- Email: [email protected]
A Behavioral Perspective on Insomnia Treatment
Sl.eep Disorders
0193-953X/87 $0.00+ .20
A Behavioral Perspective on Insomnia Treatment Arthur]. Spielman, Ph.D.,* Lauren S. Caruso, M .S .,f and Paul B. Glovinsky, PhD.:/:
Growing interest in sleep disorders has underscored the heterogeneity of conditions resulting in chronic insomnia. Difficulties initiating or maintaining sleep can reflect conditioned anxiety or muscular tension, psychiatric disorder, persistent stress, or simply ill-conceived notions regarding the proper way to sleep. These causes of insomnia are clearly amenable to behavioral intervention. Other etiologies, such as respiratory disturbance during sleep, periodic movements, or tolerance to central nervous system (CNS) medications may require surgical intervention, the use of mechanical devices, or the addition or withdrawal of pharmacologic agents. In these latter cases as well, concurrent behavioral manipulations may greatly enhance therapeutic effectiveness. We will briefly review the procedures. and underlying rationales for the major behavioral treatments of insomnia. Our focus is on sleep as a complex behavior, one that is organized temporally, responsive to stimuli encountered during wakefulness, and composed of both physiologic and cognitive components. Sleep is seen as subject to the same processes of conditioning that govern waking behaviors. Its potential for modification through conditioning forms the basis for several of the interventions to be described. Less widely appreciated is the influence that conditioning exerts on pharmacotherapy. The application of behavioral principles to hypnotic drug therapy will be introduced following discussion of treatments now in current practice.
*Associate Professor and Director of the Sleep Disorders Center, Department of Psychology, The City College of the City University of New York; Adjunct Associate Professor, Sleep Disorders Program for the Middle Aged and Older Adult, Department of Geriatrics, Mount Sinai Medical Center, New York tFellow, Department of Psychology, The City College of the City University of New York, New York, New York :j:Adjunct Assistant Professor, Department of Psychology, The City College of the City University of New York. New York New Ym:k__
Psychiatric Clinics of North America- Vol. 10, No. 4, December 1987
541
-'«"
542
ARTHUR
J.
--··
SPIELMAN , LAURENS. C ARUSO AND P AUL
fitHAV1:0flAL
TREATf.1iEI~ T S
FOR
B.
GLOVINSKY
iI~SOiviNiA
Relaxation and Biofeedback Training
Progressive relaxation is foremost among those treatments that attempt to influence sleep by modifying the quality of preceding wakeful periods. Both laboratory studies and clinical experience suggest that many patients complaining of insomnia exhibit increased autonomic activity and muscular tension. 34 Relaxation techniques lower arousal to the point at which sleep mechanisms can engage; they permit rather than generate sleep. Most behavioral treatments of insomnia take advantage of the interdependence of sleep and wakefulness. The patient may be prescribed exercises, told to be more active during the daytime, or asked to remain out of bed until a relatively late hour. In each case, qualitative or quantitative changes are made to wakefulness with the expectation that these will in turn affect sleep. Several decades of sleep research have borne out this clinical suspicion. The depth of sleep in terms of its slow-wave component appears responsive both to the extent of accumulated prior wakefulness 11 and to the degree of physical exertion/temperature elevation involved in that wakefulness. 20 If wakefulness is interrupted by naps, nocturnal sleep may be less efficient. 27 Progressive relaxation aims to make the patient more aware of when muscular tension is heightened or relaxed. 21 The patient is first asked to contract muscles in a particular group: to hold the tension and clearly register the sensation. Then the muscles are relaxed; special attention is paid to the feeling of tension ebbing, to the sensation of doing nothing rather than "trying" to relax. This procedure is repeated in sequential fashion until all major muscle groups have been treated. As a result of this training, patients learn to avoid the tonic muscular tension that is an obstacle to sleep. Just as patients with longstanding pathologic daytime sleepiness are unaware of brief sleep episodes ("microsleeps") and lose their ability to discriminate levels of alertness, 10 the chronically hyperaroused insomniac may be unable to discern momentary reductions in level of arousal. Because relaxation training requires fine discrimination of behavior states, these patients are hampered in their efforts to relax. Biofeedback studies have shown that physiologic activity can be modified more easily if sensory information concerning that activity is available. For example, someone who is not able to easily recognize decreased frontalis muscle tone might nonetheless be capable of influencing the pitch of a tone that is linked to that muscle's activity. In the treatment of insomnia biofeedback has been employed to condition muscular and electroencephalographic activity. Biofeedback devices have been used to modify the 4 to 7 Hz theta production characteristic of the transition to sleep, as well as to strengthen a 12 to 14 cps "sensorimotor rhythm," which may be homologous to the spindle activity seen in nonREM stage 2 sleep. 19 Cognitive Treatments
For many persons suffering from insomnia, anxious concerns, a racing mind or obsessional thoughts are the most distressing signals that sleep will
A
':
'
l·l
.,I I
BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATMENT
543
not come easily. Worries may reflect realistic reactions to external events such as disputes, financial difficulties, or health problems. Anxiety can be triggered just as effectively by imagined or exaggerated crises. In a surprising number of instances, patients will insist that the only worries they have relate to whether or not sleep will be forthcoming. Whatever the source of anxiety, a vicious cycle is likely; concerns over not sleeping lead to a disrupted night's sleep, which, in turn, reinforces the conviction that the insomnia will persist. Cognitive strategies are helpful in allaying these concerns and in providing a degree of calm that permits sleep mechanisms to engage. In addition to the specific ways that each treatment promotes sleep, cognitivelybased strategies have in common the means to distract or desensitize the insomniac. As the principle of reciprocal inhibition postulates, one cannot be calm and anxious at the same time. 58 The task facing the clinician becomes one of preventing the patient from setting a mental agenda that inhibits sleep, thereby removing barriers to endogenously produced sleep. For example, it has been suggested that the strategies be employed to promote 15 minutes or so of mental ease as one of the key factors leading to successful sleep. 17 Autogenic training, which employs suggested sensations of heaviness and warmth in the limbs, has been directed successfully against insomnia clinically as well as in controlled studies. 4 Self-hypnosis and meditation make use of focused attention in conjunction with imaging and regulated breathing. These techniques have been employed to counter an array of anxiety disorders, although trials for insomnia are few. 60 Systematic desensitization59 is a hybrid technique involving both imagination and muscular relaxation. A hierarchy of imagined events is constructed; each level is increasingly aversive to the patient. The patient is thoroughly instructed in relaxation methods and then guided through the imagined events. The next step in the hierarchy is not attempted until preceding items can be visualized calmly. Desensitization is more effective when the hierarchy contains items that are specific to the patient. Thus instead of using 'Tm worried about not getting enough sleep" the clinical history might suggest "It's already 3:30 in the morning, I've been lying awake for nearly 2 hours and I have an algebra test first thing tomorrow." This emphasis on the specifics of a patient's experience is a hallmark of cognitive strategy, transforming what is often treated as a generic problem into a targeted domain for clinical intervention. After overcoming the myriad worries presented by waking life, cognitive treatments of insomnia must address an issue that is specific to falling asleep: namely, the anxiety that attends relinquishing control. Sleep onset usually takes place in an environment of relative sensory deprivation, with few visual, auditory, or kinesthetic cues. The racing mind of the insomniac may serve as an attempt to remain involved with the external world and its issues, rather than surrender to the hypnogogic state. The jumble of concerns, reminders, forebodings, and rehashings that take place while lying in bed may be chaotic and unproductive, yet the alternative-allowing the mind to drift into a state in which fantasy and a destructuralized egQ reign53-may be even less acceptable.
~
---
544
ARTHUR
J. SPIELMAN,
LAURENS. CARUSO AND PAUL
B. GLOVINSKY
Stimulus Control i nstructions
The experience of falling asleep repeatedly during the day in specific contexts despite adequate nocturnal sleep suggests that stimuli can become discriminative cues promoting sleep onset. 37 For example, clinicians may become sleepy with a particular psychotherapy patient. 30 Discussions of this phenomenon focus on the therapist's adaptation to particular conflicts triggered during the session. Once established, the speed and ease with which sleepiness recurs suggests that conditioning has taken place; previously neutral contextual cues associated with the patient are now soporific. Discriminative cues may also inhibit sleep onset. At home bedtime rituals and the sleeping environment may interfere with an insomniac's sleep, yet in the sleep laboratory with wires attached from head to toe this patient may experience "the best night of sleep in years." This reversal of the "first-night effect"- poor sleep usually expected on the first night in the lab---suggests that in the unique setting of the laboratory, the insomniac has been relieved of the discriminative cues that interfere with sleep. In addition to the foregoing anecdotal support, systematic studies in animals have shown that as a result of conditioning previously neutral cues can become capable of producing either rapid sleep onset or arousal. 37• 49 Persons suffering from insomnia repeatedly pair the act of going to bed with a restless, frustrating wait for sleep. Eventually, an association is formed between the various rituals attending bedtime (such as the clock reading 11:00 P.M., the end of television news, tooth brushing) and the experience of being unable to sleep. These rituals, as well as the bed itself, may become cues for arousal. Stimulus control instructions3 have been devised that aim to ensure that bedtime becomes associated with rapid sleep onset. These instructions are as follows : 1. Go to bed only when you are sleepy. 2. If you are not asleep within about 10 minutes of geting into bed, get out of bed. When you return to bed the same rule holds; if you are not sleeping within 10 minutes, get out of bed. 3. Set your alarm and get up at the same time every morning. 4. Do not nap. 5. Do not use the bed for anything except sleep (sexual activity is permitted there).
