9.
10.
11. 12.
19, 449, 1117 (1972); F. Lamprecht and J. T. Coyle, Brain Res. 41, 503 (1972). B. A. Campbell and P. D. Mabry, Psychopharmacologia 31, 253 (1973); C. Kellog and P. Lundberg, ibid. 23, 187 (1972). S. M. Snyder and J. L. Meyerhoff, Ann. N. Y. A cad. Sci. 205, 310 (1973). L. L. Iversen, Science 188, 1084 (1975); J. W. Kebabian, F. E. Bloom, A. L. Steiner, P. Greengard, ibid. 190, 157 (1975). S. K. Sharpless, Fed. Proc. Fed. Am. Soc. Exp.
Biol. 34, 1990 (1975); R. K. Mishra, E. L. Gardner, R. Katzman, M. H. Makman, Proc. Natl. Acad. Sci. U.S.A. 71, 3883 (1974). 13. P. H. Wender, Life Sci. 14, 1605 (1974). 14. B. A. Shaywitz, D. J. Cohen, M. B. Bowers, Jr. Pediatr. Res. 9,285 (1975). 15. Supported by research grants NS12384 and a grant from the Charles H. Hood Foundation (to B.A.S.). 12 September 1975
Sleep During Transcendental Meditation Abstract. Five experienced practitioners of transcendental meditation spent appreciable parts of meditation sessions in sleep stages 2, 3, and 4. Time spent in each sleep stage varied both between sessions for a given subject and between subjects. In addition, we compare electroencephalogram records made during meditation with those made during naps taken at the same time of day. The range of states observed during meditation does not support the view that meditation produces a single, unique state of consciousness. In 1970, Wallace reported several physiological changes observed during transcendental meditation (TM) (1). His results were replicated and extended by Wallace, Benson, and Wilson (2) and they were subsequently made available to a wider audience (3). They found, in meditating subjects, reduced oxygen consumption, increased skin resistance, increased alpha activity in the electroencephalogram (EEG), decreased heart rate, and decreased blood lactate. Although many of these changes take place in ordinary relaxed wakefulness and in sleep, Wallace and his co-workers postulated that, during most of the meditation period, experienced practitioners of TM enter a single, unique state of consciousness, a 'wakeful hypometabolic state,' that differs from ordinary relaxed or sleep states. The Stanford Research Institute estimates that, from a few hundred in 1965, the number of practitioners of TM has increased to more than 240,000 as of June 1973. Estimates from the TM organization
indicate that this number now exceeds 900,000 (4). The findings of Wallace and his co-workers are often cited to prospective meditators and may have played an important role in producing this increase. We have found that meditators spend considerable time in sleep stages 2, 3, and 4 during meditation; their subjective reports of sleep confirm our analysis of the EEG records. Further, our data suggest that the meditation period is not spent in a single, unique, wakeful, hypometabolic state. The five subjects we observed had at least 2.5 years of experience with TM, and four of them were teachers of the technique. All were male Caucasians between the ages of 20 and 30, accustomed to meditating for 40-minute periods twice each day, and not in the habit of napping. Subjects reported, on the average, 7.8 hours of sleep per night. Psychophysiological measures were made on each subject during ten sessions, each of which lasted 40 minutes. During five of these sessions, the subjects were
Table 1. Percentage of time spent in stages 2, 3, or 4 during each session. Subject
551
I 2 3 4 5
0 49 0 37
Meditation 2 90 0 0 90 0
session 3 59 0 0 59 86
Nap session 4 78 26 78 74 31
1
37 78 86 18 95
2 41 92 31 38 95
3 59 79 83 95 93
4 62 58 89 88 78
Table 2. Percentage of time spent in each stage, averaged over sessions.*
Meditation 1 2 3 4 5
w
1
19 44 53 37 43
12 46 15 6 17
*These percentages do not sum to
308
Nap
1 W 3,4 27 42 32 17 6 0 7 14 16 15 15 12 28 27 31 8 23 1 15 7 100 because some epochs were scored as movement time. 2
2
40 62 31 51 54
