The cognitive theory of dreaming holds that dreams are produced by the same mental systems you use while awake. Rather than treating dreams as random neural noise or disguised wishes, this framework views them as a form of thinking that happens during sleep, shaped by your knowledge, memories, and concerns. The theory was first proposed by Calvin Hall in the 1950s and later expanded by David Foulkes and other researchers who approached dreaming as a cognitive process that could be studied the same way psychologists study perception, memory, and language.
Core Ideas Behind the Theory
The cognitive theory rests on several connected assumptions. First, dreaming draws on the same mental machinery as waking thought. Your brain doesn’t switch to some exotic mode when you fall asleep; it continues processing information, just under different conditions. Second, because dreaming relies on cognitive systems, it has prerequisites. Not every sleeping brain dreams in the same way. The capacity for rich, narrative dreaming develops over childhood alongside other cognitive abilities, and it can be disrupted by specific types of brain damage.
Hall originally argued that dream content reflects the dreamer’s “concepts,” their understanding of themselves and the world around them. If you spend your days worried about a relationship, those concerns will surface in your dreams, not as coded symbols hiding forbidden desires, but as relatively straightforward expressions of what’s on your mind. Dream symbolism, in this view, is expressive rather than defensive. A dream about falling doesn’t need to be decoded through a psychoanalytic lens; it more likely reflects a feeling of losing control that you already recognize in waking life.
The Continuity Hypothesis
One of the most influential ideas to come out of the cognitive approach is the continuity hypothesis: the idea that dream content is continuous with waking experience. If you spend a lot of time doing something, thinking about something, or feeling a certain way during the day, those activities and emotions tend to show up in your dreams.
Empirical research broadly supports this. Studies using correlational designs have found that individual differences in how people spend their waking hours are reflected in their dream content. A striking recent example involved analyzing Reddit posts around the start of the Russo-Ukrainian war. When researchers filtered those posts to include only dream reports, topic modeling revealed an increase in war-related dream content, a real-time demonstration of waking concerns bleeding into dreams at a population level.
That said, the relationship isn’t a simple replay. Dreams don’t typically recreate your day like a recording. They pull fragments of experience and recombine them, sometimes in bizarre ways. And not all waking activities show up equally. Some experiences are incorporated more readily than others, and researchers have noted that the continuity hypothesis in its broad, general form needs more specific formulation to account for these differences.
How Dreams Develop in Children
One of the strongest arguments for the cognitive theory comes from developmental research. David Foulkes conducted landmark longitudinal studies following children from ages two through adolescence, waking them during REM sleep and collecting dream reports. What he found challenged the assumption that all sleeping people dream the same way.
Very young children’s dream reports were brief, static, and relatively simple, often just an image or a short scene. As children developed more sophisticated cognitive abilities (spatial reasoning, narrative skill, a stable sense of self), their dreams became longer, more complex, and more storylike. This pattern mirrors what you’d expect if dreaming were a cognitive skill that matures alongside other mental abilities, not something that arrives fully formed at birth. If dreams were simply random activations, there would be no reason for their complexity to track so closely with cognitive development.
Dreams and Memory Processing
Modern cognitive science has added a layer to the theory by connecting dreaming to memory consolidation. Sleep doesn’t just passively store memories. It transforms them, extracting general patterns and integrating new information into what you already know.
One influential model proposes that during sleep, recently encoded information gets “interleaved” into related memory networks. The brain simultaneously reactivates recent and older memory fragments, gradually associating new content into existing knowledge structures. This process of pulling together fragments from different times and contexts may explain why dreams so often feature bizarre combinations of people, places, and events from different parts of your life. That strangeness isn’t a malfunction. It may be exactly the kind of associative process that helps your brain find connections between experiences and build broader understanding from individual episodes.
Evidence for Problem Solving in Dreams
The cognitive theory also predicts that dreams should be capable of something useful, and recent experimental work supports this. In a study published in Neuroscience of Consciousness, researchers gave participants unsolved puzzles and then played associated sound cues during REM sleep. The sound cues reliably increased dreaming about those specific puzzles. More importantly, when a puzzle was actually incorporated into a participant’s dream, that puzzle was significantly more likely to be solved the next morning.