Chronotherapy
Periodic behaviors are governed by endogenous oscillators; they are not solely dependent on external time cues (zeitgebers) for their expression. If an organism is placed in an environment free of time cues, the cyclicity of the behaviors will generally persist, although with period lengths somewhat different from those observed under entrained conditions. 35 For example, the free-running sleep-wake cycle of most human subjects is close to 25 hours. Under entrained conditions, the various physiologic functions which have been synchronized via zeitgebers will assume distinct and stable phase relationships. The body temperature minima, for example, may occur 3 hours before habitual arising time. A well-entrained circadian system is characterized not only by stabil-
A
lit,
'I
i
!
BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATMENT
545
ity, but also by large amplitudes. There should be a great difference between, for example, the level of arousal experienced at 11 A.M. versus the sleepiness present at 3 A.M. The strength and stability of circadian rhythms are affected by the convergence and salience of zeitgebers, drug intake, psychiatric disorder, and by an irregular sleep-wake schedule. Wide variation in the timing of nocturnal sleep together with frequent daytime napping tend to flatten circadian rhythms, for example, by increasing the propensity for daytime sleepiness and nocturnal arousal. 51 • 55 Attenuated rhythms are most clearly seen in shift workers and also observed in the ubiquitous problem of "Sunday night insomnia." Here, later bedtimes and arising times over the weekend have begun the process of shifting circadian rhythm peaks to later clock hours. Then, on Sunday night, when the decision is made to go to bed earlier in order to get up in time for work, these delayed rhythms will anchor sleep later in the night and give rise to long sleep latencies. Of course, in this situation, the insomniac may be quick to identify work pressures as a cause of sleep disturbance, and perhaps rightfully so, while tending to overlook the contribution of circadian factors. In some individuals, difficulty initiating sleep at an appropriate clock hour and trouble arising in the morning are primary complaints, not limited to Sunday night. Sleep once begun is fairly consolidated. In these cases, consideration should be given to the possibility of a disturbance of the sleep-wake cycle known as delayed sleep phase syndrome. 56 In this condition, a stable shift of the sleep propensity phase of the circadian sleep-wake cyde towards later clock hours is primarily responsible for producing long sleep latencies; learned associations may aggravate the condition. Chronotherapeutic techniques exploit characteristics of the circadian pacemaker in order to establish a more appropriate phase relationship between sleep propensity and clock time. Most people are more capable of delaying sleep by staying awake longer than they are of advancing sleep to an earlier hour. Therefore, if sleep arrives habitually at 3 A.M. rather than the desired midnight, it is often easier to progressively delay sleep onset by a total of 21 hours (in 3-hour increments) than advance the timing of sleep by three hours. For example, the patient who generally falls asleep at 3 A.M. might be assigned a bedtime of 6 A.M. to 1:30 P.M. on the first day of treatment. The next bedtime would be between 9 A.M. and 4:30 P.M., and so on until the desired bedtime is reached. Patients should stabilize their bedtime prior to beginning chronotherapy. In establishing a schedule, care should be taken to avoid too much time in bed (which could cause sleep onset difficulties even after a 3-hour delay). It is useful to provide a calendar with explicit instructions to the patient, as in the course of living on a 27-hour "day," one whole calendar day will be lost. Close contact should be maintained with the patient during the course of treatment, as the odd bedtimes encountered during the middle phase can be quite isolating and distressing. Patients should be cautioned that although chronothe~ can effectiv~- "reset their clQcis.,~he}'._ stilLmight_ have a tendency to "run slow" (that is, drift toward later bedtimes) and
546
ARTHUR
J.
SPIELMAN, LAUREN
s.
CARUSO AND PAUL
B.
Gi.o-vINSKY
shouid guard again st this by adheri11g 1igidly to theii· nev~r schcd~llcs. Fi nally, in some cases, late schedules are reinforced by factors that are unrelated to circadian physiology, as when an adolescent gains relative independence from family members by sleeping out of phase with them . Such factors must be addressed if the new schedule is to be maintained.
Sleep Restriction Therapy Excessive time spent in bed often perpetuates insomnia, regardless of the origins of the disorder (Figure 1). 47 It is common for insomniacs to feel that extending bedtime will help compensate for long sleep latencies or disrupted sleep, because there is more opportunity to "catch up" on sleep. Although extra time in bed will, on occasion, yield more sleep, it has potentially deleterious effects. One of the assumptions of sleep restriction therapy is that extra bedtime often leads to increased wakefulness, resulting in fragmented sleep and variability in the timing of sleep and wakefulness. 48 Recent studies have shown that the continuity of sleep and the regular timing of the sleep and waking phases of the circadian cycle are major determinants of the adequacy of sleep. 2• 6 Sleep restriction therapy aims to consolidate sleep and constrain its occurrence to a specific time by restricting time spent in bed. Although initially precluding the possibility of a "great night's sleep," it reduces the night-to-night variability generally associated with insomnia. This diffuses the anticipatory anxiety attending bedtime. Patients begin to expect relatively rapid sleep onsets and at least tolerable lengths of sleep. As with stimulus control instructions, the bed and bedtime gradually shed their associations with the aversive experience of tossing and turning all night.
E] PREDISPOSING FACTORS I] PRECIPITATING FACTORS
~
i1ll
PERPETUATING FACTORS
Cf)
z
w z
1<(
z
~
0
Cf)
z
Threshold /
.I'
.I'
.I'
.I'
"
.,.
.I'
.I'
.I'
.I'
,.
.I'
.I'
"
.I'
,.
,,,
.I'
,.
.I'
,.
,.
,.
... ' ' ' ' ... '''''' '''''' ''''''
2
Pre-clinical
3
Onset
Short-term
COURSE Figure 1.
The course of insomnia.
4
Chronic
A BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATMENT
547
Sleep restriction therapy takes the patient's subjective perception of sleep quite seriously, even when polysomnographic study has revealed a large discrepancy between that perception and actual length of sleep. Initially, patients are prescribed a time in bed equal to their estimated average amount of sleep. For example, if a patient reports spending 8 hours in bed but only sleeping for 5, then the prescribed bedtime at the start of treatment should be 5 hours. In cases of extremely low subjective length of sleep, clinical judgment may allow for, say, 4 hours in bed rather than matching the patient's estimate of sleep accumulation. A restricted schedule produces moderate sleep loss over the first weeks of treatment. As a result, daytime sleepiness may be present during the initial phase of treatment and patients should be informed of this potential side effect. When subjective sleep efficiency (subjective amount of sleep/time in bed x 100 per cent) averaged over five days exceeds 90 per cent, the patient is instructed to increase time in bed by retiring 15 minutes earlier. If sleep efficiency remains below an average of 85 per cent over 5 days, then time in bed is reduced. Through a series of these small adjustments in time in bed an optimal bedtime length is reached which maintains a high sleep efficiency without unduly curtailing length of sleep. I I.
I' t ! t
I
IMPORTANCE OF PERPETUATING FACTORS AND CONDITIONING Sleep restriction therapy, stimulus control instructions and specific sleep hygiene recommendations 18 evolved from an appreciation that factors that perpetuate insomnia (Table 1) may be operating long after the precipitating factors have subsided (see Figure 1).47 The insomnia that often accompanies mourning, for example, may become entrenched as worries over decrements in performance due to sleep loss are added to the patient's distress. The anticipatory anxiety concerning sleep is now a disturber of sleep. 25 Eventually, its importance for maintaining the insomnia may overshadow the resolving grief reaction. This process is so common in clinical experience that the contribution of perpetuating factors should be carefully assessed, even when clear precipitants of an insomnia are discerned. The role of conditioning in biofeedback and stimulus control instructions is implicit. In addition to these clinical demonstrations Pavlov37 and subsequent investigators 49· 50 have systematically studied the classical conditioning of sleep. Conditioning of sleep onset in cats was accomplished using electrical stimulation of the preoptic forebrain area as the unconditioned stimulus, producing sleep as the unconditioned response within 30 seconds. After 25 to 40 trials in which a tone, the conditioned stimulus, was presented 10 seconds prior to the electrical stimulation, the tone elicited high-voltage slow waves and sleep, now a conditioned response. This study and others suggest that other unconditioned stimuli, such as hypnotic drugs, may be used to condition sleep onset. As an empirical demonstration of this approach we have conducted a preliminary study (N = 4) of the effects of repeated hypnotic administration_ on polygraphically recorded daytime naps. The hypnotic triazolam (0.125
548
ARTHUR
J. SPIELMAN,
'fahie l.
LAUREN S . CARUSO AND PAUL
B.
