Behavior and Sleep. Subjective and Objective Confirmation of the Ability to Self-Awaken at a Self-Predetermined Time Without U sing External Means

Transcription

1 Sleep. 20(1 ): American Sleep Disorders Association and Sleep Research Society Behavior and Sleep Subjective and Objective Confirmation of the Ability to Self-Awaken at a Self-Predetermined Time Without U sing External Means William H. Moorcroft, * Krista Hennager Kayser and t Anthony J. Griggs Sleep and Dreams Laboratory, Department of Psychology, Luther College, Decorah, Iowa, U.S.A.; *The University of Iowa, Iowa City, Iowa, U.S.A.; and tdepartment of Ophthalmology, The Mayo Clinic, Rochester, Minnesota, U.S.A. Summary: The ability to awaken oneself from sleep at a preselected time without external means (such as alarm clocks) was studied using, first, subjective and, second, objective methods. First, in a telephone survey of 269 un selected adults, over one-half said that they never use an alarm clock (or other external means) or always awaken before it. Another 24% said that they sometimes awaken before the alarm. Furthermore, this ability positively correlated with age and was related to consistency in the amount of nightly sleep but not consistency in wake-up time. Second, 15 people who said they regularly self-awaken were objectively tested for this ability in their own beds using actigraphy for three consecutive nights while choosing their own wake-up times. Five awoke within 10 minutes of their target time (mostly before) on each night, five did so on two of the three nights, and of the remaining five. four did so on one night. Choice of target times varied considerably within subjects but more so for those who were more successful. Taken together these results show that many people have the ability to regularly awaken themselves from sleep at a desired time and that such an ability is of practicable utility. Key Words: Selfawakening-Actigraphy-Alarm clocks-awakening. The ability of a substantial percentage of humans to awaken without the aid of external means (hereafter referred to simply as self-awakening or awakening without an alarm) has been examined various times over the last century. These endeavors included single subject studies (1,2), surveys (3,4), sleep diary experiments (1,5-10; see also 11) and, more recently, laboratory studies using polysomnography (10,12-14). Reviews of this research have reached different conclusions. Some say that while not all people seem to have, or at least use, this ability, there is sufficient evidence to affirm its existence (5,10,11,15). Others have maintained that the ability is more apparent than real (16). The single subject studies and the surveys have produced positive results while the sleep diary research and laboratory experiments have been less definite. The surveys generally show that over one-half of Accepted for publication September Address correspondence and reprint requests to William H. Moorcroft, Sleep and Dreams Laboratory, 700 College Dr., Luther College, Decorah, IA 52101, U.S.A. 40 adults say they regularly awaken without an alarm or before it goes off (3,4). The sleep diary experiments have primarily used college students not selected for self-awakening ability and produced positive (5) but sometimes unimpressive results (6,7,8) even when some subjects were selected for their professed ability to self-awaken (9,10). For example, Carskadon et al. (6) found only 220 of 1172 nights resulted in reports of self-awakenings within ± 15 minutes of the target time and only five of the 93 subjects reported being able to come this close to the target on 50% or more of the awakenings. The sleep laboratory results have been mixed. For example, in Zepelin's laboratory study (13) in which only four of 15 subjects were self-claimed self-awakeners, the results were unimpressive with some selfawakenings close to the target time but with no individuals, even the professed self-awakeners, able to consistently awaken near the target time without error. Zung and Wilson (14), using un selected subjects, found 14 of 44 self-awakenings within 10 minutes of the randomly selected target times. In contrast, Lavie.)