3,4 10 14 41 9 36
asked to meditate in their accustomed sitting position, and in the other five sessions, they were asked to nap lying down on a bed. The first nap and the first meditation were scheduled on the first observation day. The data collected on this day are not included for analysis here because initial unfamiliarity with the laboratory situation produces atypical sleeping patterns (5). On eight subsequent days, subjects were asked either to meditate or to nap. These sessions were all conducted in the afternoon within 2 hours of the same time each day. The order in which the two types of sessions were scheduled followed an irregular pattern, and subjects were not told whether they would be asked to meditate or to nap on a particular day until they arrived in the laboratory. If a subject reported that his previous night's sleep was more than 30 minutes shorter than normal, he did not take part on that day. Subjects were asked not to consume food, coffee, or tea for at least 2 hours before each session. At the beginning of each session, electrodes were applied so that occipital, central, and frontal EEG responses, eye movements, submental (below the chin) muscle potentials, and skin resistance level could be measured (6). The subject then moved to the room where he was to meditate or nap. A 45-db white noise partially masked any disturbance from the adjoining apparatus room (7). The room in which the subject sat during meditation was dimly illuminated, but the room was dark when the subject lay down to nap. Once the recording was proceeding smoothly, the subject was asked to relax for 5 minutes with his eyes closed, and then a signal was given to begin meditation or napping. After 40 minutes, an identical signal required the subject to stop meditating or napping and to relax with his eyes closed for an additional 5 minutes before leaving the recording room. At the end of the session, the subject filled out a questionnaire on his subjective impressions of what had transpired and stated whether he had slept or become drowsy during the meditation or nap. The most striking feature of our data is that meditators spent appreciable amounts of time in EEG sleep stages 2, 3, and 4 while they were meditating (Fig. 1). Averaged over meditation sessions, we found that 39 percent of the time was spent in wakefulness (stage W), 19 percent in stage 1, 23 percent in stage 2, and 17 percent in stages 3 or 4. More than a quarter of the meditation time was spent in stages 2, 3, or 4 in 13 out of the 20 meditation sessions (Table 1). It is customary to identify stages 2, 3, and 4 as sleep and stage 1 as drowsiness (8); according to these conventional designations, our subjects were asleep durSCIENCE, VOL. 191
ing, on the average, 40 percent of their STAGE 1 meditation time. C4 -M I EE G Meditation might produce a dissociation between the EEG and consciousness that Left EOG _ Submentol would permit a subject to be awake even EMG though his EEG record indicated sleep (9). STAGE 2 However, this does not appear to have oc- C4-M1 EEGcurred in our study, because our subjects reported having slept in 12 of the 13 medi- Left EG tation sessions in which patterns of stages Submental E MG --2, 3, or 4 appeared. In addition, they re- STAGE 3 ported feeling drowsy in 18 of the 19 sesE sions during which they spent more than 30 C4-M1 EEGseconds in stage 1. The consistency of the Left EOG-reports with the EEG rcords indicates that Submental the conventional EEG criteria defining EMGSTAGE 4 sleep and drowsiness were applicable. No rapid eye movement (REM) sleep C4-M1 EEG-A was observed during either the meditations ~ N Left EOG ~ ]y or the naps. This is probably because we I second Submental 20 conducted all sessions during the afternoon EMG S*Ipv and limited each session to a length of 40 Fig. 1. Representative records from subject 5 minutes; REM sleep does not