GLOVINSKY
Factors That Perpetuate Insvrnniu
Excessive time in bed Irregular timing of retiring and arising Unpredictability of sleep Worry over daytime deficits Multiple bouts (naps, fragmentation) of sleep Maladaptive conditioning Increased caffeine consumption Hypnotic and alcohol ingestion
mg twice a day) was paired with a 200 Hz tone and other contextual cues on four to five hourly daytime nap trials, one day per week, for 8 consecutive weeks. Adaptation and baseline (placebo) were followed by 5 days of drug and a final assessment day (placebo). The results were as follows: Mean sleep latency: baseline = 11.5 minutes, first day of drug ingestion = 6. 6 minutes, and assessment = 8.4 minutes. There was no suggestion that tolerance developed over the course of drug administrations; the sleep latencies were lower on the mean of drug days 4 and 5 compared to drug days 1 and 2, in all subjects. The change in sleep latency from baseline to assessment, a 27 per cent reduction, represents a change from the normal to the borderline pathologically sleepy range on a similar protocol called the Multiple Sleep Latency Test. 7 These preliminary data suggest that pairing contextual cues with the sleep-promoting properties of a hypnotic produces a conditioned response of rapid sleep onset. Of course, this uncontrolled pilot leaves many questions unanswered. Repeated practice falling asleep and long-term adaptation to the laboratory and other nonassociative mechanisms are alternative explanations of the observed effects. Confirmation of the present findings would suggest that the mechanisms mediating sleep latency, on and off drugs as well as at night and during daytime naps, includes a conditioned component. This would require a revision of the interpretation of the Multiple Sleep Latency Test which, at present, is conceived of as reflecting physiologic sleep tendency. 7
CONDITIONING MODEL OF TOLERANCE Hypnotic drug treatment is not traditionally discussed in reviews of the behavioral treatments of insomnia. However, new and well substantiated models have demonstrated that behavioral factors modulate the response to a wide variety of drugs, suggesting that these factors are operating in the drug treatment of insomnia, whether or not appreciated. Drug tolerance has always been understood in solely physiologic terms, that is, that systemic changes within the organism resulting from drug stimulation such as receptor-site adaptation and changes in metabolic clearance are the mechanisms responsible for reduced drug effect. The feature common to all of these physiologic theories is that the organism's response to repeated drug administration is determined by the parameters of drug exposure such as dose, duration of administration, and frequency of delivery.
A BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATMENT
I'
fj
''
I
cI
:J
11 ~I
549
Countervailing theories account for drug tolerance through varying combinations of associational and nonassociational mechanisms. 1• 40• 46• 57 Our focus here is on a classical conditioning theory, which has stimulated much empirical work. 40 We also discuss implications of these behaviorally-based theories for hypnotic drug treatment. In this view, tolerance develops as the environmental cues preceding the systemic effects of the drug become drug-predictive by repeated association. As a result of such repeated pairing, the environmental cues begin to elicit conditioned responses that are opposite in direction to the unconditioned responses generated by the drug. 40 These conditioned responses have been called compensatory to distinguish them from the traditional finding with classical conditioning that the acquired (conditioned) response resembles the unconditioned response (see reference13 for an attempt to synthesize these apparently discrepant findings into the traditional model). For example, pairing environmental cues with midazolam, a benzodiazepine that has a direct pharmacologic action of sedation, results in the association of these cues with the systemic effects of the drug. During a test trial when the cues are followed by saline, the resulting conditioned response is hyperactivity. 28 The empirical finding that a conditioned response may be antagonistic or compensatory to the direct effect of the drug is the keystone of the classical conditioning theory of tolerance. This theory postulates that as contextual cues are repeated prior to each drug exposure, the compensatory conditioned response grows in strength and opposes the direct effect of the drug. 40 Tolerance is conceived of as the net effect of combining the direct response to the drug and the compensatory conditioned response. A substantial series of recent studies has demonstrated that behavioral mechanisms play a role in tolerance. These studies have demonstrated phenomena that cannot be accounted for by the extant physiologic models. For example, it has been shown repeatedly that tolerance to a wide variety of drugs depends on the contextual cues previously associated with drug delivery. 15• 28 • 29• 36 • 40• 43 Despite equal exposure to drug, subjects demonstrated greater tolerance when tested in the environment in which the drug was previously administered. Another finding consistent with behavioral models is that after tolerance has developed, experience with drug cues while drug-free affects the retention of tolerance. After acquiring drug tolerance from identical exposure to drug, groups receive different experiences during an equivalent drug-free period. One group was repeatedly presented with a placebo and the cues previously paired with drug-which constitutes extinction trials according to a conditioning model-while the other group received no placebo, drug administration, or repeated contextual cues. At the time of reassessment of the drug effect, the group which experienced the series of extinction trials was no longer tolerant, whereas the other group remained tolerant. 28· 40• 43 Other work has shown that the size of the compensatory conditioned response, when unmasked by placebo, is a direct function of the similarity of the contextual cues present during previous drug delivery and administration of placebo. 8· 29 These findings are consistent with a conditioning model and not adequately explained by physiologic theories that attribute withdrawal effects or the abstinence syndrome solely to prior drug exposure. A variety of other effects consistent with a conditioning formulation
550
AHTHUR
J.
SPIELMAN, LAUREN
S.
C ARUSO AND PAUL
B.
GLOVINSKY
have been reported , ~. :3B , 54 including the inhibition of tolerance Ly p1iu1 experiences with the drug 14· 41· 44 • 45 and the retarded development of tolerance by partial reinforcement. 41· 42 Although the associational model of tolerance has marshaled substantial support, it does not attempt to account for tolerance occurring in the absence of drug-cue contingencies. This nonassociational tolerance is based on such parameters as drug dosage and frequency of delivery. 39· 52 Furthermore, although some studies have demonstrated a compensatory conditioned response, others have not, 13 raising the question whether such antagonistic responses are necessary substrates of drug tolerance. 1
CLINICAL IMPLICATIONS Although initially effective in inducing and maintaining sleep, the benzodiazepine hypnotics show reduced efficacy with repeated administration. 12• 33 The relevance of the classical conditioning model for the pharmacotherapy of sleep is demonstrated by conditioned aspects of tolerance to drugs with sedating properties, such as barbiturates, benzodiazepines, and ethanol. 5· 16 · 31 However, sleep was not polygraphically measured in these animal studies. The conditioning perspective may yield information on ways to maintain drug efficacy by manipulating drug administration regimens. For example, both the associational and the physiologic models of drug tolerance would predict that occasional placebo trials should be superior to chronic drug administration in maintaining drug efficacy. The association model would base this prediction on the principle that intermittent extinction trials retard acquisition, whereas the physiologic model would posit that it is the reduced drug exposure that limits tolerance. The conditioning model further predicts that intermittent placebo trials would be superior in reducing conditioned tolerance compared with the same schedule of intermittent drug holidays (in which no capsule is administered). This prediction has widespread implications because the majority of insomniacs who use drugs use them intermittently32 and have been advised to do so. 12 However, nights without capsule ingestion would less effectively retard the acquisition of conditioned tolerance, compared with placebo ingestion. This prediction derives from the conditioning principle that a trial must employ the contextual cues previously associated with the unconditioned stimulus delivery in order to be an extinction trial. Behavioral principles have a wider range of application in the evaluation and treatment of insomnia than is generally recognized. In addition to generating behavioral interventions, these principles can refine the pharmacotherapy of sleep and enhance our understanding of basic sleep processes.
SUMMARY The major behavioral treatments of insomnia- progressive relaxation, biofeedback, cognitive approaches, stimulus control instructions, chronotherapy, and sleep restriction therapy- are described. The basis of these
A
BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATMENT
551
interventions are conceptualized as issuing from the interdependence of sleep and wakefulness, the temporal organization of sleei:r-wake processes, cognitive effects on arousal, the role of perpetuating factors in chronic insomnia, and conditioning. A pilot study of the conditioning of rapid sleep onset with the aid of a hypnotic provides a preliminary demonstration of the application of conditioning to the pharmacotherapy of sleep. It is predicted that the commonly accepted view of sleep latency as solely reflecting physiological sleep tendency, will require modification to include the effects of conditioning. The current pattern of hypnotic usage, an issue of widespread concern, is subjected to a behavioral analysis based on a new model of conditioned tolerance. The intermittent administration of placebo within a hypnotic regimen is predicted to be especially beneficial in sustaining hypnotic efficacy. ACKNOWLEDGEMENT Our thanks to Dr. Saul Rothenberg for his useful comments on an earlier version of this article.
