Biological Psychology 9 (1979) 155-161 © North-Holland Publishing Company

SLEEP LEARNING DURING STAGE 2 AND REM SLEEP * ANDREW J. TILLEY **

Department of Psychology, University of Hull, Hull, U.K. Accepted for publication 31 October 1979

Pairs of subjects were presented with a 20-item picture series at bedtime. In the latter part of the night, a tape-recorded series of 10 words, the verbal equivalents of half the original series of pictures, was repeated 10 times during either Stage 2 or REM sleep. Morning recall and recognition for repeated words was found to be facilitated following repetition during Stage 2 sleep, but relatively unaffected following repetition during REM sleep. However, adjusting for recall, the number of additional words elicited through recognition was found to be significantly greater for REM repeated words than for Stage 2 repeated words. It was suggested that retrieval limitations, perhaps as a result of REM state dependency, rather than storage inhibition may be the main locus of the initial recall failure. By comparison, Stage 2 sleep would seem to present both a lower barrier to memory storage and retrieval compatibility with wakefulness. 1. Introduction This paper deals with the possibility of reinforcing pre-sleep learning by repetition trials during sleep. Particular interest is focused on REM and Stage 2 sleep. Despite the lack of standardised procedures, some degree of sleep learning or hypnopedia has been reported in all but one of the dozen or so major studies *** relating to this phenomenon that fulf'd the absolutely essential requirement of having monitored and recorded the EEG. Two main conclusions can be drawn from these studies: (1) Subsequent retention is normally only possible when the information input occurs in the presence of low-amplitude EEG activity exhibiting alpha frequencies. * This research was supported by a Science Research Council Studentship. ** Present address: Medical Research Council, Applied Psychology Unit (Psychophysiology Section), 5 Shaftesbury Road, Cambridge, England. *** Simon and Emmons (1956); Emmons and Simon (1956); Koukkou and Lehman (1968, 1973); Portnoff, Bakeland, Goodenough, Karaean and Shapiro (1968); Evans and Orchard (1969); Lehman and Koukkou (1971, 1974); Tani and Yoshii (1970); Bruce, Evans, Fenwick and Spencer (1970); Evans, Gustafson, O'Connell, Orne and Shor (1970); Jus and Jus (1972); Cooper and Hoskovec (1972); Levy, Coolidge and Stoab (1972); Johnson (1972). 155

156

A.J. Tilley / Sleep learning

(2) The probability of subsequent retrieval is directly related to the duration of the post-input EEG activation. In a comprehensive review of the literature, Aarons (1976)points out that sleep assisted instruction (SAI), as he has labelled the topic, is apparently only successful when 'contaminated' by pre-sleep learning trials, incidently a common design feature of Soviet studies in this field (Dodge and Lamont, 1969). In attempting to explain this apparently limiting factor, Aarons speculates that SAI constituted of material studied prior to sleep might harmoniously reinforce natural processes of sleep mentation. Such stimulation may act to prime ongoing mentation for sleep rehearsal of the learning material, or perhaps it reinforces the course of memory consolidation during sleep (cf. Tilley and Empson, 1978). In a similar fashion, stimulation during sleep may facilitate retention through direct reactivation of attributes of the stimulus events originally associated with their memory (cf. Neisser, 1967). In view of the fact that REM sleep is characterised by low-amplitude EEG activity in which alpha frequencies are sometimes present, it would seem reasonable to hypothesise that this stage of sleep Should be more compatible with processing incoming information from external sources than NREM sleep. Therefore, bearing in mind the apparently potentiating effects of pre-sleep learning trials, the aim of the present experiment was to compare the post-sleep retention of material presented at bedtime following repetition during either REM or NREM (Stage 2 in this case) sleep. Dillon and Babor (1970) and Dillon and Bowles (1976) have reported closely related studies, but both have serious methodological flaws. In the first study, it was reported that retention was improved following repetition during sleep. However, sleep was only defined behaviourally, not electrophysiologically, so the possibility arises that the subjects may not have been fully asleep on all occasions during 'sleep' presentation. In the second study, it was reported that repetition of a list of paired associates during NREM sleep improves post-sleep learning acquisition rates, whereas repetition during REM sleep tended to impair acquisition rates, although not significantly so. But the former condition was not restricted to any one NREM sleep stage in particular, thus potentially confounding possible stage specific effects related to known differential arousal thresholds between the various NREM sleep stages (Rechtschaffen, Hauri and Zeitlin, 1966). Methodological improvements and tighter controls have been introduced into the present study. 2. Method

