Eye Movements During Sleep & Dream Activity PDF

Journal of Experimental Psychology Vol. S3, No. 5, 1957 THE RELATION OF EYE MOVEMENTS DURING SLEEP TO DREAM ACTIVITY: AN OBJECTIVE METHOD FOR THE STUDY OF DREAMING WILLIAM DEMENT i AND NATHANIEL KLEITMAN 2 Department of Physiology,...

Journal of Experimental Psychology Vol. S3, No. 5, 1957 THE RELATION OF EYE MOVEMENTS DURING SLEEP TO DREAM ACTIVITY: AN OBJECTIVE METHOD FOR THE STUDY OF DREAMING WILLIAM DEMENT i AND NATHANIEL KLEITMAN 2 Department of Physiology, University of Chicago The study of dream activity and possibility of unintentional cuing by its relation to physiological variables E. (b) The subjective estimate of during sleep necessitates a reliable the duration of dreams was compared method of determining with precision with the length of eye movement when dreaming occurs. This knowl- periods before awakening, reasoning edge, in the final analysis, always that there should be a positive cor- depends upon the subjective report relation if dreaming and eye move- of the dreamer, but becomes relatively ments were concurrent, (c) The pat- objective if such reports can be sig- tern of the eye movements was related nificantly related to some physio- to the dream content to test whether logical phenomena which in turn can they represented a specific expression be measured by physical techniques. of the visual experience of dreaming Such a relationship was reported or merely a random motor discharge by Aserinsky and Kleitman (1) who of a more active central nervous observed periods of rapid, conjugate system. eye movements during sleep and METHOD found a high incidence of dream recall The Ss for the experiments were seven adult in Ss awakened during these periods males and two adult females. Five were studied and a low incidence when awakened intensively while the data gathered from the at other times. The occurrence of other four were minimal with the main intent of these characteristic eye movements confirming the results on the first five. and their relation to dreaming were In a typical experiment, S reported to the laboratory a little before his usual bedtime. He confirmed in both normal Ss and was instructed to eat normally but to abstain schizophrenics (4), and they were from alcoholic or caffeine-containing beverages shown to appear at regular intervals on the day of the experiment. Two or more in relation to a cyclic change in the electrodes were attached near the eyes for depth of sleep during the night as registering changes in the corneoretinal potential fields as the eyes moved. Two or three elec- measured by the EEC (5). trodes were affixed to the scalp for recording This paper represents the results brain waves as a criterion of depth of sleep. of a rigorous testing of the relation The S then went to bed in a quiet, dark room. between eye movements and dream- All electrode lead wires were further attached to the top of the head and from there to the lead ing. Three approaches were used: box at the head of the bed in a single cord to (a) Dream recall during rapid eye minimize the possibility of entanglement and movement or quiescent periods was allow S a free range of movement. The po- elicited without direct contact be- tentials were amplified by a Model III Grass tween E and S, thus eliminating the Electroencephalograph in an adjoining room. The electroencephalograph was run continuously 1 Postdoctoral Public Health Service Research throughout the sleep period at a paper speed of Fellow of the National Institute of Neurological 3 or 6 mm. per sec. which allowed easy recog- Diseases and Blindness. nition of eye-movement potentials. A faster 2 Aided by a grant from the Wallace C. and speed (3 cm./sec.) was used for detailed ex- Clara A. Abbott Memorial Fund of the Uni- amination of the brain waves although the versity of Chicago. slower speed permitted at least an approximate 339 340 WILLIAM DEMENT AND NATHANIEL KLEITMAN TABLE 1 through a stage similar to that ac- SUMMARY OF EXPERIMENTS companying the rapid eye movement periods occurring later in the night. Nights Awak- Average Average These findings concerning associated 5s Slept Nightly enings Awakenings Sleeping Time EEG patterns were identical with previous observations on uninter- DN 6 50 8.3 7:50 IR 12 65 5.4 4:20 rupted sleep (5). KC 17 74 4.4 6:00 An accurate appraisal of the mean WD 11 77 7.0 6:30 duration of the REM periods was PM 9 55 6.1 6:20 impossible since most were terminated KK 2 10 5.0 6:00 SM 1 6 6.0 6:40 artificially by an awakening. How- DM 1 4 4.0 7:00 MG 2 10 5.0 6:10 ever, those that were not so termi- nated varied between 3 and SO min. Totals 61 351 S.7 6:00 in duration with a mean of about 20 min., and they tended to be longer estimation of the gross pattern. The criteria of the later in the night they occurred. eye-movement potentials and their differentia- The eyes were not constantly in tion from brain wave artifacts have been dis- motion during such periods; rather, cussed at length elsewhere (1,4). At various times during the night Ss were the activity occurred in bursts of one awakened to test their dream recall. The return or two, up to fifty or a hundred to sleep after such an awakening invariably took movements. A single movement was less than 5 min. Table 1 is a summary of the generally accomplished in.1-.2 sec. experiments showing the number of nights each and was followed by a fixational pause S slept and the number of awakenings. In all, 21% of the awakenings fell in the first 2 hr. of of varying duration. The amount, sleep, 29% in the second two, 28% in the third pattern, and size of the movements two, and 22% in the fourth two. varied irregularly from period to period. RESULTS The REM periods occurred at The occurrence of rapid eye move- fairly regular intervals throughout the ments?