REFERENCES
!1
l. Baker TB, Tiffany ST: Morphine tolerance as habituation. Psycho! Rev 92:78-108, 1985 2. Bonnet MH: Performance and sleepiness as a function of frequency and placement of sleep disruption. Psychophysiology 23:263-271, 1986 3. Bootzin R: A stimulus control treatment for insomnia. Proc Am Psycho! Assoc l(Abstr) 395-396, 1972 4. Bootzin RR, Nicassio PM: Behavioral treatments for insomnia. In Progress in Behavior Modification. Vol 6. New York, Academic Press, 1978 5. Cappell H, Roach C, Poulos CX: Pavlovian control of cross-tolerance between pentobarbital and ethanol. Psychopharmacology 74:54-57, 1981 6. Carskadon MA, Brown ED, Dement WC: Sleep fragmentation in the elderly: Relationship to daytime sleep tendency. Neurobiol Aging 3:321-327, 1982 7. Carskadon MA, Dement WC: The multiple sleep latency test: what does it measure? Sleep 5:67-72, 1982 8. Crowell CR, Hinson RE, Siegel S: The role of conditional drug responses in tolerance to the hypothermic effects of ethanol. Psychopharmacology 73:51- 54, 1981 9. Dafters R, Hetherington M, McCartney H: Blocking and sensory preconditioning effects in morphine analgesic tolerance: Support for a Pavlovian conditioning model of drug tolerance. Q J Exp Psycho! 35:1-11, 1983 10. Dement WC, Carskadon MA, Richardson G: Excessive daytime sleepiness in the sleep apnea syndrome. In Guilleminault C, Dement WC (eds): Sleep Apnea Syndromes. New York, Alan R. Liss, 1978 11. Dement WC, Greenberg S: Changes in total amount of stage four sleep as a function of partial sleep deprivation. Electroencephalogr Clin Neurophysiol 20:523-526, 1966 12. Drugs and insomnia: The use of medications to promote sleep. Consensus conference. Office of Medical Applications of Research, National Institutes of Health. JAMA 251:2410--2414, 1984 13. Eikelboom R, Stewart J: Conditioning of drug-induced physiological responses. Psycho! Rev 89:507-528, 1982 14. Fanselow MS, German C: Explicitly unpaired delivery of morphine and the test situation: Extinction and retardation of tolerance to the suppressing effects of morphine on locomotor activity. Behav Neural Biol 35:231-241, 1982 15. File S: Development and retention of tolerance to the sedative effects of chlordiazepox-,ae:ROieOI apparatus cues. Eur J Pharmacol 81:637--043, 1982 16. Greeley J, Cappell H: Associative control of tolerance to the sedating and hypothermic effects of chlordiazepoxide. Psychopharmacology 86:487-493, 1985
_J
- -552
ARTHUR
j. SPIELMAN, LAURENS. CARUSO AND PAUL B. GLOVINSKY
17. Hauri P: The cognitive-be havioral treatment of insomnia. In Karasu B (ed): Psychiatric Treatment: Treatment of Sleep Disorders in Psychiatric Practice. State of the Art and Science circa 1987. American Psychiatric Press, in press, 1987 18. Hauri P: The Sleep Disorders. Ed 2. Kalamazoo, Michigan, Upjohn, 1982 19. Hauri P: Treating psychophysiologic insomnia with biofeedback. Arch Gen Psychiatry 38:752-758, 1981 20. Horne JA, Staff LHE: Exercise and sleep: Body heating effects. Sleep 6: 36--46, 1983 21. Jacobson E: You Can Sleep Well. New York, Whittesey, 1938 22. Kales A, Bixler EO, Kales JD, et al: Comparative effectiveness of nine hypnotic drugs: Sleep laboratory studies. J Clin Pharmacol 17:207-213, 1977 23. Kales A, Bixler EO, Scharf MB, Kales JD. Sleep laboratory studies of flurazepam: A model for evaluating hypnotic drugs. Clin Pharmacol Ther 1976, 19:576--583 24. Kales A, Bixler EO, Soldatos CR, et al: Quazepam and flurazepam: Long-term use and extended withdrawal. Clin Pharmacol Ther 32:781-788, 1982 25. Kales A, Caldwell AB, Preston TA, et al: Personality patterns in insomnia. Arch Gen Psychiatry 33:1128--1134, 1976 26. Kales A, Kales JD, Bixler EO, et al: Hypnotic efficacy of triazolam: Sleep laboratory evaluation of intermediate-term effectiveness. J Clin Pharmacol 16:399--405, 1976 27. Karacan I, Williams RL, Finley WW, et al: The effects of naps on nocturnal sleep: Influence on the need for stage 1-REM and stage 4 sleep. Biol Psychiatry 2:391-399, 1970 28. King DA, Bouton ME, Musty RE: Associative control of tolerance to the sedative effects of a short-acting benzodiazepine. Behav Neurosci 101:104--114, 1987 29. Krank MD, Hinson RE, Siegel S: Conditional hyperalgesia is elicited by environmental signals of morphine . Behav Neural Biol 32:148--157, 1981 30. McLauglin JT: The sleepy analyst. J Am Psychoanal Assoc 23:363-382, 1975 31. Melchior CL, Tabakoff B: Modification of environmentally cued tolerance to ethanol in mice. J Pharmacol Exp Ther 219:175-180, 1981 32. Mellinger GD, Balter MB, Uhlenhuth EH: Insomnia and its treatment: Prevalence and correlates. Arch Gen Psychiatry 42:225-232, 1985 33. Mitler MM, Seidel WF, Van Den Hoed J, et al: Comparative hypnotic effects of flurazepam, triazolam, and placebo: A long-term simultaneous nighttime and daytime study. J Clin Pharmacol 4;1:2-13, 1984 34. Monroe LJ: Psychogical and physiological differences between good and poor sleepers. J Abnorm Psycho! 72:255-264, 1967 35. Moore-E de MC, Czeisler CA, Richardson GS: Circadian timekeeping in health and disease, Part l. Basic properties of circadian pacemakers. N Engl J Med, 309:469--476, 1983 36. Mucha RF, Volkovskis C , Kalant H: Conditioned increases in locomotor activity produced with morphine as an unconditioned stimulus, and the relation of conditioning to acute morphine effect and tolerance. J Comp Physiol Psycho! 95:351-362, 1981 37. Pavlov IP: Lectures on Conditioned Reflexes. Chapter 32. London, Lawrence and Wishart, 1928 38. Schnur P, Bravo F, Trujillo M: Tolerance and sensitization to the biphasic effects of low doses of morphine in the hamster. Pharmacol Biochem Behav 19:435-439, 1983 39. Seamen SF: Growth of morphine tolerance: The effect of dose and interval between doses. In N Brush FR, Overmeier JB (eds): Affect, Conditioning and Cognition: Essays on the Determinants of Behavior. Hillsdale, New Jersey, Erlbaum, 1985, pp 249-262 40. Siegel S: Evidence from rats that morphine tolerance is a learned response. J Comp Physiol Psycho! 89:498--506, 1975 41. Siegel S: Morphine tolerance acquisition as an associative process. J Exp Psycho!: Animal Behav Processes 3:1-13, 1977 42. Siegel S: Tolerance to the hyperthermic effect of morphine in the rat is a learned response. J Comp Physiol Psycho! 92:1137- 1149, 1978 43. Siegel S, Hinson RE, Krank MD: Modulation of tolerance to the lethal effect of morphine by extinction. Behav Neural Biol 25:257-262, 1979 44. Siegel S, Hinson RE, Krank MD: Morphine-induced attenuation of morphine tolerance. Science 212:1533-1534, 1981 45. Siegel S, Hinson RE, Krank MD: The role of predrug signals in morphine analgesic tolerance: Support for a Pavlovian conditioning model of tolerance. J Exp Psycho!: Animal Behav Processes 4:188--196, 1978
A
BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATME NT
553
46. Solomon RL: The opponent-process theory of acquired motivation: The cost of pleasure and the benefits of pain. Am Psycho! 350:691-712, 1980 47. Spielman AJ: Assessment of insomnia. Clin Psycho! Rev 6:11-25, 1986 48. Spielman AJ, Saskin P, Thorpy MJ: Treatment of chronic insomnia by restriction of time spent in bed. Sleep 10:45-56, 1987 49. Sterman MD, Clemente CD, Wyrwicka W: Forebrain inhibitory mechanisms: Conditioning of basal forebrain induced EEG synchronization and sleep. Exp Neurol 7:404--417, 1963 50. Sterman MB, Szymusiask R, McGinty D: Quantitative analysis of basal forebrain stimulation effects: Sleep induction, sleep conditioning, and state distribution. Presented at the 5th International Congress of Sleep Research, Copenhagen, Denmark, July, 1987 51. Taub JM: Behavioral and psychophysiological correlates of irregularity in chronic sleep routines. Biol Psycho! 7:37-53, 1978 52. Tiffany ST, Baker TB: Morphine tolerance in rats: Congruence with a Pavlovian paradigm. J Comp Physiol Psycho! 95:747-762, 1981 53. Vogel G, Foulkes D, Trosman H: Ego functioning and dreaming during sleep onset. Arch Gen Psychiatry 14:238--248, 1966 54. Walter TA, Riccio DC: Overshadowing effects in the stimulus control of morphine tolerance. Behav Neurosc 97:658--662, 1983 55. Webb WB, Agnew HW: Sleep efficiency for sleep-wake cycles of varied length. Psychophysiology 12:637--641, 1975 56. Weitzman ED, Czeisler CA, Coleman RM, et al: Delayed sleep phase syndrome. Arch Gen Psychiatry 38:737-746, 1981 57. Wikler A: Dynamics of drug dependence: Implications of a conditioning theory for research and treatment. Arch Gen Psychiatry 28:611--616, 1973 58. Wolpe J: Psychotherapy by Reciprocal Inhibition. Stanford, California, Stanford University Press, 1958 59. Wolpe J, Lazarus, AA: Behavior Therapy Techniques. New York, Pergamon Press, 1966 60. Woolfolk RL, McNulty TF: Relaxation treatment of insomnia: A component analysis. J Clin Consult Psycho! 51:495-503, 1983 Sleep Disorders Center Department of Psychology The City College of The City University of New York 138th Street and Convent Avenue New York, New York 10031
0193-953X/87 $0.00+ .20
A Behavioral Perspective on Insomnia Treatment Arthur]. Spielman, Ph.D.,* Lauren S. Caruso, M .S .,f and Paul B. Glovinsky, PhD.:/:
Growing interest in sleep disorders has underscored the heterogeneity of conditions resulting in chronic insomnia. Difficulties initiating or maintaining sleep can reflect conditioned anxiety or muscular tension, psychiatric disorder, persistent stress, or simply ill-conceived notions regarding the proper way to sleep. These causes of insomnia are clearly amenable to behavioral intervention. Other etiologies, such as respiratory disturbance during sleep, periodic movements, or tolerance to central nervous system (CNS) medications may require surgical intervention, the use of mechanical devices, or the addition or withdrawal of pharmacologic agents. In these latter cases as well, concurrent behavioral manipulations may greatly enhance therapeutic effectiveness. We will briefly review the procedures. and underlying rationales for the major behavioral treatments of insomnia. Our focus is on sleep as a complex behavior, one that is organized temporally, responsive to stimuli encountered during wakefulness, and composed of both physiologic and cognitive components. Sleep is seen as subject to the same processes of conditioning that govern waking behaviors. Its potential for modification through conditioning forms the basis for several of the interventions to be described. Less widely appreciated is the influence that conditioning exerts on pharmacotherapy. The application of behavioral principles to hypnotic drug therapy will be introduced following discussion of treatments now in current practice.
