What Is the Moment of Sleep Onset for Insomniacs?

Transcription

1 Sleep, 6(1): 10-\ Raven Press, New York What Is the Moment of Sleep Onset for Insomniacs? Peter Rauri and Elaine Olmstead Dartmouth Medical School, Hanover, New Hampshire, U,S,A, Summary: Subjective estimates of sleep latency were compared with three EEG-assessed measures of sleep onset: (a) the traditional one, i,e" the first epoch that is scored as stage 2 sleep; (b) the beginning of the first 15 min of uninterrupted stage 2 sleep; and (c) the beginning of the first 30 min of un interrupted stage 2 sleep, A total of 56 insomniacs and 10 good sleepers were studied for 3 nights each in the laboratory, The traditional measure of sleep latency agreed best with the subjective estimates of good sleepers, Most insomniacs, however, were best able to estimate their sleep latency when the 15-min criterion was used, We suggest that for most insomniacs the subjective experience of being asleep occurs later in the EEG-defined transition from waking to sleeping than it does for good sleepers. Key Words: Sleep onset Insomnia-Subjective sleep latency. Sleep researchers have traditionally used the first sleep spindle or K-complex as an objective marker for sleep onset. Good sleepers are fairly accurate at judging sleep latency when using this criterion. Insomniacs, however, seriously overestimate their sleep latency when this criterion is used (1-5). It seems possible that insomniacs indeed overestimate their "true" sleep latency, as is traditionally claimed. Alternatively, it seems also possible that for insomniacs the experience of sleep onset actually occurs at a later time during the EEG-assessed transition from waking to sleeping. Rechtschaffen and Monroe (6) have observed that poor sleepers often report they were awake when aroused from the first 10 min of stage 2 after sleep onset, while good sleepers typically feel asleep during this same EEG-defined sleep stage. Concerning this issue, it also remains unclear whether the various subgroups of insomniacs all exhibit the same pattern of overestimating sleep latency, or whether this occurs only in some, but not all, patient subgroups. Accepted for publication August Address correspondence and reprint requests to Peter J. Hauri, Ph.D., Dartmouth-Hitchcock Sleep Disorders Center, Hinman Box 7770, Dartmouth Medical School, Hanover, New Hampshire

2 WHEN DOES SLEEP BEGIN? 11 METHODS Subjects An insomniac was defined for this study as a person with a serious complaint of inadequate sleep persisting for at least 2 years. Each insomniac included in this study had received a psychiatric and a medical evaluation resulting in an AS DC diagnosis (7), and each had slept for at least two consecutive riights in the laboratory. In our files we found records of 56 insomniacs who fitted these criteria and who were also classified in one of the four ASDC categories of interest in this study. Ofthese, 48 had been physician-referred to the sleep disorders center and 8 had been self-referred for inclusion in ongoing research. For comparison, we also included in the study 10 self-described good sleepers from the same age range as the insomniacs. The following four subgroups were chosen for analysis: (a) insomnia complaint without objective findings (n = 18); (b) psychophysiological insomnia (n = 20); (c) medical insomnia (n = 8); and (d) insomnia associated with depression (n = 10). For a detailed description of these subgroups, see the AS DC nosology (7). For a diagnosis of insomnia complaint without objective finding (subjective insomnia), patients had to have complained about serious insomnia for 2 years, while showing both a sleep efficiency of better than 90% and a sleep latency ofless than 30 min on at least two of the three laboratory nights. There were 7 females and 11 males in that group, with a mean age of 31 years. Typically these patients felt that they slept slightly better in the laboratory than at home, but that even in the laboratory they still suffered from insomnia. To be included in the other three groups of insomniacs, patients had to show less than 85% sleep efficiency (percent asleep of their own bedtime) or longer than 45 min sleep latency (time to first sleep spindle) on at least two ofthree laboratory nights. Psychophysiological insomniacs (11 females, 9 males, mean age 47) were those who were judged free from serious psychiatric or medical conditions but still slept poorly, both by complaint and by sleep laboratory criteria. Medical insomniacs (3 females, 5 males, mean age 50) included three with central sleep apneas, three with myoclonus, and one each with chronic pain and fibrositis. Finally, the insomniacs associated with depression (6 females, 4 males, mean age 54) all had been referred by physicians for a primary complaint of insomnia, not depression. Major affective disorders were excluded from this study. Rather, this group satisfied the DSM III criteria for dysthymic patients. The 10 good sleepers (5 females, 5 males, mean age 42) were solicited from the staff of our medical center. They claimed to have no problems with sleeping and were classified as psychologically and medically healthy. Procedures and analysis In the laboratory each patient followed the same bedtimes and arousal times as at home. Total time in bed was kept constant for the three lab nights, but was varied from one subject ot the next. Usually it fell between 7.5 and 8 h. Each sleeper also filled out a post-sleep questionnaire comparing laboratory sleep of that night with average sleep at home. The mean rating for each of the insomniac groups on each of the laboratory nights was slightly above average, whereas the Sleep, Vol. 6, No.1, 1983