Thirty subjects took part in the experiment. They were run in pairs, each pair spending two consecutive nights (one familiarisation, one experimental) in the sleep laboratory where they were wired-up for sleep recordings (EEG, EOG and EMG were monitored). The experiment was introduced to subjects as a study into the

A.J. Tilley / Sleep learning

157

combined effects of visual images at bedtime and monotonous speech during sleep on dream patterns. This minor and temporary deception was intended to minimise the temptation on the part of the subjects to indulge in pre-sleep rehearsal by keeping them unaware of the morning retention tests. They were told the real purpose of the experiment before leaving the sleep laboratory. While lying in bed and immediately prior to lights-out (23.15-23.30) on their second night, each subject was individually presented with a series of 20 single item pictures (children's picture book type drawings) at a rate of approx, one picture every 3 sec. (Pictures were used instead of words at this point in an attempt to increase the imagery value (Paivio, 1971) of the word equivalents presented during sleep. It was thought that this might facilitate their incorporation into dream mentation.) Immediately after lights-out, a continuous repetitive series of names was played over the intercom system at quiet speech levels. This tape loop was continued for at least 45 min or until both subjects had fallen asleep, whichever was longest. They had been forewarned of similar repetitions throughout the night. The intention of the whole exercise was to fully prepare and accustom subjects to the sound of speech during their sleep. At 0400 houis, or thereabout, the tape loop was re-started. Occasionally this caused momentary arousals, but sleep was quickly resumed. The names were continued for at least 15 min and thereafter until one subject exhibited REM sleep while the other exhibited Stage 2 sleep. (This technique was primarily devised to control for 'time-of-day' effects). The tape was then switched to another track which contained one of two repetitive word series spoken at the same rate as the name series (i.e. one every 5 sec). The word series was repeated 10 times and represented the verbal equivalent (i.e. the name of the object in the picture) of half the original series of 20 pictures. Half the subjects in each condition were presented with the word-labels corresponding to the even numbered pictures in the original series, and half with the appropriate words corresponding to the odd numbered pictures. Subjects were awakened at 0700 hours and required to complete two simple retention tests. First, they had to give a written free recall of the picture series. Second, they had to select those items from a 60-word recognition list they were confident they had seen pictures of at bedtime.

3. Results

Those subjects who did not exhibit uncontaminated, indubitable REM or Stage 2 sleep during and for a few seconds after word presentation were excluded from the results analysis - for example, if Stage 1 sleep, alpha (more than 2 see bursts) or wakefulness intruded into the sleep record. In addition to maintaining the counterbalancing between and within groups with respect to the two ten-word series, this left a ffmal complement of eight subjects per group.

