘sleep restriction therapy’ [40] were the time in bed is restricted to the reported total sleep time, as such that a mild sleep deprivation results in fast sleep onset latencies and increased sleep efficiency. In addition to these two major behavioral interventions, sleep hygiene instructions [29, 41] are recommended to ensure that poor sleep habits do not interfere with the beneficial effects of other interventions.
Finally, in the 1990′s cognitive therapy was introduced in order to directly intervene on the level of dysfunctional beliefs and attitudes about sleep [29]. These different components were then integrated into a multi-component treatment for insomnia, widely known as Cognitive Behavioral Therapy for Insomnia (CBT-I) [42]. The use of these different components in one integrated training program results in better outcome in comparison to single component treatment, and the addition of cognitive restructuring to the behavioral components causes slightly greater benefits than behavioral treatment alone [43]. Furthermore, it has been shown that CBT-I does induce a greater decrease in maladaptive beliefs and attitudes about sleep in comparison to relaxation therapy and placebo [44]. This effect is maintained at follow-up 6 months after completing the training program. Morin and colleagues [45] performed a similar study comparing a CBT-I group, pharmacotherapy, combined therapy of CBT-I and pharmaco and a placebo medication group. Again they showed that CBT-I or the combination of CBT-I with medication resulted in greater improvements of beliefs and attitudes about sleep. Furthermore, both studies showed that a decrease in maladaptive beliefs and attitudes were correlated with objective and subjective sleep improvement.
Although CBT-I is regarded the gold standard for psychological management of insomnia, there are limitations, suggesting that a further exploration of additional interventions or treatments is still required [46]. First of all, the majority of insomnia patients following CBT-I show an average improvement that does not bring them into the good sleeper range, which means that they still show some impairment after treatment [46-48]. Secondly, the effect sizes after CBT-I training are markedly lower in insomnia patients in comparison to the effect sizes resulting from CBT in other psychophysiological disorders [46]. Thirdly, although the combination of CBT-I with pharmacotherapy might result in better improvements on the short term, Hauri [49] also showed that the progress obtained after a combined therapy are not maintained over a follow-up period of 10 months in comparison to the use of CBT-I alone. Fourthly, the different components of CBT-I, such as sleep restriction and stimulus control, require a certain amount of dedication and the will to make some changes in habitual life style to produce the desired effects, which is not always obvious [48]. Finally, it has to be noted that about 19% to 26% of patients do not respond to CBT-I [46, 48]. These observations suggest that a focus on physiological and/or cognitive/behavioral components of arousal might not be enough for a substantial group of insomnia patients, and other factors may be in play causing sleep disruption.
The Neurocognitive Model: Introduction of a New Arousal Component
As a theoretical perspective on insomnia the behavioral model has been dominantly used since the 1980′s, and it has given rise to new and efficient therapeutic interventions. However, some characteristics or paradoxes observed in insomnia patients can not be fully explained by this model [50]. A first paradox refers to the phenomenon perceiving ‘sleep’ as ‘wakefulness’. There appears to be a misperception of sleep resulting in a discrepancy between polysomnographically measured sleep and the subjective report through sleep logs [51-54]. Secondly, insomnia patients tend to overestimate the time needed to fall asleep and underestimate their total sleep time [11, 52, 55]. Thirdly, when using hypnotic medication there appears to be a discrepancy between the benefits reported by the patients and the objective gains [53, 56]. Furthermore, it has been shown that the administration of benzodiazepines does not normalize sleep; in fact it decreases SWS, while insomnia patients tend to report great benefits of them. In 1997, Perlis and colleagues introduced the neurocognitive perspective, an extension of the previous discussed behavioral model, which focuses on a third arousal component, namely cortical arousal [50]. They hypothesize that the presence of high frequency EEG activity during sleep reflects a state of hyperarousal, interfering with the normal sleep onset and maintenance processes. The presence of cortical arousal makes it possible to explain some of the mentioned paradoxes as cognitive alterations may result from high frequency EEG activity. It is suggested by the authors that cortical arousal results in heightened sensory and information processing. These cognitive alterations in turn are able to clarify certain characteristics of insomnia, such as the complaint of not falling asleep, the perceived misperception between wakefulness and sleep and the overestimation of wakefulness. The past decade research concerning cortical arousal in insomnia patients has received much attention and suggests that this sleeping disorder is characterized by the presence of high frequency EEG activity during sleep onset and sleep. Nevertheless, results are still inconsistent, probably due to different methodologies and inclusion criteria. On the other hand, literature shows that the presence of physiological and/or cognitive arousal is not a uniform phenomenon as well. Indeed, the insomnia population appears to be quite heterogeneous, which might also be the case for cortical arousal.