*Associate Professor and Director of the Sleep Disorders Center, Department of Psychology, The City College of the City University of New York; Adjunct Associate Professor, Sleep Disorders Program for the Middle Aged and Older Adult, Department of Geriatrics, Mount Sinai Medical Center, New York tFellow, Department of Psychology, The City College of the City University of New York, New York, New York :j:Adjunct Assistant Professor, Department of Psychology, The City College of the City University of New York. New York New Ym:k__
Psychiatric Clinics of North America- Vol. 10, No. 4, December 1987
541
-'«"
542
ARTHUR
J.
--··
SPIELMAN , LAURENS. C ARUSO AND P AUL
fitHAV1:0flAL
TREATf.1iEI~ T S
FOR
B.
GLOVINSKY
iI~SOiviNiA
Relaxation and Biofeedback Training
Progressive relaxation is foremost among those treatments that attempt to influence sleep by modifying the quality of preceding wakeful periods. Both laboratory studies and clinical experience suggest that many patients complaining of insomnia exhibit increased autonomic activity and muscular tension. 34 Relaxation techniques lower arousal to the point at which sleep mechanisms can engage; they permit rather than generate sleep. Most behavioral treatments of insomnia take advantage of the interdependence of sleep and wakefulness. The patient may be prescribed exercises, told to be more active during the daytime, or asked to remain out of bed until a relatively late hour. In each case, qualitative or quantitative changes are made to wakefulness with the expectation that these will in turn affect sleep. Several decades of sleep research have borne out this clinical suspicion. The depth of sleep in terms of its slow-wave component appears responsive both to the extent of accumulated prior wakefulness 11 and to the degree of physical exertion/temperature elevation involved in that wakefulness. 20 If wakefulness is interrupted by naps, nocturnal sleep may be less efficient. 27 Progressive relaxation aims to make the patient more aware of when muscular tension is heightened or relaxed. 21 The patient is first asked to contract muscles in a particular group: to hold the tension and clearly register the sensation. Then the muscles are relaxed; special attention is paid to the feeling of tension ebbing, to the sensation of doing nothing rather than "trying" to relax. This procedure is repeated in sequential fashion until all major muscle groups have been treated. As a result of this training, patients learn to avoid the tonic muscular tension that is an obstacle to sleep. Just as patients with longstanding pathologic daytime sleepiness are unaware of brief sleep episodes ("microsleeps") and lose their ability to discriminate levels of alertness, 10 the chronically hyperaroused insomniac may be unable to discern momentary reductions in level of arousal. Because relaxation training requires fine discrimination of behavior states, these patients are hampered in their efforts to relax. Biofeedback studies have shown that physiologic activity can be modified more easily if sensory information concerning that activity is available. For example, someone who is not able to easily recognize decreased frontalis muscle tone might nonetheless be capable of influencing the pitch of a tone that is linked to that muscle's activity. In the treatment of insomnia biofeedback has been employed to condition muscular and electroencephalographic activity. Biofeedback devices have been used to modify the 4 to 7 Hz theta production characteristic of the transition to sleep, as well as to strengthen a 12 to 14 cps "sensorimotor rhythm," which may be homologous to the spindle activity seen in nonREM stage 2 sleep. 19 Cognitive Treatments
For many persons suffering from insomnia, anxious concerns, a racing mind or obsessional thoughts are the most distressing signals that sleep will
A
':
'
l·l
.,I I
BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATMENT
543
not come easily. Worries may reflect realistic reactions to external events such as disputes, financial difficulties, or health problems. Anxiety can be triggered just as effectively by imagined or exaggerated crises. In a surprising number of instances, patients will insist that the only worries they have relate to whether or not sleep will be forthcoming. Whatever the source of anxiety, a vicious cycle is likely; concerns over not sleeping lead to a disrupted night's sleep, which, in turn, reinforces the conviction that the insomnia will persist. Cognitive strategies are helpful in allaying these concerns and in providing a degree of calm that permits sleep mechanisms to engage. In addition to the specific ways that each treatment promotes sleep, cognitivelybased strategies have in common the means to distract or desensitize the insomniac. As the principle of reciprocal inhibition postulates, one cannot be calm and anxious at the same time. 58 The task facing the clinician becomes one of preventing the patient from setting a mental agenda that inhibits sleep, thereby removing barriers to endogenously produced sleep. For example, it has been suggested that the strategies be employed to promote 15 minutes or so of mental ease as one of the key factors leading to successful sleep. 17 Autogenic training, which employs suggested sensations of heaviness and warmth in the limbs, has been directed successfully against insomnia clinically as well as in controlled studies. 4 Self-hypnosis and meditation make use of focused attention in conjunction with imaging and regulated breathing. These techniques have been employed to counter an array of anxiety disorders, although trials for insomnia are few. 60 Systematic desensitization59 is a hybrid technique involving both imagination and muscular relaxation. A hierarchy of imagined events is constructed; each level is increasingly aversive to the patient. The patient is thoroughly instructed in relaxation methods and then guided through the imagined events. The next step in the hierarchy is not attempted until preceding items can be visualized calmly. Desensitization is more effective when the hierarchy contains items that are specific to the patient. Thus instead of using 'Tm worried about not getting enough sleep" the clinical history might suggest "It's already 3:30 in the morning, I've been lying awake for nearly 2 hours and I have an algebra test first thing tomorrow." This emphasis on the specifics of a patient's experience is a hallmark of cognitive strategy, transforming what is often treated as a generic problem into a targeted domain for clinical intervention. After overcoming the myriad worries presented by waking life, cognitive treatments of insomnia must address an issue that is specific to falling asleep: namely, the anxiety that attends relinquishing control. Sleep onset usually takes place in an environment of relative sensory deprivation, with few visual, auditory, or kinesthetic cues. The racing mind of the insomniac may serve as an attempt to remain involved with the external world and its issues, rather than surrender to the hypnogogic state. The jumble of concerns, reminders, forebodings, and rehashings that take place while lying in bed may be chaotic and unproductive, yet the alternative-allowing the mind to drift into a state in which fantasy and a destructuralized egQ reign53-may be even less acceptable.
~
---
544
ARTHUR
J. SPIELMAN,
LAURENS. CARUSO AND PAUL
B. GLOVINSKY
Stimulus Control i nstructions
The experience of falling asleep repeatedly during the day in specific contexts despite adequate nocturnal sleep suggests that stimuli can become discriminative cues promoting sleep onset. 37 For example, clinicians may become sleepy with a particular psychotherapy patient. 30 Discussions of this phenomenon focus on the therapist's adaptation to particular conflicts triggered during the session. Once established, the speed and ease with which sleepiness recurs suggests that conditioning has taken place; previously neutral contextual cues associated with the patient are now soporific. Discriminative cues may also inhibit sleep onset. At home bedtime rituals and the sleeping environment may interfere with an insomniac's sleep, yet in the sleep laboratory with wires attached from head to toe this patient may experience "the best night of sleep in years." This reversal of the "first-night effect"- poor sleep usually expected on the first night in the lab---suggests that in the unique setting of the laboratory, the insomniac has been relieved of the discriminative cues that interfere with sleep. In addition to the foregoing anecdotal support, systematic studies in animals have shown that as a result of conditioning previously neutral cues can become capable of producing either rapid sleep onset or arousal. 37• 49 Persons suffering from insomnia repeatedly pair the act of going to bed with a restless, frustrating wait for sleep. Eventually, an association is formed between the various rituals attending bedtime (such as the clock reading 11:00 P.M., the end of television news, tooth brushing) and the experience of being unable to sleep. These rituals, as well as the bed itself, may become cues for arousal. Stimulus control instructions3 have been devised that aim to ensure that bedtime becomes associated with rapid sleep onset. These instructions are as follows : 1. Go to bed only when you are sleepy. 2. If you are not asleep within about 10 minutes of geting into bed, get out of bed. When you return to bed the same rule holds; if you are not sleeping within 10 minutes, get out of bed. 3. Set your alarm and get up at the same time every morning. 4. Do not nap. 5. Do not use the bed for anything except sleep (sexual activity is permitted there).