normally oc- during a meditation session. The time scale and cur during the first 40 minutes of an after- channel gain are shown on the stage 4 record. noon nap (10). Although meditation in a laboratory might lead to a state different from that ject (from meditation to meditation) and outside the laboratory, our subjects' rat- between subjects (Tables I and 2). For ings of their meditations indicated that in 7 example, subject 2 slept in only one of his of the 13 sessions in which stage 2 was ob- four meditations, whereas subject I slept served, the subject rated his meditation as more than half the time in each of his four typical rather than atypical. Further, on a meditations. Subjects 3, 4, and 5 each had 7-point scale from 7 (very deep) to 1 at least one meditation in which they did (very light), the modal depth of meditation not sleep at all and another in which they was 5, and there was no significant correla- slept for more than three-fourths of the tion between reported depth of meditation session. What emerges from these EEG and the amount of time asleep. Thus, in findings is that meditation is an activity several meditations described as typical that gives rise to quite different states both and relatively deep, considerable amounts from day to day and from meditator to of sleep occurred. This corroborates re- meditator. Our data differ from the EEG responses ports that we have received from these and other meditators that they occasionally fall reported by Wallace (1). Only 4 of his 15 asleep while meditating in their normal subjects occasionally evidenced drowsisettings. ness, and he states, 'The EEG pattern durIf TM produces the wakeful state de- ing meditation clearly distinguishes this scribed by Wallace (1), one would expect state from the sleeping state. There are no to find less sleep during meditation than slow (delta) waves or sleep spindles, but alduring a nap period. An analysis of vari- pha-wave activity predominates.' Several ance of time spent in sleep stages 2, 3, or 4 factors may account for the differences berev_aled no significant differences between tween Wallace's data and ours. He reportmeditation and nap sessions (F = 3.2, ed on records from just one session per P > .1). Because we obtained repeated subject, presumably the first experience for measures over sessions for each subject, we the subject in the laboratory. In addition, also carried out individual t-tests on each many of his subjects meditated while subject's data. Only in subject 2 was there breathing through a mouthpiece or with a significant difference (t = 7.3, P < .01) arterial cannulae or rectal thermometers in indicating fewer EEG sleep patterns during place (1, 2). Both of these factors would meditations than during naps. Because of probably tend to activate the EEG more the high variability in the states observed, than would be expected in a normal mediwe caution against the conclusion that tation session outside the laboratory. meditation and napping produce identical Wallace's subjects meditated for 20 to distributions of EEG stages. 30 minutes, whereas our sessions lasted 40 One of the striking features of our data minutes; it could be argued that sleep was was the variability in the time spent in the more likely to occur in our longer sessions. various EEG stages both for a single sub- But when we examined the data from the V
]
23 JANUARY 1976