2 SELF-A WAKENING AT A PREDETERMINED TIME 41 et al. (12) used seven self-proclaimed self-awakeners and found that they were successful on eight of 12 awakenings. Although previous research has indicated that the ability to self-awaken at a predetermined time may exist in at least some proportion of the general population, questions about it still remain. The present research attempted to more precisely verify the existence of this ability, to determine to what extent it is utilized by adults, and to ascertain some of its determining factors. The first experiment sought to replicate previous survey results by assessing what proportion of adults report the ability to regularly self-awaken, to compliment previous findings by asking self-proclaimed self-awakeners what methods they use, and to gather correlative data on factors such as age, sex, shift work, and the consistency of hours of sleep per night. However, survey data relies on the ability of people to be reliable and to respond honestly, thus, it may not always be completely accurate, causing skepticism about any conclusions that may be derived (16). For this reason, a second experiment was undertaken to obtain objective confirmation that adults who report that they never use an alarm when sleeping in their own beds are actually able to awaken when they desire with sufficient accuracy and consistency to be of practicable value. While prior objective studies of the ability to self-awaken at a specified time have been conducted in sleep laboratories with experimenter-chosen target times (10,12-14), the subjects in this experiment slept in their own beds at home and chose their own target time in accordance with their customary routine. This enabled the assessment of the practicable utility of self-awakenings away from any interference of the unfamiliar sleep laboratory environment. This was similar to Bell's (5) research in that the subjects were in their home bed but differed in that the intended arousal times were entirely freely chosen by the subjects and that the recording of the actual awakening time was objective. Experiment 1 METHODS Telephone calls were made by randomly picking residential phone numbers from U.S.A. telephone books. If an adult did not answer, the caller asked for the person whose name was listed in the phone book. The resultant subjects were 269 un selected adults from the (mainly Midwestern) United States. There were 140 females and 129 males, ranging in age from 21 to 84. After explaining the purpose of the call and obtaining consent, each subject was asked demographic questions such as sex and age, then questions about general sleep habits and sleep patterns including whether the subject was involved in shift or rotational work, and finally specific questions about how they awaken in the morning. The responses were written down by the caller. Respondents were immediately categorized into one of four groups according to their self reported typical method of awakening. These groups were people who: 1) never use an alarm or external source, 2) use an alarm yet stated that they can awaken before the alarm goes off, 3) use an alarm but said they sometimes awaken before the alarm goes off, and 4) use an alarm and said they do not awaken before the alarm goes off. Subsequent questions depended upon to which of these groups the respondent belonged and were aimed at obtaining more detailed descriptions of the particular methods regularly used by that individual to awaken. Experiment 2 The subjects in Experiment 2 were 15 individuals from a Midwestern city in the United States who volunteered in response to a newspaper advertisement asking for people who were able to regularly awaken at a predetermined time without an alarm or were able to always awaken before the alarm. There were nine females and six males, ranging in age from 19 to 62, who successfully participated in this experiment. All subjects were fully informed about the experiment and then gave their consent to participate. A sleep diary was used containing 14 questions, some to be answered prior to falling asleep and the rest upon awakening. Importantly, this diary contained questions about the subject's intended wake-up time, which was to be completed before. going to bed at night, and if any external stimuli awakened them from sleep, which was to be recorded in the morning. An Ambulatory Monitoring, Inc. actigraph was used to record the sleep and wake periods including the time of the final awakening of the subjects. An actigraph is a movement-activated recording device programmed to store the frequency of a subject's movements per minute throughout the recording period. The actigraph was chosen for this experiment because it is relatively unobtrusive and can easily be used in the subject's own bed. Sleep is indicated with a high degree of accuracy by the absence of activity in actigraph recordings (see 17). The time each subject awoke each morning was shown by the abrupt and sustained onset of activity on the actigraph recordings. This objective indicator of time of awakening was subsequently compared to the intended time of awakening recorded prior to sleep onset. Subjects were given the actigraph with instruction to wear it at night, starting at 8:00 p.m., and to con- Sleep, Vol. 20, No. I, 1997