Chronotherapy
Periodic behaviors are governed by endogenous oscillators; they are not solely dependent on external time cues (zeitgebers) for their expression. If an organism is placed in an environment free of time cues, the cyclicity of the behaviors will generally persist, although with period lengths somewhat different from those observed under entrained conditions. 35 For example, the free-running sleep-wake cycle of most human subjects is close to 25 hours. Under entrained conditions, the various physiologic functions which have been synchronized via zeitgebers will assume distinct and stable phase relationships. The body temperature minima, for example, may occur 3 hours before habitual arising time. A well-entrained circadian system is characterized not only by stabil-
A
lit,
'I
i
!
BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATMENT
545
ity, but also by large amplitudes. There should be a great difference between, for example, the level of arousal experienced at 11 A.M. versus the sleepiness present at 3 A.M. The strength and stability of circadian rhythms are affected by the convergence and salience of zeitgebers, drug intake, psychiatric disorder, and by an irregular sleep-wake schedule. Wide variation in the timing of nocturnal sleep together with frequent daytime napping tend to flatten circadian rhythms, for example, by increasing the propensity for daytime sleepiness and nocturnal arousal. 51 • 55 Attenuated rhythms are most clearly seen in shift workers and also observed in the ubiquitous problem of "Sunday night insomnia." Here, later bedtimes and arising times over the weekend have begun the process of shifting circadian rhythm peaks to later clock hours. Then, on Sunday night, when the decision is made to go to bed earlier in order to get up in time for work, these delayed rhythms will anchor sleep later in the night and give rise to long sleep latencies. Of course, in this situation, the insomniac may be quick to identify work pressures as a cause of sleep disturbance, and perhaps rightfully so, while tending to overlook the contribution of circadian factors. In some individuals, difficulty initiating sleep at an appropriate clock hour and trouble arising in the morning are primary complaints, not limited to Sunday night. Sleep once begun is fairly consolidated. In these cases, consideration should be given to the possibility of a disturbance of the sleep-wake cycle known as delayed sleep phase syndrome. 56 In this condition, a stable shift of the sleep propensity phase of the circadian sleep-wake cyde towards later clock hours is primarily responsible for producing long sleep latencies; learned associations may aggravate the condition. Chronotherapeutic techniques exploit characteristics of the circadian pacemaker in order to establish a more appropriate phase relationship between sleep propensity and clock time. Most people are more capable of delaying sleep by staying awake longer than they are of advancing sleep to an earlier hour. Therefore, if sleep arrives habitually at 3 A.M. rather than the desired midnight, it is often easier to progressively delay sleep onset by a total of 21 hours (in 3-hour increments) than advance the timing of sleep by three hours. For example, the patient who generally falls asleep at 3 A.M. might be assigned a bedtime of 6 A.M. to 1:30 P.M. on the first day of treatment. The next bedtime would be between 9 A.M. and 4:30 P.M., and so on until the desired bedtime is reached. Patients should stabilize their bedtime prior to beginning chronotherapy. In establishing a schedule, care should be taken to avoid too much time in bed (which could cause sleep onset difficulties even after a 3-hour delay). It is useful to provide a calendar with explicit instructions to the patient, as in the course of living on a 27-hour "day," one whole calendar day will be lost. Close contact should be maintained with the patient during the course of treatment, as the odd bedtimes encountered during the middle phase can be quite isolating and distressing. Patients should be cautioned that although chronothe~ can effectiv~- "reset their clQcis.,~he}'._ stilLmight_ have a tendency to "run slow" (that is, drift toward later bedtimes) and
546
ARTHUR
J.
SPIELMAN, LAUREN
s.
CARUSO AND PAUL
B.
Gi.o-vINSKY
shouid guard again st this by adheri11g 1igidly to theii· nev~r schcd~llcs. Fi nally, in some cases, late schedules are reinforced by factors that are unrelated to circadian physiology, as when an adolescent gains relative independence from family members by sleeping out of phase with them . Such factors must be addressed if the new schedule is to be maintained.
Sleep Restriction Therapy Excessive time spent in bed often perpetuates insomnia, regardless of the origins of the disorder (Figure 1). 47 It is common for insomniacs to feel that extending bedtime will help compensate for long sleep latencies or disrupted sleep, because there is more opportunity to "catch up" on sleep. Although extra time in bed will, on occasion, yield more sleep, it has potentially deleterious effects. One of the assumptions of sleep restriction therapy is that extra bedtime often leads to increased wakefulness, resulting in fragmented sleep and variability in the timing of sleep and wakefulness. 48 Recent studies have shown that the continuity of sleep and the regular timing of the sleep and waking phases of the circadian cycle are major determinants of the adequacy of sleep. 2• 6 Sleep restriction therapy aims to consolidate sleep and constrain its occurrence to a specific time by restricting time spent in bed. Although initially precluding the possibility of a "great night's sleep," it reduces the night-to-night variability generally associated with insomnia. This diffuses the anticipatory anxiety attending bedtime. Patients begin to expect relatively rapid sleep onsets and at least tolerable lengths of sleep. As with stimulus control instructions, the bed and bedtime gradually shed their associations with the aversive experience of tossing and turning all night.
E] PREDISPOSING FACTORS I] PRECIPITATING FACTORS
~
i1ll
PERPETUATING FACTORS
Cf)
z
w z
1<(
z
~
0
Cf)
z
Threshold /
.I'
.I'
.I'
.I'
"
.,.
.I'
.I'
.I'
.I'
,.
.I'
.I'
"
.I'
,.
,,,
.I'
,.
.I'
,.
,.
,.
... ' ' ' ' ... '''''' '''''' ''''''
2
Pre-clinical
3
Onset
Short-term
COURSE Figure 1.
The course of insomnia.
4
Chronic
A BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATMENT
547
Sleep restriction therapy takes the patient's subjective perception of sleep quite seriously, even when polysomnographic study has revealed a large discrepancy between that perception and actual length of sleep. Initially, patients are prescribed a time in bed equal to their estimated average amount of sleep. For example, if a patient reports spending 8 hours in bed but only sleeping for 5, then the prescribed bedtime at the start of treatment should be 5 hours. In cases of extremely low subjective length of sleep, clinical judgment may allow for, say, 4 hours in bed rather than matching the patient's estimate of sleep accumulation. A restricted schedule produces moderate sleep loss over the first weeks of treatment. As a result, daytime sleepiness may be present during the initial phase of treatment and patients should be informed of this potential side effect. When subjective sleep efficiency (subjective amount of sleep/time in bed x 100 per cent) averaged over five days exceeds 90 per cent, the patient is instructed to increase time in bed by retiring 15 minutes earlier. If sleep efficiency remains below an average of 85 per cent over 5 days, then time in bed is reduced. Through a series of these small adjustments in time in bed an optimal bedtime length is reached which maintains a high sleep efficiency without unduly curtailing length of sleep. I I.
I' t ! t
I
IMPORTANCE OF PERPETUATING FACTORS AND CONDITIONING Sleep restriction therapy, stimulus control instructions and specific sleep hygiene recommendations 18 evolved from an appreciation that factors that perpetuate insomnia (Table 1) may be operating long after the precipitating factors have subsided (see Figure 1).47 The insomnia that often accompanies mourning, for example, may become entrenched as worries over decrements in performance due to sleep loss are added to the patient's distress. The anticipatory anxiety concerning sleep is now a disturber of sleep. 25 Eventually, its importance for maintaining the insomnia may overshadow the resolving grief reaction. This process is so common in clinical experience that the contribution of perpetuating factors should be carefully assessed, even when clear precipitants of an insomnia are discerned. The role of conditioning in biofeedback and stimulus control instructions is implicit. In addition to these clinical demonstrations Pavlov37 and subsequent investigators 49· 50 have systematically studied the classical conditioning of sleep. Conditioning of sleep onset in cats was accomplished using electrical stimulation of the preoptic forebrain area as the unconditioned stimulus, producing sleep as the unconditioned response within 30 seconds. After 25 to 40 trials in which a tone, the conditioned stimulus, was presented 10 seconds prior to the electrical stimulation, the tone elicited high-voltage slow waves and sleep, now a conditioned response. This study and others suggest that other unconditioned stimuli, such as hypnotic drugs, may be used to condition sleep onset. As an empirical demonstration of this approach we have conducted a preliminary study (N = 4) of the effects of repeated hypnotic administration_ on polygraphically recorded daytime naps. The hypnotic triazolam (0.125
548
ARTHUR
J. SPIELMAN,
'fahie l.
LAUREN S . CARUSO AND PAUL
B.