3 12 P. HAURI AND E. OLMSTEAD good sleepers, on the average, felt that they slept slightly worse than average in the laboratory situation. Each subject was recorded and scored according to the standard Rechtschaffen and Kales (8) procedures. In addition, respiration, EKG, and tibialis activity were recorded on at least one laboratory night. EEG-assessed sleep latency was scored according to three different criteria. (a) traditional: elapsed time from lights out to the first epoch scored stage 2, i.e., to the first sleep spindle or K-complex followed by at least 30 s scored stage 2. (b) 15 min: from lights out to the beginning of the first 15 min of solid sleep, stage 2 or deeper. Solid means that these 15 min were not interrupted by any epoch scored either as awake or as stage 1 sleep. (c) 30 min: from lights out to the beginning of the first 30 min of solid sleep (criteria as in b). The three sleep latency criteria are identical if a person falls asleep once and then experiences no arousal for at least 30 min. The more the early sleep is riddled by awakenings and stage 1 arousals, the more discrepant the three criteria become. For each night in the laboratory, a subject's EEG-assessed sleep latency according to the three sleep onset criteria was subtracted from his or her subjective sleep latency as estimated in the morning. Means of these discrepancy scores were then calculated for each group and for each of the three nights. These means are shown in Table 1, together with their standard deviations. TABLE 1. Differences between subjective and EEG-assessed sleep latencies Difference scores (min) Subjective minus EEG-assessed sleep latencies Night Type of sleeper (#) Traditional a 15 Min b 30 Mine Good sleepers 1.9 (4.1) -8.4 (19.7) (49.3) (9.7) (40.6) (67.5) (5.7) (24.8) (167.9) Insomnia complaint (22.8) 5.8 (22.6) (46.0) without objective (20.6) 1.2 (21.9) (66.3) findings (19.1) -2.7 (19.7) (101.4) Psychophysio (30.9) -8.9 (37.0) (109.6) logical insomniacs (35. J) 2.6 (42.0) (92.7) (40.5) -3.3 (28.2) (85.9) Medical insomniacs (68.8) 13.9 (49.9) (194.8) (89.3) 2.0 (46.7) (206.9) (62.3) -1.9 (43.5) (233.3) Depressed I 54.4 (109.4) 44.0 (108.6) (62.6) insomniacs (112.5) 63.6 (100.6).2 (101.2) (47.0) 14.4 (45.4) (89.5) a.b.e EEG-assessed sleep latency is the time elapsed between lights out and (a) the first epoch scored stage 2; (b) the beginning of the first IS min of stage 2 sleep not interrupted by epochs scored as awake or stage 1; (c) the beginning of the first 30 min of uninterrupted sleep (criteria as in b). Standard deviations are indicated in parentheses. Boldface: best match. Sleep, Vol. 6, No. I, 1983

4 WHEN DOES SLEEP BEGIN? 13 RESULTS As reported in Table 1, the traditional marker of sleep onset (the first sign of stage 2) fits extremely well with the subjective estimate of sleep latency in good sleepers. Indeed, according to this criterion, good sleepers could estimate their sleep latency with a mean accuracy of 1-2 min. All insomniacs, however, overestimated by a wide margin their sleep latency according to this measure, as had been expected from the literature discussed earlier. According to the 15-min criterion, good sleepers seriously underestimated their sleep onset. However, the insomniacs without objective findings, the psychophysiological insomniacs, and the medical insomniacs au did quite well by this criterion. Indeed, these insomniacs were about as accurate in their sleep latency estimates using this criterion as the good sleepers were using the traditional one, considering that the insomniacs' task was much more difficult. Insomniacs often have to estimate sleep latencies lasting 0.5 to 2 h, and they did so with an accuracy of a few minutes. The good sleepers, on average, had to estimate sleep latencies lasting only a few minutes, a much simpler task. The depressed insomniacs had more difficulty in estimating their sleep latencies than the three other types of insomniacs. For nights 1 and 2, a 30-min criterion seemed to fit their subjective estimates best, but on night 3 a 15-min criterion seemed to fit best. It is well known that depressives often show a disturbed perception of elapsed time, even when fully awake, and their difficulty with estimating sleep latency may simply be an extension of that finding. Table 1 also shows that the standard deviations of the discrepancy scores are relatively high. This indicates that it is, indeed, very difficult to estimate with accuracy how long it takes to fall asleep. Important for purposes of this study is the fact that the standard deviations reported for insomniacs do not increase as one moves from the traditional criterion to the I5-min one. In other words, Table 1 indicates that the 15-min criterion seems to yield about the same random error variance as the traditional criterion for sleep latency, but that for insomniacs it eliminates the systematic bias. DISCUSSION It is often claimed that the EEG is an objective or true measure of sleep. Such a statement overlooks the fact that originally certain EEG waves were called "sleep" or "wake" because they correlated highly with both behavioral and subjective indices of sleep or wakefulness (9). However, although the EEG is a very convenient marker for normal wakefulness or sleep, the correlation is not perfect. Indeed, as early as 1963 Kleitmann (10) discussed 12 different studies specifically showing that behavioral and EEG indices of sleep often disagreed, and he severely questioned the adequacy of using EEG patterns as the sole objective indices of sleep. For example, atropine in dogs results in an EEG sleep pattern while the dogs are unmistakably awake (11). In humans, certain types of brain damage slow the EEG to sleep-like frequencies during obvious behavioral wakefulness (12). Ether narcosis, on the other hand, yields a pattern of EEG wakefulness during obvious behavioral sleep. In such instances we claim that, although Sleep. Vol. 6, No. I, 1983