3 42 W H. MOORCROFT ~ ~ 40 >-.!: ;- 30 Cl <t ; 20 Q) :2! 10 FIG. 1. groups. after sometimes before no alarm before Wake-up Relative to Alarm Mean age (years) of subjects in each of the four wake-up 90 c ~ 80 I!! ~ 70 C Co Q) 60 Q) en >- 40 U c 30.S! 1/1 'iii 20 c ~ 0 after sometimes before no alarm before Wake-up Relative to Alarm FIG. 2. Percent of subjects in each wake-up group who report a consistency in the amount of sleep per night. tinue to wear it until about 10 minutes after they got up in the morning. They were also asked to fill out the top section of the sleep diary before going to bed and the rest of it after getting up in the morning. They were instructed to do all of this for three successive nights. Otherwise they were allowed to go about their regular activities and routines. Data on the actigraphs were periodically downloaded to an IBM-P'e. computer at the Mayo Clinic Sleep Disorders Center in Rochester, MN from which the wake-up time was determined to the minute. Experiment 1 RESULTS The results of Experiment 1 indicated that 23% of the 269 respondents never use an alarm; that is, they regularly awaken themselves from sleep at a time determined prior to going to sleep. There were also respondents who use an alarm but consistently awaken before the alarm (29%) and some who can do this but only some of the time (24%). Only 24% of the respondents consistently relied on an alarm or other external source to awaken; yet, over one-half of them related that they had experienced awakening at least once before their alarm. When respondents in the first three groups were asked how they awakened without or prior to an alarm, the most common response involved a sense of having some kind of an internal clock (26%). This was sometimes coupled with the bedtime ritual of actively focusing on a specific time to awaken. Other common responses included habit or daily routine (16%) and stress or nervousness about something anticipated the next day (4%). The remaining 54% consisted of various other responses including having no idea of how they were able to do this. Additional responses from Sleep, Vo!' 20, No. J, 1997 some of the subjects reflected their perceptions that they could also program themselves to awaken from brief naps after a set time had elapsed. Figure 1 shows that age was found to be a factor in the ability to awaken without an alarm; those never using an alarm were significantly older [F (3, 265) = 17.25, P < 0.05; followed by Tukey's HSD], However, sex [X 2 (3) = 1.02, P > 0.05] and shift-work [X 2 (3) = 6.52, P > 0.05] did not significantly influence this ability. Figure 2 shows the percent of respondents in each group who report a consistency in the amount of sleep per night, which was also found to be a significant factor [X 2 (3) = 40.34, P < 0.05]; specifically, the respondents who never use an alarm are significantly more consistent in their duration of nightly sleep [X 2 (1) = 7.58, P < 0.05], In contrast, consistency in the time of awakening may not be a determining factor because many subjects in response to a question stated that they felt they could easily change their awakening time such as when needing to catch an early flight. Experiment 2 The results were analyzed by comparing the intended wake time that the subjects noted in their sleep diary before they went to sleep to the actual awakening time as indicated on the actigraph recording. The difference between these times was derived for each of the three mornings. A positive difference indicated that the subject had overslept while a negative difference indicated that the subject had awoken before the intended time. There were 44 nights of usable data because the actigraph failed on one night. The overall average was minutes (overslept by 3 minutes and 20 seconds) with a standard deviation of minutes. A paired t test did not reject the null hypothesis [t (43) = -0.63, p < 0.05] suggesting that the intended and actual awakening times did not signifi- 1-\.

4 SELF-A WAKENING AT A PREDETERMINED TIME 43 I/) Cl c: c: CIl -" ;: '" <t: '0... CIl.c E :::I Z FIG. 3. > Self-awakenings Relative to Target (in minutes) Frequency of awakenings relative to target times. cantly differ from one another. The 95% confidence interval for the difference between actual and intended awakening times was from to minutes showing that most subjects were able to awaken themselves very close to the intended time even after a mean of 7.06 hours of sleep. Twenty-eight of the nights resulted in awakenings within 15 minutes of target and 35 were within 30 minutes. One-half of the 44 awakenings were within 7 minutes of the target (mean = 3 minutes, 27 seconds). Twenty-five awakenings were before target, 19 after (see Fig. 3). Five subjects were within 10 minutes of their target each of their three consecutive nights; they can be considered consistently successful. Another five succeeded awakening within 15 minutes of their target on two of the three nights; they can be considered moderately successful. Of the remaining subjects, four succeeded only once within 15 minutes of target. The subjects who were more accurate were also slightly more likely to awaken before rather than after target time. Twelve of the subjects chose different awakening times on at least one of their three experimental nights (mean difference = 81.4 minutes). It is noteworthy that the subjects who were more successful at awakening close to their target times were primarily the ones who were not as consistent in their choice of target times. DISCUSSION Overall, experiments 1 and 2 indicate that many people have the ability to awaken themselves from sleep at a predetermined time, often with considerable accuracy and reliability. Both the subjective and objective data confirm that this ability does exist in some people. These data not only confirm past research findings (3-5,18,19) but extend them by objectively showing that some people who say they rely on their ability to self-awaken in their own beds at their own chosen wake-up times are usually quite good at it. > 30 Although a random phone survey may not accurately represent the entire U.S.A. adult population, the results of Experiment 1 are nevertheless informative. There are a surprisingly large number of people (76% in this sample) who are able, at least some of the time, to cognitively program themselves to awaken without an alarm or to awaken before the alarm before it goes off. In