first 20 minutes of each of our subject's meditations, the discrepancies remained: In the first 20 minutes, an average of 42.5 percent of the time was spent in sleep stages 2, 3, and 4. In three other studies EEG responses were recorded during transcendental meditation: Younger et al. report that advanced meditators spent 41 percent of their meditations in sleep stages 1 and 2 (11); Wada and Hamm also found sleep stages 1 and 2 in the EEG records of both experienced and inexperienced meditators (12); Banquet recorded EEG responses during meditation but did not present an analysis by sleep stage (13). He did, however, mention the presence of 'short bursts of large amplitude delta waves identical to those of sleep stage 4.' The results of Younger et al., of Wada and Hamm, and of this experiment raise the question of whether the beneficial effects reported for meditation (14) are due to the sleep that occurs during meditation or to some other feature of that process. ROBERT R. PAGANO, RICHARD M. ROSE ROBERT M. STIVERS STEPHEN WARRENBURG Department ofPsychology, University of Washington, Seattle 98195 References and Notes 1. R. K. Wallace, Science 167, 1751 (1970). 2. H. Benson, A. F. Wilson, Am. J. PhysioL 221, 795 (1971). 3. R. K. Wallace and H. Benson, Sci. Am. 226 (No. 2), 84 (1972). 4. D. P. Kanellakos and J. S. Lukas, The Psychobiology of Transcendental Meditation: A Literature Review (Stanford Research Institute, Menlo Park, 1973), p. iii; Fundamentals of Progress: Scientific Research on Transcendental Meditation (Maharishi International Univ. Press, New York, 1975). 5. See, for example, H. W. Agnew, Jr., W. B. Webb, R. L. Williams, Psychophysiology 2, 263 (1966). 6. Electrodes for the EEG were placed according to the international 10-20 system, and recordings were taken between each of leads 02, C4, and F4 and a reference electrode on the opposite mastoid. A Beckman Dynograph (type RM) was used to record the data, and the records were scored according to the criteria of Rechtschaffen and Kales [A. Rechtschaffen and A. Kales, Eds., A Manual of Standardized Terminology, Techniques, and Scoring System for Sleep Stages of Human Subjects (Public Health Service Publ. No. 204, Govemnment Printing Office, Washington, D.C., 1968), pp. 8-15]. The scorers had no knowledge of the condition under which a record was made; the agreement between the two scorers averaged 93 percent. 7. In order to make additional comparisons between meditation and naps beyond those that we report here, a 45-db, 600-hertz tone of 0.5-second duration was presented on an irregular schedule averaging one presentation per minute. This tone was found to evoke EEG responses without disturbing the course of meditation, as judged by pilot subjects. In this report we present only the sleep stage and sleep report data. The galvanic skin responses and the responses to the tone have not yet been an___
alyzed.
8. F. Snyder and J. Scott, in Handbook of Psychophysiology, N. S. Greenfield and R. A. Stemback, Eds. (Holt, Rinehart & Winston, New York, 1972), p. 645. 9. E. Green, in Biofeedback and Self-Control 1972, D. Shapiro et al., Eds. (Aldine, Chicago, 1973), p. 164; A. Jus and K. Jus, in Proceedings of the Third International Congress of Psychiatry, B. Stoukis, Ed. (Montreal, 1961), pp. 473-482. 10. L. Maron, A. Rechtschaffen, E. A. Wolpert, Arch. Gen. Psychiatry 11, 503 (1964); S. A. Lewis, Br. 309
J..Psychiatry 115, 107 (1969); 1. Karacan, W. W. Firnley, R. L. Williams, C. J. Hursch, Biol. Psychiatry 2,261 (1970). 11. J Younger, W. Adriance, R. Berger, Percept. Mot. Sk rills 40, 953 (1975). 12. J..A. Wada and A. E. Hamm, paper presented at the 27th annual meeting, American EEG Society, Boiston, Mass., 15 and 16 June 1973. 13. J. P. Banquet, Electroencephalogr. Clin. Neuro-
phjysiol. 35, 143 (1973).
14. Sete, for example, H. Benson and R. K. Wallace [in
Dr-ug Abuse: Proceedings of the International
Conference, C. J. D. Zarafonetis, Ed. (Lea & Febiger, Philadelphia, 1972), p. 369]; S. Nidlich,
W. Seeman, T. Dreskin [.. Couns. Psychol. 20,
565 (1973)1; and J. Robbins and D. Fisher [Tranquility Without Pills (Wyden, New York, 1972)]. 15. We thank A. Lubin, G. Chatrian, D. Barash, and A. Marlatt for constructive comments. We thank the Student International Meditation Society Seattle Center for their cooperation in helping us locate subjects. 17 June 1975; revised 12 November 1975.