5 14 P. HAURI AND E. OLMSTEAD the EEG waves appear sleeplike or wakelike, the person is not truly asleep or awake. The data presented here suggest that a similar dissociation between the true state of the individual and the EEG state may occur during sleep onset in insomniacs. In the early studies on the EEG as an indicator of sleep and wakefulness (9), only normal volunteers were used, and their subjective estimates of sleep onset correlated highly with the first appearance of stage 2 sleep. This has led to an exaggerated confidence in our traditional measure of sleep latency. However, it may be that the first stage 2 is purely a convenient marker of sleep onset in normals, usually, but not necessarily, associated with the time when the basic process of sleeping starts. In the insomniac we often see frequent alternations between stage 2 and wakefulness around sleep onset. The data in Table 1 might indicate that the basic process of sleeping does not start in insomniacs until the brain is ready to settle into at least 15 min of solid sleep. In practical terms, we see no advantage to be gained in accepting as a fact the statement, found in much of the literature, that insomniacs habitually overestimate the time it takes them to fall asleep. Rather, we find that most of them are quite able to estimate sleep latency accurately once a criterion has been developed that is suitable for them. We suggest that the beginning of the first 15 min of uninterrupted sleep, stage 2 or better, may be such a criterion. It seems worth noting that, in this aspect, the insomniacs without objective findings follow the same pattern as do the other insomniacs. This finding suggests that such patients are not just normal sleepers who complain fortuitously about poor sleep. Rather, they may be people who experience the onset of sleep in the same way as do other insomniacs, i.e., long after the EEG has produced the first sleep spindles. In sum, we suggest that the traditional measure of sleep latency may be invalid for insomnia. We recommend that the first sign of solid sleep (i.e., the first epoch scored as stage 2 that is followed by at least 15 min of uninterrupted sleep) be used instead as the marker of sleep onset in insomniacs. Acknowledgments: This work was supported by grant MH from the National Institute of Mental Health. Technical support was provided by L. Percy and B. Hayes, and medical supervision by Drs. C. Hellekson and M. Sateia. REFERENCES 1. Baekeland F, Hoy P. Reported vs. recorded sleep characteristics. Arch Gen Psychiatry 1971;24: Johns MW. Preliminary communication: factor analysis of objective and subjective characteristics ofa night's sleep. Psychol Med 1975;5: Bixler EO, Kales A, Leo LA, Slye T. A comparison of subjective estimates and objective sleep laboratory findings in insomniac patients. Sleep Res 1973;2: 143 (Abstract). 4. Carskadon MA. Dement WC, Mitler MM, Guilleminault C, Zarcone VP, Spiegel R. Self-reports versus sleep laboratory findings in 122 drug-free subjects with complaints of chronic insomnia. Am J Psychiatry 1976;133: Frankel BL, Coursey RD, Buchbinder R, Snyder F. Recorded and reported sleep in chronic primary insomnia. Arch Gen Psychiatry 1976;33: Rechtschaffen A, Monroe LJ. Laboratory studies of insomnia. In: Kales A, ed, Sleep physiology and pathology, a symposium. Philadelphia: J. B. Lippincott Co., 1969: Sleep, Vol. 6. No.1, 1983

6 WHEN DOES SLEEP BEGIN? Association of Sleep Disorders Centers. Diagnostic classification of sleep and arousal disorders, first edition, prepared by the Sleep Disorders Classification Committee, HP Roffwarg, Chairman. Sleep 1979;2: Rechtschaffen A, Kales A, eds. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Brain Information Service/Brain Research Institute, University of California at Los Angeles, Loomis AL, Harvey EN, Hobart GA. Cerebral states during sleep as studied by human brain potentials. J Exp Psychol 1937;21: Kleitman N. Sleep and wakefulness, second edition. Chicago: The University of Chicago Press, Wikler A. Pharmacologic dissociation of behavior and EEG "sleep patterns" in dogs: morphine, N-allylnormorphine, and atropine. Proc Soc Exp Bioi Med 1952;79: Hamoen AM. Signs of sleep in the EEG of waking patients. Electroencephalogr Clin Neurophysiol 1954;6: Sleep, Vol. 6, No, 1, 1983