Written by Dr. Sarah Mitchell, PhD, sleep researcher at the Stanford Sleep Research Center, this article explores the fascinating developmental neuroscience behind children's dream life — why infants spend half their sleep in REM, why the most terrifying nightmares cluster at age four, and why a seven-year-old's dreams are cognitively different from a three-year-old's, even though both children sleep more than adults.
The REM-Rich World of Early Life
Before any discussion of dream content, it is worth establishing the extraordinary fact of infant sleep architecture. A premature infant born at 28 weeks gestational age spends approximately 80% of total sleep time in a state researchers call active sleep — the developmental precursor of REM sleep, characterised by rapid eye movements, irregular breathing, twitching of limbs and facial muscles, and the motor inhibition that in adults prevents sleepers from acting out their dreams. By full-term birth, this proportion has fallen to around 50%. Through the toddler and preschool years it gradually declines to 35-40%. By adolescence it approaches adult levels of 20-25%.
This trajectory — from maximally REM-rich in the most immature nervous systems to proportionally REM-reduced in mature ones — is not coincidental. Jerome Siegel at UCLA and Marcos Frank at the University of Pennsylvania have argued, supported by substantial animal research, that REM sleep in early development serves primarily as an endogenously generated stimulator of brain maturation. During REM, the brainstem generates bursts of neural activity — pontine-geniculate- occipital (PGO) waves in animals, and their human equivalents — that propagate through the developing sensory cortices, driving the formation and refinement of synaptic connections. In the womb and in early infancy, when external sensory experience is limited, this internally generated activation may substitute for experience in shaping neural architecture.
Matthew Walker at Berkeley describes this process as the brain "self-programming" during REM — using its own activity to wire the circuits that will later process vision, language, and social information. On this view, the extraordinary REM richness of infant sleep is not merely a curiosity of developmental neuroscience but a functional imperative: the developing brain needs massive amounts of REM precisely because it is building itself. The gradual REM reduction through childhood tracks, roughly, the progressive completion of cortical development.
David Foulkes and the Cognitive Development of Dreaming
The most important researcher in the study of children's dreams is David Foulkes, a psychologist who spent decades conducting laboratory-based systematic studies of dream development across childhood. His methodology — awakening children from documented REM sleep in a controlled laboratory setting and recording their dream reports verbatim, across multiple nights, across multiple years — was far more rigorous than the parent-report and retrospective interview methods that dominated earlier research.
Foulkes's central and initially surprising finding: children do not dream the way adults do. When children between the ages of three and five are awakened from REM sleep — the sleep stage most strongly associated with vivid narrative dreaming in adults — they rarely report narrative experiences. What they describe, when they report anything at all (which is far less frequent than adults), is closer to static imagery: a single animal, a person, an object, a colour. There is no story. There is rarely a self as a character within the scene. The first-person experiential dream — in which the dreamer moves through a world and experiences events — does not reliably emerge in Foulkes's data until approximately ages five to seven.
Between ages seven and nine, dream reports begin to approach adult complexity: characters, setting, narrative sequence, emotional experience, and a self-as-protagonist who acts and is acted upon all appear with increasing regularity. Foulkes connected this developmental trajectory explicitly to the broader cognitive development of spatial cognition and self-representation — abilities that develop through the same age range according to Piagetian and post-Piagetian developmental frameworks. For Foulkes, dreaming is not an innate capacity triggered by REM sleep but a cognitive achievement, built upon the same mental architecture that supports waking imagination and self-concept.
Why Nightmares Peak Between Ages Three and Six
If Foulkes's research shows that narrative dreaming barely exists before age five, why do parents of three- and four-year-olds frequently report their children waking in terror from apparent nightmares? The answer likely involves a combination of factors that do not require full narrative dreaming to produce distressing sleep experiences.
First, partial arousal disorders — including night terrors (pavor nocturnus), confusional arousals, and sleepwalking — peak in early childhood and are distinct from REM nightmares. Night terrors occur during slow-wave sleep, not REM, and produce screaming, agitation, and apparent terror without subsequent dream recall because they do not involve the cortical narrative processing of REM. Parents experiencing their child's night terror — eyes open, screaming, unable to be comforted, with no memory of the episode the following morning — are witnessing a slow-wave arousal phenomenon, not a nightmare in the clinical sense.
True REM nightmares, with their associated dream recall, do cluster in the three-to-six age range, however — with surveys finding 50-75% prevalence of frequent nightmares in this population. This clustering coincides with the emergence of true narrative dreaming (which provides the narrative vehicle for nightmare experience), with the developmental period of maximum imaginative engagement with fears and monsters (the classic 'under-the-bed monster' phase), and with relative prefrontal immaturity. The prefrontal cortex — which provides executive regulation of the amygdala's alarm responses in adults — is profoundly underdeveloped through early childhood, meaning children have limited internal regulatory resources for contextualising frightening dream imagery. When the dreaming preschooler's amygdala fires in response to threatening content, there is little prefrontal dampening available. See also our article on the causes and meaning of nightmares for the adult parallel.
Imagination, Reality, and the Blurry Line in Early Childhood
A crucial factor in children's vivid and sometimes terrifying dream experience is the developmental reality boundary — the cognitive distinction between what is real and what is imagined. Adults have a robust, automatic capacity to locate experiences as real or imagined, past or present. This capacity is built over childhood and is incomplete until middle childhood (approximately seven to nine years).
For a four-year-old, the monster that appeared in last night's dream may retain a quality of felt reality that a six-year-old's would not. The sleeping brain's default-mode-dominant, uncritical processing does not yet have a robust 'this was a dream' tagging system. This is why young children frequently report dream events as having happened — 'Grandma came and took away my toys last night' — in a way that parents recognise as confusing dream and memory.