addition, it was found that age and consistency of nightly sleep duration are related to the ability to awaken at a predetermined time without using an alarm. The average age of the group that consistently did not use an alarm was older than the average age of the other three groups. People who are more consistent in their nightly sleep durations may be more confident in their ability to awaken themselves and/or be more successful in doing so and therefore never use an alarm. In contrast, consistency in the time of awakening from morning to morning does not appear to be necessary for success. The results of Experiment 2 offer objective evidence that most people who say they can wake up in the morning at or very close to the time they determined when going to sleep in their own beds frequently do so even though a full night of sleep had intervened. This supports and extends previous research (10,12,14). It supports it by again showing that this ability exists and is reasonably reliable. It extends it since the subjects freely chose their own awakening times, slept in their own beds, and the actual time of awakening was objectively determined. There are limitations to this ability, however. Some people appear to be better at it than others while some seem to lack the ability entirely. The reasons for these differences are not clear. There are limitations even within those people who do have the ability. For one thing, some (6,9), but not all (10,14), prior research has shown that the ability to self-awaken is either absent or poor early in the sleep period or after a short amount of nocturnal sleep. (Ability to awaken at a desired time after short naps may be an exception based on verbal reports from some of our subjects, but to our knowledge this has not yet been tested.) Conversely, self-awakening has been reported to be more reliable closer to the habitual awakening time (11,20 but not 14) (although our data from Experiment 2 is not consistent with this). These data also help explain why some prior reports have shown poor self-awakening ability (6,13). The subjects used in some prior experiments were not screened for this ability. If, as our survey results show, one-half of adults self-awakening regularly and the other one-half not at all or only sporadically then a random selection of subjects would not be expected to produce impressive results. Experiments in which subjects were screened for self-awakening ability, includ- Sleep, Vol. 20, No.1, 1997

5 44 W. H. MOORCROFT ing this one, have been more positive in their results (e.g. 9,12). Furthermore, some recent reports that did not produce impressive results used mostly college students as subjects (e.g. 6,9). But college students may not be the best subjects for normative sleep research since they tend to be chronically sleep deprived and sleep-phase delayed (21). Research using polysomnography in sleep labs has not produced impressive results either (16). Perhaps self-awakening may be disrupted by the lab and polysomnography procedures. Laboratory polysomnography has been reported to alter some characteristics of sleep cognition (22,23). Indeed, in Experiment 2, when the self-awakening ability of the subjects was objectively assessed in their own beds using actigraphy, the results were more positive than those previously reported using polysomnography in a sleep laboratory. Thus, both the nature of the subjects and methods of assessment may be factors contributing to prior poor results. A third factor may involve the sources of the target time. In our experiment, subjects picked their own wake-up times, whereas in many prior experiments they were chosen by the experimenters. Self-chosen times may provide more motivation for success and less of an arbitrary challenge. Fourth, the criterion for success is another factor that determines if experimental support for this ability is found. That is, just how close to the target must a subject come in order to be considered successful? Likewise, should waking before the target be considered better than oversleeping by the same amount of time? In one review of self-awakening research (16) awakening minutes from target was not considered very good, even though well over 400 minutes may have elapsed since sleep onset. In our second experiment two out of three subjects awoke themselves within 20 minutes of their target, most of them prior to it. It seems to us that this is accurate enough to be of practicable value to many people. Moreover, onehalf of these awakenings were within 7 minutes of the target (with a mean of less than 3lh minutes) and more than two-thirds of these were before or at their selfchosen target time which can be considered very successful and highly valuable. The ability to self-awaken has been discounted (16) since it is not universal, has never been shown to be totally reliable in the laboratory for those in whom it does occur, and the wake-ups occur in or near REM and thus appear to be coincidences of the nature of this stage of sleep. However, these same arguments would eliminate lucid dreaming and rapid eye movement (REM) behavior disorder as legitimate since they are REM-dependent and may occur occasionally in only some people. Another criticism (16) of the ability to self-awaken on schedule is that the person may Sleep. Vol. 20. No sometimes awaken prior to the target time, perhaps check the time, and return to sleep to later awaken close to the target. While this may occur on occasion, and may provide additional external information regarding the time, the sleepers nevertheless do usually subsequently awaken at about the target time. The additional awakenings may weaken the robustness of the phenomenon but do not discredit it as being legitimate and of practicable value. It should not be surprising that a large proportion of humans possesses the ability to self-awaken