—Discrete periods during night. The frequency of occurrence seemed to be relatively constant and which their eyes exhibited rapid movements were observed in all nine characteristic for the individual. DM Ss every night they slept. These and WD averaged one eye-movement periods were characterized by a low- period every 70 min. and every 75 voltage, relatively fast pattern in the min. respectively. KC averaged one EEG. The interspersed periods in eye-movement period every 104 min. which rapid eye movements were The other Ss fell between these two absent showed EEG patterns in- extremes. The average for the whole dicative of deeper sleep, either a group was one REM period every 92 predominance of high-voltage, slow min. activity, or frequent, well-defined Despite the considerable disturb- sleep spindles with a low-voltage ance of being awakened a number of background. No REM's were ever times, the frequency and regularity observed during the initial onset of with which REM periods occurred sleep although the EEG always passed was almost exactly comparable to 3 that seen previously in a study of In most of the remaining text the following uninterrupted sleep (5). If the awak- abbreviations will be used: REM's (rapid eye movements) and NREM's (no rapid eye move- ening occurred during a NREM ments). period, the return to sleep was never  EYE MOVEMENTS DURING SLEEP 341 associated with REM's, nor was the TABLE 2 time of onset of the next REM period INSTANCES OF DREAM RECALL AFTER AWAK- markedly changed from that which ENINGS DURING PERIODS OF RAPID EYE would have been expected in the MOVEMENTS OR PERIODS OF No RAPID EYE MOVEMENTS absence of an awakening. An awak- ening during an REM period gen- Rapid Eye No Rapid Eye erally terminated the REM's until Movements Movements the next period, and the sequence of 5 EEC changes, excluding the brief Dream No Dream No Recall Recall Recall Recall period of wakefulness, was the same as that following an REM period DN 17 9 3 21 IR 26 8 2 29 that ended spontaneously. Excep- KG 36 4 3 31 tions occurred when S was awakened WD 37 5 1 34 during an REM period in the final PM 24 6 2 23 KK 4 1 0 5 hours of sleep when the period was SM 2 2 0 2 likely to be quite long if uninter- DM 2 1 0 1 MG 4 3 0 3 rupted. On these occasions, the REM's sometimes started up again Totals 152 39 11 149 when S fell asleep. It seemed as though a period of heightened CNS activity had not run its normal course mentary impressions of content, were and, although S was able to fall considered negative. asleep, he continued to dream. The awakenings were done either Eye movement periods and dream during REM periods or at varying recall.—For all awakenings to elicit increments of time after the cessation dream recall, the arousing stimulus of eye movements during the inter- was the ringing of an ordinary door- spersed periods of NREM's. The Ss, bell placed near the bed and suffi- of course, were never informed when ciently loud to ensure immediate awakened whether or not their eyes awakening in all levels of sleep. The had been moving. Ss then spoke into a recording device Table 2 shows the results of the near the bed. They were instructed attempts to recall dreams after the to first state whether or not they had various awakenings. The REM or been dreaming and then, if they could, NREM awakenings for PM and KC to relate the content of the dream. were chosen according to a table of When S had finished speaking E, random numbers to eliminate any who could hear their voices, occasion- possibility of an unintentional pat- ally entered the room to further tern. For DN, a pattern was fol- question them on some particular lowed: first three REM awakenings, point of the dream. There was no then three NREM awakenings, and communication between S and E in so on. WD was told he would be any instance, it must be emphasized, awakened only when the recording until S had definitely commited him- indicated that he was dreaming, but self. The Ss were considered to have REM and NREM awakenings were been dreaming only if they could then interspersed randomly The relate a coherent, fairly detailed type of awakenings for IR was description of dream content. As- chosen according to the whim of E sertions that they had dreamed with- The Ss uniformly showed a high out recall of content, or vague, frag- incidence of dream recall following 342 WILLIAM DEMENT AND NATHANIEL KLEITMAN REM awakenings and a very low In general, Ss were best able to incidence of recall following awak- make an emphatic statement that enings during periods of NREM's they had not been dreaming when the regardless of how the awakenings NREM awakenings were done during were chosen. In particular, DN was an intermediate stage of sleep as not more accurate than the others indicated by a brain-wave pattern of although there was a