The effect was specific. Participants who dreamed about the cued puzzles solved them at higher rates than uncued puzzles (with an odds ratio of 0.32, meaning cued puzzles were roughly three times more likely to be solved). Participants who didn’t incorporate the cues into their dreams saw no benefit. This suggests it’s not just sleeping on a problem that helps. The dream itself, the active cognitive processing during sleep, contributes to finding solutions.
What the Brain Is Doing During Dreams
Neuroimaging research has mapped out a distinctive pattern of brain activity during dreaming that aligns well with the cognitive theory. During REM sleep, areas involved in emotion, visual processing, and memory are highly active. These include limbic structures like the amygdala and hippocampus, visual processing areas in the back of the brain, and medial prefrontal cortex.
At the same time, several regions are notably quiet. The dorsolateral prefrontal cortex, which supports logical reasoning, working memory, and self-monitoring, is deactivated. So are areas involved in spatial orientation and self-reflection, like the posterior cingulate cortex and inferior parietal cortex. This explains a lot about what dreams feel like. You can have vivid emotional experiences and complex visual scenes, but you rarely stop to question whether what’s happening makes sense. The thinking machinery is running, but the editor is off duty.
This pattern also explains why dreams are so hard to remember. The combination of reduced prefrontal activity and changes in brain chemistry during REM sleep creates conditions that work against forming lasting memories. Unless you wake up during or immediately after a dream, the experience typically fades.
Metacognition and Lucid Dreaming
Lucid dreaming, the state of becoming aware that you’re dreaming while still asleep, offers a useful test case for cognitive theories. In ordinary dreams, metacognition (the ability to reflect on your own mental state) is largely absent. You accept impossible events without question. In lucid dreams, that reflective capacity comes back online.
Brain imaging shows that lucid dreaming involves a “hybrid” state. Standard markers of REM sleep are maintained, but frontal brain regions associated with self-awareness show increased activity. People who lucid dream frequently have greater gray matter volume in frontopolar cortex, the same region that activates during metacognitive tasks while awake. This finding reinforces the cognitive theory’s central claim: dreaming and waking thought share neural infrastructure. When more of that infrastructure is active during sleep, the dream experience becomes more like waking consciousness.
How Cognitive Theory Differs From Other Models
The cognitive approach sits between two older theories. Freud’s psychoanalytic model treated dreams as disguised expressions of unconscious wishes, requiring interpretation to decode hidden meaning. The activation-synthesis hypothesis, proposed by Hobson and McCarley in the 1970s, swung to the opposite extreme, arguing that dreams are the brain’s attempt to make sense of essentially random neural firing during sleep.
The cognitive theory takes a middle path. Dreams aren’t meaningless noise, but they aren’t encrypted messages either. They’re the natural output of a brain that keeps processing information during sleep, drawing on memory, emotion, and learned patterns of thought. The content of your dreams tells you something real about your concerns and mental life, not because it’s symbolically coded, but because your dreaming mind is working with the same material your waking mind uses.
Modern Tools and the Future of Dream Science
Recent advances have given researchers tools that earlier cognitive dream theorists could only imagine. Neural decoding using EEG can now identify specific content features in dreams. For instance, whether a dream contains faces or locations can be predicted from patterns of brain oscillation in particular regions, and the level of anger in a dream report corresponds to measurable frontal brain activity. These findings provide direct evidence that dream content maps onto the same neural processes as waking perception and emotion.
Targeted memory reactivation, the technique of playing sounds or cues during sleep, has opened the door to experimentally manipulating dream content. In one clinical application, people with nightmare disorder mentally rehearsed a new ending for a recurring nightmare while listening to a piano chord. When that chord was played again during their sleep, participants reported fewer nightmares and more positive dreams. Computational analysis of large dream-report databases, including mining online platforms for thousands of reports, has made it possible to identify patterns in dream content across populations and in response to real-world events. Together, these methods are turning dreaming from something that could only be studied through subjective morning reports into a phenomenon that can be observed, measured, and experimentally tested.