158

A.J. Tilley / Sleep learning

Table 1 Mean number of items recalled

REM Stage 2

Repeated

Non-repeated

4.37 6.12

4.37 4.25

Tables 1 and 2 show respectively the morning recall and recognition scores following verbal repetition of half a 20-item pre-sleep learning list during either REM or Stage 2 sleep. Two-way analysis of variance with repeated measures on one factor revealed that the repetition of words during sleep significantly improves recall (F = 5.34, df 1, 14, p < 0.05). However, a significant interaction term (F = 5.34, df 1, 14, p < 0.05) suggests that this main effect is due to improved recall following repetition during Stage 2 sleep, whereas repetition during REM sleep appears to be neither beneficial nor detrimental to subsequent recall (see table 1). Direct comparisons substantiate this position. Repeated words are recalled significantly better than non-repeated words in the Stage 2 condition (F= 10.68, df 1, 14,p 0.05). Furthermore, words repeated during Stage 2 sleep are recalled significantlybetter than words repeated during REM sleep (F = 5.7, df 1, 14,p < 0.05). Reflecting recall performance, repetition during sleep also significantly improves recognition (F = 11.67, df 1,14, p < 0.05). The effect is most pronounced in the Stage 2 condition where repeated words are recognised significantly better than non-repeated words (F= 14.93, df 1,14, p < 0.05), whereas the difference in the REM condition is non-significant (F = 0.93, df 1,14, p > 0.05). The recognition of words repeated during Stage 2 sleep, on the other hand, is not significantly higher than the recognition of words repeated during REM sleep (F = 1.3, df 1,14, p > 0.05). The above recognition scores are undoubtedly contaminated by the preceding recall test. Table 3, therefore, shows the number of additional words elicited through recognition. Significantly more unrecalled REM repeated words are recognised than unrecalled Stage 2 repeated words (F = 5.12, df 1,14, p < 0.05). This

Table 2 Mean number of times recognised

REM Stage 2

Repeated

Non-repeated

8.12 9.00

7.62 7.00

A.J. TiUey / Sleep learning

159

Table 3 Mean number of unrecalleditems reeognised

REM Stage 2

Repeated

Non-repeated

3.75 2.88

3.25 2.75

would tend to suggest that the initial recall failure may be primarily due to retrieval difficulties as opposed to storage limitations.

4. Discussion The reinforcement of pre-sleep learning by the repetition of material during either Stage 2 or REM sleep only significantly improves retention in the former condition. There is evidence, however, that the limitations on REM sleep learning are largely related to retrieval difficulties during wakefulness rather than storage restrictions during REM sleep. The present results complement the findings of Dillon and Bowles (1976) who reported that pre-exposure during REM sleep facilitated learning acquisition rates. They concluded that 'facilitation resulted from an increased readiness to respond induced by the pre-presentation during REM. This improved readiness would then translate into more rapid consolidation...'. An explanation along similar lines may account for the improved retention following repetition during Stage 2 sleep found in the present study. For example, information presented during Stage 2 sleep may reinforce or reactivate memory storage processes which subsequently leads to improved retention. But whatever the explanation, it is safe to say that the conditions of storage during Stage 2 sleep are compatible with the conditions of retrieval during wakefulness. The same may not apply to REM sleep. The apparent failure of REM sleep learning might have been accounted for in terms of storage inhibition, perhaps as a result of attenuated information processing capacity acting as a buffer against endogenous REM processes. However, since retrieval is improved when extra cues are provided by recognition, it follows that some information is being processed and stored during REM sleep but that its retrieval is restricted during wakefulness. This retrieval inhibition may in turn signify a state dependent learning phenomenon operating as a transfer barrier between REM sleep and wakefulness. In addition to uncovering a possible state dependent learning phenomenon, the present experiment has also demonstrated that, under the present experimental conditions and from a visual analysis of the sleep records, neither low-amplitude, desynchronised EEG activity nor alpha frequencies are essential prerequisites for sleep learning. These EEG characteristics are normally only associated with relaxed

160

A.Z Tilley / Sleep learning

wakefulness, Stage 1 sleep or REM sleep and are not a distinguishing feature o f Stage 2 sleep. However, they may still optimise sleep learning in the absence o f prior learning trials during wakefulness. In conclusion, it is clear that the controversial file on the sleep learning issue should be re-opened and examined in more detail, particularly with a view to specifying the most conducive experimental conditions and parameter setting for successful sleep learing. On a broader front, it is also important to continue the investigation o f the cognitive functioning o f the brain during sleep to find out how this differs from the waking state and its relevance thereto.

