Dement and Kleitman (1957)

Psychology Being Investigated

• EEG (electroencephalogram) records brain waves and electrical voltage during sleep to identify REM and nREM stages. It is objective and reliable.
• EOG (electrooculogram) can record the direction and speed of eye movements during sleep by attaching electrodes to the skin around the eyes.
• Dreaming is a subjective experience of imagery while we are asleep.

1. REM ➜ Rapid Eye Movement is when dreaming occurs. A person is paralysed, their eyes flicker and they can breathe but all other muscles are paralysed.
2. NREM ➜ Non-Rapid Eye Movement are the other stages of sleep and they differ in brain activity.

Procedure

Dement and Kleitman (1957) is a classic study that explored sleep and dreaming using EEG and EOG recording alongside participants’ own verbal dream reports.

1. During the day, the participant ate normally (excluding coffee and alcohol) and then arrived at the laboratory just before their normal bedtime.
2. They then went to bed in a quiet (dark) room.
3. The participant went to sleep with electrodes attached beside the eyes (EOG) and on the scalp (EEG), both feeding signals to the experimenter’s room.
4. The EEG wires became a single cord/ponytail (to stop entanglement).
5. The EEG was run continuously through the night (at speeds of 3 or 6 mm per sec).
6. After the electrodes had been fitted, the participants had to go off to sleep.
7. They were then awoken from different rapid-eye movement (REM) and non-rapid-eye movement (nREM) of sleep throughout the night.
8. Participants were woken by a doorbell sound (placed near the bed) in REM/nREM, or they were randomly woken 5 or 15 minutes after REM began.
9. Participants were asked if they dreamed (or not).
10. If they had, they were asked to “please describe the dream that you have just experienced”, into a recording device near their bed.
11. Then, they would go off to sleep again.
12. Sometimes the experimenter would enter the room and ask them more questions to clarify some aspect of the participant’s dream.
13. Initially, participants were asked to estimate the length of time in REM sleep to the nearest minute.
14. In the revised procedure participants were given a fixed choice and asked if they were dreaming for 5 or 15 minutes.

Sample

• 9 adult participants were predominantly male (7 males and 2 females).
• The five main participants were studied intensely: spending between 6 and 17 nights in the laboratory.
• The other four participants were used to confirm the results of the main participants: spending only 1 to 2 nights in the lab.
•  All participants were identified by their initials in the study.
• The researchers do not provide information about how the participants were recruited or selected, nor do they mention any specific demographic details beyond gender.

IV 1: Sleep Stage of Awakening

1. Participants were woken either in REM or nREM (but not told which). They confirmed whether they were having a dream and described the content in a recorder.

• IV = Timing of awakenings (during REM or NREM periods)
• DV = whether a dream was reported.
• DV = report of dream narration content.

IV 2: Duration of REM Sleep

1. Participants were woken randomly either 5 or 15 minutes after REM began, using the forced-choice procedure described above.
2. The number of words in the dream narrative was counted (although this was affected by how expressive the participant was).

• IV = randomly woken after 5 or 15 minutes.
• DV = dream duration:
  • estimate which duration (5 or 15 minutes) they had been dreaming
  • the number of words in the dream narrative was counted as a measure of dream length.

IV 3: Pattern of Eye Movement during REM Sleep

1. The direction of eye movements was detected using electrodes around the eyes (EOG).
2. The eye movement patterns were: mainly vertical, mainly horizontal, both vertical and horizontal, and very little or no movement.
3. Participants were woken after the persistence of a single eye movement pattern for more than one minute and asked to report their dream.

• IV = eye movement pattern type (not manipulated by researchers, so natural experiment)
• DV = report of dream content.

High incidence of dream recall during REM periods (152 out of 191 awakenings) compared to nREM periods (11 out of 160 awakenings)

When woken in nREM participants returned to nREM, but when woken in REM they typically didn’t dream again until the next REM phase (except sometimes in the final REM phase).

DV 2: Duration Estimate of REM Sleep & Dream Narrative Length

The accuracy of estimation of 5 or 15 minutes of REM was very high (88% and 78%, respectively). REM duration and the number of words in the narrative were significantly positively correlated.

Participants could distinguish between shorter and longer dream durations with considerable accuracy.