GLOVINSKY
Factors That Perpetuate Insvrnniu
Excessive time in bed Irregular timing of retiring and arising Unpredictability of sleep Worry over daytime deficits Multiple bouts (naps, fragmentation) of sleep Maladaptive conditioning Increased caffeine consumption Hypnotic and alcohol ingestion
mg twice a day) was paired with a 200 Hz tone and other contextual cues on four to five hourly daytime nap trials, one day per week, for 8 consecutive weeks. Adaptation and baseline (placebo) were followed by 5 days of drug and a final assessment day (placebo). The results were as follows: Mean sleep latency: baseline = 11.5 minutes, first day of drug ingestion = 6. 6 minutes, and assessment = 8.4 minutes. There was no suggestion that tolerance developed over the course of drug administrations; the sleep latencies were lower on the mean of drug days 4 and 5 compared to drug days 1 and 2, in all subjects. The change in sleep latency from baseline to assessment, a 27 per cent reduction, represents a change from the normal to the borderline pathologically sleepy range on a similar protocol called the Multiple Sleep Latency Test. 7 These preliminary data suggest that pairing contextual cues with the sleep-promoting properties of a hypnotic produces a conditioned response of rapid sleep onset. Of course, this uncontrolled pilot leaves many questions unanswered. Repeated practice falling asleep and long-term adaptation to the laboratory and other nonassociative mechanisms are alternative explanations of the observed effects. Confirmation of the present findings would suggest that the mechanisms mediating sleep latency, on and off drugs as well as at night and during daytime naps, includes a conditioned component. This would require a revision of the interpretation of the Multiple Sleep Latency Test which, at present, is conceived of as reflecting physiologic sleep tendency. 7
CONDITIONING MODEL OF TOLERANCE Hypnotic drug treatment is not traditionally discussed in reviews of the behavioral treatments of insomnia. However, new and well substantiated models have demonstrated that behavioral factors modulate the response to a wide variety of drugs, suggesting that these factors are operating in the drug treatment of insomnia, whether or not appreciated. Drug tolerance has always been understood in solely physiologic terms, that is, that systemic changes within the organism resulting from drug stimulation such as receptor-site adaptation and changes in metabolic clearance are the mechanisms responsible for reduced drug effect. The feature common to all of these physiologic theories is that the organism's response to repeated drug administration is determined by the parameters of drug exposure such as dose, duration of administration, and frequency of delivery.
A BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATMENT
I'
fj
''
I
cI
:J
11 ~I
549
Countervailing theories account for drug tolerance through varying combinations of associational and nonassociational mechanisms. 1• 40• 46• 57 Our focus here is on a classical conditioning theory, which has stimulated much empirical work. 40 We also discuss implications of these behaviorally-based theories for hypnotic drug treatment. In this view, tolerance develops as the environmental cues preceding the systemic effects of the drug become drug-predictive by repeated association. As a result of such repeated pairing, the environmental cues begin to elicit conditioned responses that are opposite in direction to the unconditioned responses generated by the drug. 40 These conditioned responses have been called compensatory to distinguish them from the traditional finding with classical conditioning that the acquired (conditioned) response resembles the unconditioned response (see reference13 for an attempt to synthesize these apparently discrepant findings into the traditional model). For example, pairing environmental cues with midazolam, a benzodiazepine that has a direct pharmacologic action of sedation, results in the association of these cues with the systemic effects of the drug. During a test trial when the cues are followed by saline, the resulting conditioned response is hyperactivity. 28 The empirical finding that a conditioned response may be antagonistic or compensatory to the direct effect of the drug is the keystone of the classical conditioning theory of tolerance. This theory postulates that as contextual cues are repeated prior to each drug exposure, the compensatory conditioned response grows in strength and opposes the direct effect of the drug. 40 Tolerance is conceived of as the net effect of combining the direct response to the drug and the compensatory conditioned response. A substantial series of recent studies has demonstrated that behavioral mechanisms play a role in tolerance. These studies have demonstrated phenomena that cannot be accounted for by the extant physiologic models. For example, it has been shown repeatedly that tolerance to a wide variety of drugs depends on the contextual cues previously associated with drug delivery. 15• 28 • 29• 36 • 40• 43 Despite equal exposure to drug, subjects demonstrated greater tolerance when tested in the environment in which the drug was previously administered. Another finding consistent with behavioral models is that after tolerance has developed, experience with drug cues while drug-free affects the retention of tolerance. After acquiring drug tolerance from identical exposure to drug, groups receive different experiences during an equivalent drug-free period. One group was repeatedly presented with a placebo and the cues previously paired with drug-which constitutes extinction trials according to a conditioning model-while the other group received no placebo, drug administration, or repeated contextual cues. At the time of reassessment of the drug effect, the group which experienced the series of extinction trials was no longer tolerant, whereas the other group remained tolerant. 28· 40• 43 Other work has shown that the size of the compensatory conditioned response, when unmasked by placebo, is a direct function of the similarity of the contextual cues present during previous drug delivery and administration of placebo. 8· 29 These findings are consistent with a conditioning model and not adequately explained by physiologic theories that attribute withdrawal effects or the abstinence syndrome solely to prior drug exposure. A variety of other effects consistent with a conditioning formulation
550
AHTHUR
J.
SPIELMAN, LAUREN
S.
C ARUSO AND PAUL
B.
GLOVINSKY
have been reported , ~. :3B , 54 including the inhibition of tolerance Ly p1iu1 experiences with the drug 14· 41· 44 • 45 and the retarded development of tolerance by partial reinforcement. 41· 42 Although the associational model of tolerance has marshaled substantial support, it does not attempt to account for tolerance occurring in the absence of drug-cue contingencies. This nonassociational tolerance is based on such parameters as drug dosage and frequency of delivery. 39· 52 Furthermore, although some studies have demonstrated a compensatory conditioned response, others have not, 13 raising the question whether such antagonistic responses are necessary substrates of drug tolerance. 1
CLINICAL IMPLICATIONS Although initially effective in inducing and maintaining sleep, the benzodiazepine hypnotics show reduced efficacy with repeated administration. 12• 33 The relevance of the classical conditioning model for the pharmacotherapy of sleep is demonstrated by conditioned aspects of tolerance to drugs with sedating properties, such as barbiturates, benzodiazepines, and ethanol. 5· 16 · 31 However, sleep was not polygraphically measured in these animal studies. The conditioning perspective may yield information on ways to maintain drug efficacy by manipulating drug administration regimens. For example, both the associational and the physiologic models of drug tolerance would predict that occasional placebo trials should be superior to chronic drug administration in maintaining drug efficacy. The association model would base this prediction on the principle that intermittent extinction trials retard acquisition, whereas the physiologic model would posit that it is the reduced drug exposure that limits tolerance. The conditioning model further predicts that intermittent placebo trials would be superior in reducing conditioned tolerance compared with the same schedule of intermittent drug holidays (in which no capsule is administered). This prediction has widespread implications because the majority of insomniacs who use drugs use them intermittently32 and have been advised to do so. 12 However, nights without capsule ingestion would less effectively retard the acquisition of conditioned tolerance, compared with placebo ingestion. This prediction derives from the conditioning principle that a trial must employ the contextual cues previously associated with the unconditioned stimulus delivery in order to be an extinction trial. Behavioral principles have a wider range of application in the evaluation and treatment of insomnia than is generally recognized. In addition to generating behavioral interventions, these principles can refine the pharmacotherapy of sleep and enhance our understanding of basic sleep processes.
SUMMARY The major behavioral treatments of insomnia- progressive relaxation, biofeedback, cognitive approaches, stimulus control instructions, chronotherapy, and sleep restriction therapy- are described. The basis of these
A
BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATMENT
551
interventions are conceptualized as issuing from the interdependence of sleep and wakefulness, the temporal organization of sleei:r-wake processes, cognitive effects on arousal, the role of perpetuating factors in chronic insomnia, and conditioning. A pilot study of the conditioning of rapid sleep onset with the aid of a hypnotic provides a preliminary demonstration of the application of conditioning to the pharmacotherapy of sleep. It is predicted that the commonly accepted view of sleep latency as solely reflecting physiological sleep tendency, will require modification to include the effects of conditioning. The current pattern of hypnotic usage, an issue of widespread concern, is subjected to a behavioral analysis based on a new model of conditioned tolerance. The intermittent administration of placebo within a hypnotic regimen is predicted to be especially beneficial in sustaining hypnotic efficacy. ACKNOWLEDGEMENT Our thanks to Dr. Saul Rothenberg for his useful comments on an earlier version of this article.