References Aarons, L. (1976). Sleep assisted instruction. Psychological Bulletin, 83, 1-40. Bruce, D.J., Evans, C.R., Fenwick, P.B.C. and Spencer, V. (1970). Effect of presenting novel verbal material during slow-wave sleep. Nature, 225,873-874. Cooper, L.M. and Hoskovec, J. (1972). Hypnotic suggestion for learning during stage 1-REM sleep. American Journal of Clinical Hypnosis, 15, 102-111. Dillon, D.J. and Babor, T.F. (I970). Intervening activity and the retention of meaningful verbal material. Psychonomic Science, 19,369-370. Dillon, D.J. and Bowles, E.G. (1976). The learning of meaningful verbal material following prepresentation during either REM or non-REM sleep the previous night. Research Communications in Psychology, Psychiatry, and Behaviour, 1, 3 1 5 - 326. Dodge, C.H. and Lament, E. (1969). Sleep learning in the USSR. (ATD Report 68-91-108-6). Aerospace Technology Division, NTIS No. AD-682527). Emmons, W.H. and Simon, C.W. (1956). The non-recall of material presented during sleep. American Journal of Psychology, 69, 76-87. Evans, F.J. and Orchard, W. (1969). S|eep learning: The successful waking recall of material presented during sleep. Psychophysiology, 6,269. Evans, F.J., Gustafson, L.A., O'Connell, D.N., Orne, M.J. and Shot, R.E. (1970). Verbally induced behavioural responses during sleep. Journal of Nervous and Mental Disorders, 150. 171-187. Johnson, W.H. (1972). Suggestibility, time, stage of sleep, and the effects upon subsequent learning of hearing paired associates during sleep. Dissertation Abstracts International, 33, 5018B. Jus, K. and Jus, A. (1972). Experimental studies on memory disturbances in humans in pathological conditions. International Journal of Psychology, 2,205-218. Koukkou, M. and Lehman, D. (1968). EEG and memory storage in sleep experiments with humans. Electroeneephlography and Clinical Neurophysiology, 25,455-462. Koukkou, M. and Lehman, D. (1973). Human learning and EEG analysis in sleep experiments. In: Jovanovic, V.J. (Ed.). The Nature of Sleep. Fischer: Stuttgart. Lehman, D. and Koukou, M. (1971). Das EEG des mensehen biem lernen yon neumen und bekanntem material. Archive fur Psychiatrie und Nervenkrankheiten, 215, 22-32. Lehman, D. and Koukkou, M. (1974). Computer analysis of EEG wakefulness-sleep patterns during learning of novel and familiar sentences. Electroencephalography and Clinical Neurology, 37, 73-84. Levy, C.M., Coolidge, F.L. and Stoab, L.C. (1972). Paired associate learning during EEGdefined sleep: A preliminary study. Australian Journal of Psychology, 24, 219-225. Neisser, V. (1967). Cognitive Psychology. Appleton-Century-Crofts: New York.

A.J. Tilley / Sleep learning

161

Paivio, A. (1971). Imagery and Verbal processes. Holt, Rinehart and Winston: New York. Portnoff, G., Bakeland, F., Goodenough, G.R., Karacan, I. and Shapiro, A. (1968). Retention of verbal materials perceived immediately prior to onset of NREM sleep. Perceptual and Motor Skills, 22, 751-758. Rechtschaffen, A., Hauri, P. and Keitlin, M. (1966). Auditory awakening thresholds in REM and NREM stages. Perceptual and Motor Skills, 22,927-924. Salzarulo, P. and CipoUi, C. (1974). Spontaneously recalled verbal material and its linguistic organisation to different stages of sleep. Biological Psychology, 2, 47-57. Simon, C.W. and Emmons, W.H. (1956). Responses to material presented during various levels of sleep. Journal of Experimental Psychology, 51, 89-97. Tani, K. and Yoshii, N. (1970). Efficiency of verbal learning during sleep as related to EEG pattern. Brain Research, 17,277-285. TiUey, A.J. and Empson, J.A.C. (1978). REM sleep and memory consolidation. Biological Psychology, 6,293-300.

Sleep learning during stage 2 and REM sleep.

Biological Psychology 9 (1979) 155-161 © North-Holland Publishing Company SLEEP LEARNING DURING STAGE 2 AND REM SLEEP * ANDREW J. TILLEY ** Departme...
357KB Sizes 0 Downloads 0 Views