Does Dream Content Predict Cognitive Abilities?

A small sample study conducted on Stuart Fogel’s student population, who is a cognitive neuroscientist and sleep researcher at the Royal Hospital Institute of Mental Health Research in Ottawa, Canada and director of the University of Ottawa Sleep Research Laboratory.

by Eric M. Fortier, B.A. | 2017

Some evolutionary theorists propose that dreaming serves an adaptive function in preparing us for danger in waking life. Others, such as Freud and Jung, propose a compensatory view, in which dreams function as an outlet, or theater, for the expression of deep, sometimes unconscious, or under-satiated wished and needs. The most empirically-supported hypothesis is the continuity theory, in which dream content is found to reflect the content of everyday waking life (Blagrove & Pace-Schott, 2010). Recently, research suggests a combination of these might hold truth, and that REM dreams reflect the use of existing experience to solve cognitively challenging learning tasks requiring innovative solutions (Smith, 2012).

Dream frequency, length and intensity are related to verbal and general mental ability. Certain cognitive abilities such as reasoning (but not verbal ability) have been linked to other features of sleep. Yet little research has examined the relation between cognitive abilities and dream content (Blagrove & Pace-Schott, 2010; Peigneux, Fogel, & Smith, 2011).

We propose the use of Hall & Van Der Castle’s dream scoring tool to analyze dream content, in combination with the Cambridge Brain Sciences cognitive test battery, for which the neuropsychological correlates are known, to identify the features of dreaming most related to cognitive abilities. Based on the idea that features of sleep are related to cognitive abilities, but not verbal ability, and that REM is beneficial to learning complex cognitive tasks, we hypothesize that reasoning and working memory would be positively related to coded good fortune and success, and negatively related to misfortune.

In total, data for both cognitive tests and dream reports from 60 participants was collected from a predominantly female population of Canadian psychology students. Participants were recruited via Dr. Stuart Fogel’s student population and asked to begin recording their dreams in a dream journal by their beds each morning. During the semester, they were then asked to set time aside to complete the Cambridge Brain Sciences cognitive test battery, consisting of twelve computerized cognitive tasks designed to assess aspects of reasoning, working memory, and verbal ability. Participants were then trained how to score their own dreams according to the Hall & Van De Castle’s content analysis tool, and asked to select a dream no longer than five sentences long, analyze its content, and enter the scores for each content category into a standardized spreadsheet template known as DreamSAT (an automated dream data entry system and statistical analysis tool). Scores were aggregated, and bivariate two-tailed correlations were then run between each cognitive testing scores and dream content scores for each participant using SPSS.

Two-tailed bivariate correlations revealed a statistically significant role for working memory and reasoning ability, but not verbal ability, in predicting features of dream content. Most notably, good fortune was moderately positively correlated with reasoning (r = .43, p = .001) and working memory (r = .32, p = .014). Reasoning ability also showed a moderate positive correlation with success (r = .30, p = .021), and importantly, a moderate negative correlation with misfortune (r = -.31, p = .12). Aside from being positively related to good fortune, working memory showed a highly statistically significant moderate negative correlation with friendliness (r = -.34, p = .007) and a weaker though less statistically significant negative correlation with coded emotion (r = -.26, p = .046). No significant correlations were found between verbal ability and categories of dream content. All cognitive capacities were highly correlated with one another. The relationship between scores for reasoning and good fortune can be seen in a scatterplot below (Figure 1), followed by full results in Table 1.

Figure 1. Moderate positive relationship between Reasoning and Good Fortune

Table 1
Bivariate Correlations Between Dream Content and Three Cognitive Measures

Measure Reasoning Working memory Verbal
CHAR .031 .040 .082
AGGRESSION .167 .004 .148
FRIENDLINESS -.084 -.343** -.169
SEX .060 -.027 -.063
ACTIVITIES .036 -.041 .044
SUCC .297* -.083 .152
FAIL -.170 -.020 .048
MISFORTUNE -.322* -.076 -.087
G-FORT .427** .315* .217
CODED_EMOT -.219 -.258* .033
DREAMER_EMOT .063 -.057 .206
SET -.120 -.144 -.228
OBJ .149 -.003 .051

**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).

Most notably, reasoning and working memory both predicted good fortune, with reasoning also related to greater success and less misfortune. This may be interpreted a number of ways. In line with the continuity hypothesis, it may could be interpreted to mean that good fortune and success represent a greater portion of daily life for those with higher reasoning and working memory ability.

Another interpretation returns to the idea that REM sleep functions to enhance learning on cognitively complex tasks requiring novel strategies. As seen in the famous mirror drawing task and the subsequent driving imagery in dream content, the act of “crashing” in the dream may be attributed to misfortune, and successful driving may be attributed to good fortune. In this sense, those with greater cognitive ability and would see a lower rate of crashes (misfortune) and higher rate of good fortune, for example, by running into less obstacles, or as a result of more often finding a successful strategy. The successful solving of the conceived learning task, according to the metaphorical function of dreams, may simply be perceived as a lucky result, (especially considering the dreamer has no conscious control over external dream events except in rare cases) and therefore coded as good fortune. These results appear to support the initial hypothesis that cognitive capacities would be related to dream content.

Limitations and Future Directions
There was limited time to train participants on how to code using the Hall and Van de Castle method, leaving many somewhat misinformed and confused. Future experiments should consider using independent dream coders who are properly trained on the Hall and Van de Castle method. It may be important to remove participation incentives to increase participant integrity and overall engagement, as some anonymous users admitted in candid discussion that they had invented their dream reports and falsified their cognitive tests in order to obtain the incentive in a shorter period of time. Finally, neuroimaging techniques could reveal a great deal about the neuropsychological relationship between post-learning REM sleep and structured dream content, and help develop a richer understanding of which neurophysiological functions of REM sleep might enhance complex learning.

Enter your email to receive an update when something new is posted.


Blagrove, M., & Pace-Schott, E. F. (2010). Trait and Neurobiological Correlates Of Individual Differences In Dream Recall And Dream Content. International Review of Neurobiology, 92, 155-180. doi:10.1016/s0074-7742(10)92008-4

Smith, C. (2012). Sleep States, Memory Processing, and Dreams. Sleep Medicine Clinics, 7(3), 455-467. doi:1.1016/j.jsmc.2012.06.008

Peigneux, P., Fogel, S., & Smith, C. (2011). Memory Processing in Relation to Sleep. Principles and Practice of Sleep Medicine (6th ed), 335-347. doi:10.1016/b978-1-4160-6645-3.00029-3