A Novel Approach to Dream Content Analysis Reveals Links Between Learning-Related Dream Incorporation and Cognitive Abilities

The extent to which newly learned material is incorporated into dreams is related to individual differences in cognitive abilities. Using a computational linguistic approach to measure semantic similarity between waking experiences and dream reports, the study found that the richness of incorporation of a spatial or motor memory task into early hypnagogic dreams correlated with reasoning abilities (r=0.41), while incorporation into later dreams correlated with verbal abilities (r=0.46). Memory performance improvements were related to dream incorporation (r=0.47) for early but not late dreams, but this relationship was not mediated by cognitive abilities, suggesting a direct link between reasoning and dream incorporation.

Study at a glance

Abstract

Can dreams reveal insight into our cognitive abilities and aptitudes (i.e., “human intelligence”)? The relationship between dream production and trait-like cognitive abilities is the foundation of several long-standing theories on the neurocognitive and cognitive-psychological basis of dreaming. However, direct experimental evidence is sparse and remains contentious. On the other hand, recent research has provided compelling evidence demonstrating a link between dream content and new learning, suggesting that dreams reflect memory processing during sleep. It remains to be investigated whether the extent of learning-related dream incorporation (i.e., the semantic similarity between waking experiences and dream content) is related to inter-individual differences in cognitive abilities. The relationship between pre-post sleep memory performance improvements and learning-related dream incorporation was investigated (N=24) to determine if this relationship could be explained by inter-individual differences in intellectual abilities (e.g., Reasoning, Short Term Memory and Verbal abilities). The extent of dream incorporation using a novel and objective method of dream content analysis, employing a computational linguistic approach was used to measure the semantic relatedness between verbal reports describing the experience on a spatial (e.g., maze navigation) or a motor memory task (e.g., tennis simulator) with subsequent hypnagogic reverie dream reports and waking “daydream” reports, obtained during a daytime nap opportunity. Consistent with previous studies, the extent to which something new was learned was related (r=0.47) to how richly these novel experiences were incorporated into the content of dreams. This was significant for early (the first 4 dream reports) but not late dreams (the last 4 dream reports). Notably, here, we show for the first time that the extent of this incorporation for early dreams was related (r=0.41) to inter-individual differences in Reasoning abilities. On the other hand, late dream incorporation was related (r=0.46) to inter-individual differences in verbal abilities. There was no relationship between performance improvements and intellectual abilities, and thus, inter-individual differences in cognitive abilities did not mediate the relationship between performance improvements and dream incorporation; suggesting a direct relationship between Reasoning abilities and dream incorporation. This study provides the first evidence that learning-related dream production is related to inter-individual differences in cognitive abilities.