Carl Jung would have recognised this boundary permeability as characteristic of what he called the 'participation mystique' of early consciousness — the mode of experience in which inner and outer, imagined and real, self and world are not yet clearly differentiated. The developmental achievement of reality testing is precisely the progressive construction of this boundary, and until it is in place, the dream world has a claim on waking reality that adults have long since revoked.
Brain Development and Dream Vividness
The extraordinary vividness that children report in their dreams — when they begin reporting dreams at age five and above — has a neurological correlate in the relative activity levels of different brain systems during REM sleep. In adults, REM sleep is characterised by high activity in the limbic system (particularly the amygdala and anterior cingulate cortex), the visual cortex, and the motor cortex, combined with relative suppression of the dorsolateral prefrontal cortex. This pattern produces vivid, emotionally intense imagery with impaired critical faculties — the classic dream state.
In children, the prefrontal cortex is even less available than in adults — both because it is structurally immature and because it is proportionally less dominant in the overall network architecture of the young brain. This means that the limbic and sensory systems that generate dream imagery operate with even less top-down regulation in children than in adults, potentially explaining why children's dreams, when they occur, can be experienced with an intensity that adults rarely match. The causes of vivid dreams in adults typically involve factors that move adult brain states toward the developmental baseline that children occupy naturally.
Parenting Strategies for Childhood Nightmares
Evidence-based guidance for parents managing children's nightmare frequency and impact draws on developmental psychology, attachment theory, and cognitive behavioural approaches. The following strategies have support in the research literature:
Respond promptly and calmly.When a child wakes frightened, rapid caregiver presence is physiologically calming: the social safety system (mediated through the vagus nerve and oxytocin) actively downregulates cortisol and amygdala arousal. Do not leave a distressed child to 'self-soothe' from a nightmare; the emotional system that is activated needs co-regulation before it can achieve self-regulation.
Validate and name the emotion."That sounds really scary. It makes sense you're frightened." Emotional validation reduces the secondary distress of feeling alone in the experience and engages the social-verbal cortex in a way that itself moderates limbic arousal.
Use daytime 'mastery' techniques.During the day — never in the hours before bed — invite the child to draw the nightmare creature, give it a funny name, change the ending of the dream story, or imagine the monster becoming friendly. This uses the same imaginative plasticity that generated the nightmare to rewrite it, restoring the child's sense of agency. Deirdre Barrett at Harvard has documented the effectiveness of this approach with older children capable of deliberate imagery.
Assess for recurring themes. Occasional nightmares are developmentally normal. Frequent nightmares with recurring themes — especially themes of abandonment, pursuit by specific figures, or scenes that mirror real events — warrant assessment for anxiety, exposure to frightening content, or potential trauma. A simple dream record kept by the parent can help identify patterns for discussion with a paediatrician or child psychologist.
For parents seeking a comprehensive and research-grounded guide to sleep development from birth through adolescence, Healthy Sleep Habits, Happy Child by Marc Weissbluth remains one of the most authoritative and practically useful resources available.
Frequently Asked Questions
Why do babies and young children spend so much time in REM sleep?
Premature infants spend up to 80% of sleep time in REM; full-term newborns approximately 50%. Jerome Siegel at UCLA and Marcos Frank at the University of Pennsylvania showed that REM in early life serves primarily as an endogenous stimulator of brain development, generating PGO waves that drive the formation and pruning of synaptic connections in the developing visual and sensory cortices. Internally generated neural activation during REM may substitute for external sensory experience in shaping neural architecture, explaining why the most immature nervous systems require the most REM sleep.
At what age do children begin having true narrative dreams?
David Foulkes's laboratory research found that children younger than approximately five years old rarely report narrative dreams from REM sleep — they describe static imagery rather than stories. Between five and seven, simple narratives and occasional self-as-character appear. True adult-style narrative dreaming reliably emerges around age seven to nine, in parallel with the spatial cognition and self-representation skills developing through middle childhood. Foulkes concluded that dreaming is a cognitive achievement, not an innate capacity triggered by REM sleep.
Why are nightmares so common in children aged 3 to 6?
The nightmare peak at ages three to six coincides with the emergence of narrative dreaming, maximum imaginative engagement with fears and monsters, profound prefrontal cortex immaturity (limiting emotion regulation during frightening dreams), and the incomplete development of the reality-imagination boundary that makes dream experiences temporarily feel as real as waking ones. Night terrors — distinct from nightmares and occurring during slow-wave rather than REM sleep — also peak in this age range, often being misidentified as nightmares by parents.
How should parents respond to a child's nightmares?
Research supports prompt, calm caregiver presence — which physiologically downregulates cortisol and amygdala arousal through the social safety system — combined with emotional validation. Avoid both dismissing the experience and amplifying its reality. During the following day, daytime mastery techniques — drawing the nightmare, changing the ending, giving scary elements funny names — use childhood imagination to rewrite the experience and restore agency. Chronic severe nightmares warrant assessment for underlying anxiety, trauma exposure, or sleep-disordered breathing.
Do children dream in colour?
Available evidence suggests children who report dreams do describe colour, consistent with the adult finding that the majority of people dream in colour. Eva Murzyn at the University of Dundee showed that dream colour experience is partly shaped by the visual media environment during formative years — people who grew up with black-and-white television reported more monochrome dreams. Children today growing up with full-colour digital environments from birth have richly chromatic baseline visual experience, which is presumed to be reflected in their dream imagery.