• Participants were more accurate at estimating 5 mins compared to 15 mins
• On 88% of trials (45/51) the participants estimated 5 mins correctly
• On 78% of trials (47/60) the participants estimated 15 mins correctly

Significant positive correlations between REM period duration and dream narrative length for all participants (r ranging from 0.40 to 0.71)

DV 3: Report of Dream Content Corresponding to Eye Movement Patterns

Eye movement patterns during REM sleep tended to correspond with dream content, although the clearest examples rested on only a handful of observations.

These results suggest that eye movements during REM sleep reflect the dreamer’s gaze within the dream, supporting a direct link between physiological activity and subjective dream experiences.

Conclusions

1. Dreaming is reported from REM but not nREM sleep.
2. People can judge the duration of their time in REM with accuracy.
3. Dreams are not instant events but are reported/experienced in real time.
4. Eye movements appear to correspond to the content of a dream: this suggests that eye movements are not purely random but are related to dream imagery.
5. Dement and Kleitman’s recall data show only that dreams are remembered more often from REM sleep, not that dreaming is confined to it. This leaves open the possibility that non-REM sleep also carries mental activity that is simply harder to recall.

Strengths

Objective quantitative data

One strength of the study was the use of quantitative data. Researchers calculated the percentages of dream recall by participants following awakenings from rapid-eye movement (REM) and non-rapid-eye movement (nREM) stages of sleep.

Similarly, they calculated the percentage of accuracy of their dream duration estimates when awoken from either 5-minute or 15-minutes of sleep.

This helped researchers objectively compare in which stage of sleep participants reported a greater recall of dreams.

Likewise, they objectively compared when their accuracy of dream duration estimates between the two durations of awakenings was better, without any bias. This increased the internal validity of findings.

High level of control

Researchers controlled for confounding variables by:

1. Prohibiting alcohol and caffeine before the experiment
2. Using a consistent doorbell sound to wake participants
3. Conducting the study in the same quiet, dark room throughout
4. Tying the electrode wires into a single cord so they could not tangle as participants slept

These controls, together with the objective EEG and EOG readings, increased the study’s internal validity.

The study used electroencephalography (EEG) to measure brain activity, allowing precise, repeatable measurements.

Replicable

Because the procedure, apparatus, and waking stimulus were all standardised (see High Level of Control above), other researchers can repeat the study. The EOG also gave an objective, bias-free measure of eye movements, which further supports its reliability.

High external validity: Inclusion of waking validation study to support findings

To enhance the external validity of the findings on eye movement patterns during REM sleep, researchers conducted an additional study.

They asked 20 naive participants and 5 experimental subjects to observe distant and close-up activity while awake, recording their eye movements: horizontal, vertical, or mixed.

The patterns observed in this waking state were comparable to those during sleep, strengthening the study’s conclusions.

Weaknesses

Low generalisability

The sample was very small: only nine participants took part, and just five of these were studied intensively.

You could argue that physiological processes are likely to be the same in all people but this may not be the case. There was no diversity in age, ethnicity and sex.

Their specific sleep patterns, such as how long and how often each stage lasted, may not match the wider population’s.

This reduces the population validity of the findings.

Small number of observations for each eye movement pattern

The study reported only 3 instances of vertical eye movements, 1 instance of horizontal movements, 21 instances of mixed movements, and 10 instances of little or no movement.

This small sample size for each pattern limits the generalizability of the findings.

It increases the possibility that the observed correlations between eye movements and dream content could be due to chance or individual differences.

A larger number of observations for each pattern would provide more robust and reliable results.

Low ecological validity

The research studied participants who went to sleep in a laboratory with electrodes stuck to their heads. It is unlikely that this bears much relation to sleep in a normal environment!

It is possible that being in such an artificial condition meant that their sleep was disturbed, and if this was the case, the researchers would not have been studying normal sleep patterns.

The participants were also woken up several times during the night and asked about their dreams. Again, this is unlikely to happen normally and may have had an effect on the way the participants slept.

Research outside the lab, however, could not measure brain activity and eye movements with this precision.

Difficulty in precisely matching eye movements to specific dream content

The study relied on broad categories of eye movements (vertical, horizontal, mixed, little/none) rather than detailed, moment-by-moment analysis.