pattern he spindling with a low-voltage back- might have learned, and WD was not ground. When aroused during a less accurate although he was de- deep stage of sleep characterized by liberately misled to expect to have high-voltage, slow waves in the EEG, been dreaming every time he was Ss often awoke somewhat bewildered. awakened. Over a narrow range, In this state they frequently felt that some Ss appeared better able to recall they must have been dreaming al- dreams than others. though they could not remember the Table 3 compares the results of the dream or, on the other hand, that first half of the series of REM awak- they had not been asleep at all. They enings with the last half. Practice sometimes had a great variety of was certainly not a significant factor feelings to describe—such as pleasant- as only one S showed any degree of ness, anxiety, detachment, etc., but improvement of recall on later nights these could not be related to any as compared with the early ones. specific dream content. The incidence of dream recall Most of the instances of inability dropped precipitously almost imme- to recall dreaming after awakenings diately upon cessation of REM's. In during REM periods occurred in the 17 NREM awakenings that were done early part of the night. Of 39 nega- within 8 min. after the end of a REM tive reports in the entire study, 19 period, 5 dreams were recalled. Al- occurred after awakenings during though small, this was a much higher REM periods falling in the first 2 hr. incidence of dream recall than oc- of sleep, 11 after REM awakenings curred when the NREM awakenings during the second 2 hr., S in the third followed the end of REM periods 2 hr., and 4 in the last 2 hr. There by more than 8 min. In the latter was no such variation relating to category only 6 dreams were recalled awakenings during the interspersed in 132 awakenings. periods of ocular quiescence, the incidence of dream recall being uni- TABLE 3 formly low, regardless of whether the COMPARISON OF FIRST HALF OF SERIES OF RAPID early or late part of the night was EVE MOVEMENT AWAKENINGS WITH SECOND HALF being considered. Length' of rapid eye movement periods First Half Second Half and subjective dream-duration esti- 5 mates.—If the length of the REM Dream No Dream No periods were proportional to the Recall Recall Recall Recall subjectively estimated duration of the DN 12 1 5 8 dreams, it would further help to IR 12 S 14 3 establish the relatedness of the two KC 18 2 18 2 WD 19 2 18 3 and would give some information PM 12 3 12 3 about the rate at which dreaming Total 73 13 67 19 progresses. At first, Ss were awakened at  EYE MOVEMENTS DURING SLEEP 343 TABLE 4 ever, the lengths of the dream narra- RESULTS OF DREAM-DURATION ESTIMATES tives still showed a significant relation- AFTER S OR 15 MIN. OF RAPID ship to the duration of REM periods EYE MOVEMENTS before awakening. Table S shows the correlations between minutes of 5 Minutes 15 Minutes REM's and lengths of dream nar- 5 ratives for each S. The number of Right Wrong Right Wrong words in the narrative was the meas- DN 8 2 S S urement of length. Of the 152 IR 11 1 7 3 dreams recalled, 26 were not included KC 7 0 12 1 WD 13 1 IS 1 because poor recording did not allow PM 6 2 8 3 complete transcription. Dream nar- Total 4S 6 47 13 ratives recalled after 30 or as much as SO min. of REM's were not a great deal longer than those after IS min. various increments of time after the although Ss had the impression that REM's had begun and were requested they had been dreaming for an un- to estimate to the nearest minute the usually long time. This was perhaps amount of time they had been dream- due to inability to remember all the ing. This proved to be too difficult, details of very long dreams. although the estimates were always Specific eye-movement patterns and of the same order of magnitude as the visual imagery of the dream.—The lengths of the REM periods, and were quality and quantity of the REM's occasionally exactly right. themselves showed endless variation. A series was then done in which Ss There was much or little movement, were awakened either S or IS min. big or small movements, and so on. after the onset of REM's and were As has been stated, the movements required on the basis of their recall occurred in bursts of activity sepa- of the dream to decide which was the rated by periods of relative inactivity. correct duration. The 5- or IS-min. However, the brain-wave stage during periods were chosen on the basis of a the whole period remained the same random series. Table 4 shows the whether there was much or little results of these awakenings. All Ss movement at any given moment of were able to choose the correct dream the period. duration with high accuracy except It was hypothesized that the move- DN. This S, however, made most ments represented the visual imagery of his incorrect choices by estimating of the dream, that is, that they IS min. to be 5 min. This is con- sistent with the interpretation that TABLE 5 the dream was longer, but he was only CORRELATION BETWEEN DURATION OF REM PERIODS IN MINUTES AND NUMBER OF able to recall the latter fraction and WORDS IN DREAM NARRATIVES thus thought it was shorter than it actually was. Number Subjects r P In addition to depending on the of Dreams amount of actual dreaming, the DN IS.60