REFERENCES
!1
l. Baker TB, Tiffany ST: Morphine tolerance as habituation. Psycho! Rev 92:78-108, 1985 2. Bonnet MH: Performance and sleepiness as a function of frequency and placement of sleep disruption. Psychophysiology 23:263-271, 1986 3. Bootzin R: A stimulus control treatment for insomnia. Proc Am Psycho! Assoc l(Abstr) 395-396, 1972 4. Bootzin RR, Nicassio PM: Behavioral treatments for insomnia. In Progress in Behavior Modification. Vol 6. New York, Academic Press, 1978 5. Cappell H, Roach C, Poulos CX: Pavlovian control of cross-tolerance between pentobarbital and ethanol. Psychopharmacology 74:54-57, 1981 6. Carskadon MA, Brown ED, Dement WC: Sleep fragmentation in the elderly: Relationship to daytime sleep tendency. Neurobiol Aging 3:321-327, 1982 7. Carskadon MA, Dement WC: The multiple sleep latency test: what does it measure? Sleep 5:67-72, 1982 8. Crowell CR, Hinson RE, Siegel S: The role of conditional drug responses in tolerance to the hypothermic effects of ethanol. Psychopharmacology 73:51- 54, 1981 9. Dafters R, Hetherington M, McCartney H: Blocking and sensory preconditioning effects in morphine analgesic tolerance: Support for a Pavlovian conditioning model of drug tolerance. Q J Exp Psycho! 35:1-11, 1983 10. Dement WC, Carskadon MA, Richardson G: Excessive daytime sleepiness in the sleep apnea syndrome. In Guilleminault C, Dement WC (eds): Sleep Apnea Syndromes. New York, Alan R. Liss, 1978 11. Dement WC, Greenberg S: Changes in total amount of stage four sleep as a function of partial sleep deprivation. Electroencephalogr Clin Neurophysiol 20:523-526, 1966 12. Drugs and insomnia: The use of medications to promote sleep. Consensus conference. Office of Medical Applications of Research, National Institutes of Health. JAMA 251:2410--2414, 1984 13. Eikelboom R, Stewart J: Conditioning of drug-induced physiological responses. Psycho! Rev 89:507-528, 1982 14. Fanselow MS, German C: Explicitly unpaired delivery of morphine and the test situation: Extinction and retardation of tolerance to the suppressing effects of morphine on locomotor activity. Behav Neural Biol 35:231-241, 1982 15. File S: Development and retention of tolerance to the sedative effects of chlordiazepox-,ae:ROieOI apparatus cues. Eur J Pharmacol 81:637--043, 1982 16. Greeley J, Cappell H: Associative control of tolerance to the sedating and hypothermic effects of chlordiazepoxide. Psychopharmacology 86:487-493, 1985
_J
- -552
ARTHUR
j. SPIELMAN, LAURENS. CARUSO AND PAUL B. GLOVINSKY
17. Hauri P: The cognitive-be havioral treatment of insomnia. In Karasu B (ed): Psychiatric Treatment: Treatment of Sleep Disorders in Psychiatric Practice. State of the Art and Science circa 1987. American Psychiatric Press, in press, 1987 18. Hauri P: The Sleep Disorders. Ed 2. Kalamazoo, Michigan, Upjohn, 1982 19. Hauri P: Treating psychophysiologic insomnia with biofeedback. Arch Gen Psychiatry 38:752-758, 1981 20. Horne JA, Staff LHE: Exercise and sleep: Body heating effects. Sleep 6: 36--46, 1983 21. Jacobson E: You Can Sleep Well. New York, Whittesey, 1938 22. Kales A, Bixler EO, Kales JD, et al: Comparative effectiveness of nine hypnotic drugs: Sleep laboratory studies. J Clin Pharmacol 17:207-213, 1977 23. Kales A, Bixler EO, Scharf MB, Kales JD. Sleep laboratory studies of flurazepam: A model for evaluating hypnotic drugs. Clin Pharmacol Ther 1976, 19:576--583 24. Kales A, Bixler EO, Soldatos CR, et al: Quazepam and flurazepam: Long-term use and extended withdrawal. Clin Pharmacol Ther 32:781-788, 1982 25. Kales A, Caldwell AB, Preston TA, et al: Personality patterns in insomnia. Arch Gen Psychiatry 33:1128--1134, 1976 26. Kales A, Kales JD, Bixler EO, et al: Hypnotic efficacy of triazolam: Sleep laboratory evaluation of intermediate-term effectiveness. J Clin Pharmacol 16:399--405, 1976 27. Karacan I, Williams RL, Finley WW, et al: The effects of naps on nocturnal sleep: Influence on the need for stage 1-REM and stage 4 sleep. Biol Psychiatry 2:391-399, 1970 28. King DA, Bouton ME, Musty RE: Associative control of tolerance to the sedative effects of a short-acting benzodiazepine. Behav Neurosci 101:104--114, 1987 29. Krank MD, Hinson RE, Siegel S: Conditional hyperalgesia is elicited by environmental signals of morphine . Behav Neural Biol 32:148--157, 1981 30. McLauglin JT: The sleepy analyst. J Am Psychoanal Assoc 23:363-382, 1975 31. Melchior CL, Tabakoff B: Modification of environmentally cued tolerance to ethanol in mice. J Pharmacol Exp Ther 219:175-180, 1981 32. Mellinger GD, Balter MB, Uhlenhuth EH: Insomnia and its treatment: Prevalence and correlates. Arch Gen Psychiatry 42:225-232, 1985 33. Mitler MM, Seidel WF, Van Den Hoed J, et al: Comparative hypnotic effects of flurazepam, triazolam, and placebo: A long-term simultaneous nighttime and daytime study. J Clin Pharmacol 4;1:2-13, 1984 34. Monroe LJ: Psychogical and physiological differences between good and poor sleepers. J Abnorm Psycho! 72:255-264, 1967 35. Moore-E de MC, Czeisler CA, Richardson GS: Circadian timekeeping in health and disease, Part l. Basic properties of circadian pacemakers. N Engl J Med, 309:469--476, 1983 36. Mucha RF, Volkovskis C , Kalant H: Conditioned increases in locomotor activity produced with morphine as an unconditioned stimulus, and the relation of conditioning to acute morphine effect and tolerance. J Comp Physiol Psycho! 95:351-362, 1981 37. Pavlov IP: Lectures on Conditioned Reflexes. Chapter 32. London, Lawrence and Wishart, 1928 38. Schnur P, Bravo F, Trujillo M: Tolerance and sensitization to the biphasic effects of low doses of morphine in the hamster. Pharmacol Biochem Behav 19:435-439, 1983 39. Seamen SF: Growth of morphine tolerance: The effect of dose and interval between doses. In N Brush FR, Overmeier JB (eds): Affect, Conditioning and Cognition: Essays on the Determinants of Behavior. Hillsdale, New Jersey, Erlbaum, 1985, pp 249-262 40. Siegel S: Evidence from rats that morphine tolerance is a learned response. J Comp Physiol Psycho! 89:498--506, 1975 41. Siegel S: Morphine tolerance acquisition as an associative process. J Exp Psycho!: Animal Behav Processes 3:1-13, 1977 42. Siegel S: Tolerance to the hyperthermic effect of morphine in the rat is a learned response. J Comp Physiol Psycho! 92:1137- 1149, 1978 43. Siegel S, Hinson RE, Krank MD: Modulation of tolerance to the lethal effect of morphine by extinction. Behav Neural Biol 25:257-262, 1979 44. Siegel S, Hinson RE, Krank MD: Morphine-induced attenuation of morphine tolerance. Science 212:1533-1534, 1981 45. Siegel S, Hinson RE, Krank MD: The role of predrug signals in morphine analgesic tolerance: Support for a Pavlovian conditioning model of tolerance. J Exp Psycho!: Animal Behav Processes 4:188--196, 1978
A
BEHAVIORAL PERSPECTIVE ON INSOMNIA TREATME NT
553
46. Solomon RL: The opponent-process theory of acquired motivation: The cost of pleasure and the benefits of pain. Am Psycho! 350:691-712, 1980 47. Spielman AJ: Assessment of insomnia. Clin Psycho! Rev 6:11-25, 1986 48. Spielman AJ, Saskin P, Thorpy MJ: Treatment of chronic insomnia by restriction of time spent in bed. Sleep 10:45-56, 1987 49. Sterman MD, Clemente CD, Wyrwicka W: Forebrain inhibitory mechanisms: Conditioning of basal forebrain induced EEG synchronization and sleep. Exp Neurol 7:404--417, 1963 50. Sterman MB, Szymusiask R, McGinty D: Quantitative analysis of basal forebrain stimulation effects: Sleep induction, sleep conditioning, and state distribution. Presented at the 5th International Congress of Sleep Research, Copenhagen, Denmark, July, 1987 51. Taub JM: Behavioral and psychophysiological correlates of irregularity in chronic sleep routines. Biol Psycho! 7:37-53, 1978 52. Tiffany ST, Baker TB: Morphine tolerance in rats: Congruence with a Pavlovian paradigm. J Comp Physiol Psycho! 95:747-762, 1981 53. Vogel G, Foulkes D, Trosman H: Ego functioning and dreaming during sleep onset. Arch Gen Psychiatry 14:238--248, 1966 54. Walter TA, Riccio DC: Overshadowing effects in the stimulus control of morphine tolerance. Behav Neurosc 97:658--662, 1983 55. Webb WB, Agnew HW: Sleep efficiency for sleep-wake cycles of varied length. Psychophysiology 12:637--641, 1975 56. Weitzman ED, Czeisler CA, Coleman RM, et al: Delayed sleep phase syndrome. Arch Gen Psychiatry 38:737-746, 1981 57. Wikler A: Dynamics of drug dependence: Implications of a conditioning theory for research and treatment. Arch Gen Psychiatry 28:611--616, 1973 58. Wolpe J: Psychotherapy by Reciprocal Inhibition. Stanford, California, Stanford University Press, 1958 59. Wolpe J, Lazarus, AA: Behavior Therapy Techniques. New York, Pergamon Press, 1966 60. Woolfolk RL, McNulty TF: Relaxation treatment of insomnia: A component analysis. J Clin Consult Psycho! 51:495-503, 1983 Sleep Disorders Center Department of Psychology The City College of The City University of New York 138th Street and Convent Avenue New York, New York 10031