Dreams can involve complex and rapidly changing scenes, making it challenging to definitively link specific eye movements to particular dream elements.

The temporal resolution of awakening (after 1 minute of a pattern) may not capture the full complexity of eye movements during a dream sequence.

Individual differences in how people visually experience or recall dreams could affect the accuracy of matching eye movements to content.

The study doesn’t account for potential symbolic or abstract visual experiences in dreams that might not directly correspond to eye movements.

Demand Characteristics: Potential for retrospective bias in dream reports

Participants were awakened and asked to report their dreams after the fact, which can lead to retrospective bias.

Memory of dreams can fade quickly upon waking, potentially leading to incomplete or inaccurate recall.

Participants might unintentionally fill in gaps in their memory or alter details based on expectations or suggestions.

The act of waking up and reporting could itself influence the recalled content of the dream.

There’s a possibility of demand characteristics, where participants might report what they think the researchers want to hear.

Was the study useful?

Later research complicates this neat picture. Foulkes (1985) found that dream-like mental activity is also reported from non-REM sleep, especially at sleep onset, so recall alone may hide non-REM dreaming.

Neuropsychological studies go further. Solms (1997, 2000) showed dreaming can stop after forebrain damage even when REM sleep continues, and can continue when REM-generating brainstem regions are lost. REM and dreaming, it turns out, are controlled by different brain mechanisms.

Many other researchers have replicated Dement and Kleitman’s conclusions. However, one methodological issue should be considered.

Regarding the first research question, Dement and Kleitman conclude that dreaming occurs in REM rather than non-REM sleep.

What they actually demonstrated is that dreams are recalled more often from REM sleep, not that dreaming only happens there. Non-REM sleep may still carry dreaming that is simply harder to recall.

Almost nothing was known then. The link between eye movements and dreaming was pure guesswork before this study.

Dement and Kleitman’s work added real new knowledge, and the EEG itself was still a fairly new tool at the time. Nearly seventy years on, it is easy to underestimate what a breakthrough this represented: it was this kind of Core Study that first showed dreams could be studied objectively.

Dement and Kleitman’s research generated very many other studies into sleep and dreaming and there have been many useful findings.

Issues and Debates

The application of psychology to everyday life

The EEG can detect REM/NREM sleep stages, which can be useful in diagnosing sleep-related problems. This information can then be used to develop appropriate treatments for sleep disorders.

For people with sleep disorders, the EEG’s ability to detect REM/NREM sleep is particularly valuable. A psychologist can analyze the EEG output to compare the patient’s sleep patterns with those of a ‘normal’ sleeper, helping to identify specific abnormalities in the sleep cycle.

The study also opened up research into why REM matters. Waking people every time they entered REM caused them to attempt it more and more often.

Once left undisturbed, they showed a rebound of extra REM sleep (Dement, 1960). This REM-deprivation effect became one of the strongest pieces of evidence that REM meets a specific physiological need.

Nature versus nurture

The study leans towards a biological explanation of dreaming, emphasizing the role of natural physiological processes (REM sleep, eye movements).

However, it doesn’t exclude the possibility that dream content could be influenced by learned or cultural factors.

Reductionism versus holism

The approach is somewhat reductionist, breaking down the complex experience of dreaming into measurable physiological components (eye movements, EEG patterns).

However, by relating these components to subjective dream reports, it also attempts a more holistic understanding of the dreaming experience.

Determinism versus free-will

The strong correlation between eye movements and dream content suggests a degree of physiological determinism in the dreaming process.

However, the variability in dream content and recall also leaves room for individual differences and possibly some element of “choice” in dream experiences.

Individual and situational explanations

The study considers both individual differences (e.g., variation in dream recall ability) and situational factors (e.g., time of night, duration of REM sleep).

It suggests that dream experiences are influenced by both individual physiology and the specific sleep situation.

Idiographic versus nomothetic

The study combines nomothetic elements (looking for general patterns across participants) with idiographic aspects (detailed analysis of individual dream reports).

This balance allows for both general conclusions about dreaming and recognition of individual variations.

Past Paper Question (Cambridge Exam)

Study Summary Sheet