at a time chosen prior to bedtime. Widespread access to alarm clocks has only been available during a recent, small segment of human existence (24). It is easy to see that at least some humans who could self-awaken at a desirable time would be at an advantage over those who could not, which would cause this ability to be favored during human evolution. This ability would not require precise accuracy to be useful, just to be reasonably close. Our results provide no consistent, single explanation of how successful self-awakenings are accomplished. In fact, the majority of our survey subjects who selfawaken do not know how they are able to do it. Popular explanations gathered from the rest of the successful respondents include habit, regular schedule, stress due to the schedule of the next day, and visualization of a clock set to wake-up time or some other kind of pre-sleep rehearsal. Some researchers (12,13 but not 14 or apparently 5) have implicated the nature and timing of REM sleep as being responsible for selfawakenings; others have maintained that light, restless sleep is necessary for this ability (5,9,11,20). REM sleep has a somewhat regular periodic onset that might serve as an elapsed time indicator. REM sleep is also implicated as a time when the brain is more aroused and thus may be more capable of the cognitive activity necessary to recognize the time, recall the instructions to awaken at the target, and be able to implement them (16) or it simply may be easier to self-awaken from REM (12). Also, the duration of REM tends to be longer later in the sleep period when there is greater accuracy of self-awakening (9). Furthermore, the frequency and duration of REM are such, especially near the end of a typical sleep period, that any spontaneous awakenings that occur during it will coincidentally be close to the target wake-up time (16). There is evidence to show that sleep is lighter or at least more restless on nights when a person is intending to self-awaken (9,11,20). Either the lighter sleep makes success more likely or the stress and anxiety of fear over consequences of failure to awaken on time causes sleep to lighten. If light sleep is indeed necessary for success then this weakens the practicality of self-awakening because the soundness of sleep has to 1 I JI I 1, ~,

6 SELF-A WAKENING AT A PREDETERMINED TIME 45 be compromised in order to self-awaken on schedule. On the other hand, Zung and Wilson (14) postulate a specific arousal system primarily responsible for selfawakening ability that is independent of specific sleep stage or biological circadian rhythms. Bell (5), combining these explanations, proposes that the ability to self-awaken has two components. The first is the ability to "induce in oneself a greater potential for awakening" (p. 507) when desiring to self-awaken at a specific time that leads to generally more awakenings during sleep. The second ability is to use these awakenings to keep track of the passage of time based on "internal, biological processes" (p. 507). According to Bell, an individual would need a high degree of both in order to self-awaken successfully. Overall, it can be concluded that self-awakening at a self-chosen time does occur with considerable accuracy in at least some people in their own beds. It is sufficiently accurate to be of functional, practicable use to them on a regular basis. Differences among people in this ability or less than perfect accuracy does not affect its validity; it is a legitimate phenomenon. Acknowledgements: The help of The Mayo Clinic Sleep Disorders Center, especially Joyce Wisbey, for downloading the actigraph records is gratefully acknowledged. REFERENCES I. Brush EN. Observations on the temporal judgement during sleep. Am J Psych 1930;42: Hall WW. The time sense. J Ment Sci 1927;73: Child CM. Statistics of "unconcious cerebration". Am J Psych 1892;5: Clauser G. Die kopfuhr. Stuttgart: F Enke, Bell CR. Awakening from sleep at a pre-set time. Percept Mot Skills 1980;50: Carskadon MA,Herman KB, Davis SS. Arousals in college students: can they beat the clock. Sleep Res 1990; 19: Hawkins J. Sleep disturbance in intentional self-awakening: behavior-genetic and transient factors. Percept Mot Skills 1989; 69: Hawkins J, Shaw, P. Sleep satisfaction and intentional selfawakening: an alternative protocol for self-report data. Percept Mot Skills 1990;70: Omwake KT, Loranz M. Study of ability to wake up at a specified time. J. Appl Psycho I 1933;17: Tart CT. Waking from sleep at a pre-selected time. J Am Soc Psychosom Dent Med 1976;17:3-16. II. Fraisse P. The psychology of time. New York: Harper & Row Lavie P, Oksenberg A, Zomer J. It's time, you must wake up now. Percept Mot Skills 1979;49: Zepelin H. REM sleep and the timing of self-awakenings. Bull Psychon Soc 1986;24: Zung WK, Wilson WD. Time estimation during sleep. Bioi Psychiat 1971;3: Kleitman N. Sleep and wakefulness. Chicago: University of Chicago Press, Zepelin H. Internal alarm clock. In: Carskadon, MA, ed. Encyclopedia of sleep and dreams. New York: Macmillan, 1993: Cole RJ, Kripke DF, Gruen W, Mullaney DJ, Gillin Jc. Automatic sleep/wake identification from wrist activity. Sleep 1992; 15, Edlund M. Psychological time and mental illness. New York: Gardner Press, Hoagland H. The physiological control of judgments of duration: evidence for a chemical clock. J Gen PsychoI1933;9: Elder J. Influence of assigned hour for waking. Psych Bull 1941; 38: Moorcroft WHo Sleep, dreaming, and sleep disorders: an introduction, 2nd edition. Langham, MD: University Press of America, Piccione P, Thomas S, Roth T, Kramer M. Incorporation of the laboratory situation in dreams. Sleep Res 1976;5: Van de Castle RL. Our dreaming mind. New York: Ballantine, Herman KB. Alarm clocks. In: Carskadon, MA, ed. Encyclopedia of sleep and dreams. New York: Macmillan, 1993: Sleep, Vol. 20, No. J, 1997