Str(ess-Induced Hyperphagia and Obesity in Rats: A P'ossible Model for Understanding Human Obesity
is not limited to eating, but appears to be
appropriate to the particular goal object at
instance, if rat pups are present during tail pinch, maternal behavior occurs hand. For
(7). A similar stimulus tail shock induces copulatory behavior in naive male rats in the
presence
of
a
receptive female, and
ag-
gression in the presence of another male
(8). When all goal objects have been removed from the testing arena, the incidence of grooming (washing of the face and flanks) and nail pulling of the paws increases markedly. There also appears to be
an increased tendency to vocalize and to
attempt to escape, which suggests that the Ab)stract. Mild tail pinch administered to rats several times daily in the presence of performance of a goal-directed behavior sweet*ened milk induced immediate hyperphagia and led to considerable gain in body may mask the perception of aspects of the weighit. Parallels are drawn with stress-induced hyperhagia and altered affective states in pinch related to stress. obese humans. There are some interesting, although indirect, parallels betweeneatinglin rats inHy'perphagia, and the obesity which typ- conditions, there is an enhanced reactivity duced by tail pinch and hyperphagia in huically accompanies it, is one of the major to other goal objects; for example, there is mans. Both may reflect a stress-related inprobllems of modern society. Although the a higher incidence of other oral activities crease in responsiveness to environmental causa.l factors underlying hyperphagia are such as smoking, and there is an increase stimuli. This study provides evidence for a nume rous, in many cases they appear to be in sexual activity (1, 4). Stress-related hy- more direct parallel: in rats with access to relate d to stress (1). For example, many perphagia may therefore be a food-direct- a highly palatable fluid food, chronic stress obese patients tend to eat when they are ed manifestation of a more general hyper- induced by repeated tail pinch leads to emotiionally tense or during other unpleas- responsivity to environmental stimuli (3). dramatic hyperphagia, weight gain, and Recent research in this laboratory has visible obesity. ant states, such as depression and boredom. Reports suggest that eating dimin- shown that a mild nonspecific stress, tail Twenty-four adult female rats (Spragueishes or prevents these states (2). Usually pinch, reliably induces eating in sated rats Dawley) were determined to readily ingest there is a well-defined pattern to the hyper- (5). This behavior occurs with short la- milk from a hand-held drinking burette phagiia, and, in most cases, a definite fin- tency, appears quite normal, and proceeds during mild tail pinch. They were ranickint ess exists. While bland foods are not without obvious pain. Such animals are domly assigned to one of three experimeneaten to excess, palatable foods stimulate finicky, displaying an increased preference tal conditions: (i) surgically intact (N = 6), hyperrphagia (1, 3.). When food is not read- for highly palatable fluids and familiar (ii) ovariectomized (N = 10), or (iii) ovariectomized and injected daily with hormone ily avtailable, such as when obese individ- foods (6). The behavior observed during tail pinch replacement (1 Ag of estradiol benzoate) uals Eare forced to adhere to strict dieting (N= 8). Three animals from each of these conditions were assigned to the pinched 100 group, and the rest served as weight> Best tail pinch animal matched controls. Both groups were indiPoorest tail pinch animal Best control ovidually housed and allowed free access to 80 - Poorest -l control -_ w sweetened milk (9) and tap water. In addiT,al pinc period tion, the experimental animals received six daily pinch sessions (10 to 15 minutes each, L. Iw ! spaced at equal intervals throughout the 60 24-hour cycle) in the presence of a handSD held burette containing milk (10) for up to five consecutive days. The tail was pinched a l. _ u; - >/between I and 2 inches from the tip with 40 ' a hand-held, 25-cm hemostat, padded at /Iw the tips with foam rubber. Testing was / .iSDO conducted in a wire mesh cage 15 inches (38 cm) square. Only slight pressure was required to induce ingestion of milk (it rarely ap__ @ _',> ,~ - | _ _',/' ----~~~~---~~= 9' _ - proached the first notch of the hemostat) and infrequently resulted in any indication (for example, squealing) that the animals were attending to the pinch. Immediately I I g I I I I J I after the beginning of thepninch, animals -20 0 10 2 4 6 8 12 14 16 would explore the testing arena for a few seconds and then begin licking the milk Time (days) Fig. 1 Weight changes (in grams) for the best and poorest animals in both tail pinch and control tube. Ingestion usually proceeded contingroup,s. Data reflect comparisons before, during, and after the period of testing. uously for several minutes at a time and
|
20
0
310
SCIENCE, VOL. 191