Why do we dream
Short answer
Moderate evidenceDreams are most likely the conscious experience of the brain consolidating emotional and procedural memories during REM sleep, with a possible additional role in threat rehearsal. The evidence is strong for the memory function and weaker (but real) for the emotional and threat-simulation functions.
Key points
- Dreams happen in all sleep stages but are most vivid and narrative in REM sleep.
- REM dreaming is reliably associated with consolidation of emotional and procedural memories — both well-replicated in lab studies.
- Antonio Zadra and Robert Stickgold's NEXTUP theory: REM lets the brain explore weak associations between memories, which may aid creative problem-solving.
- Threat simulation theory (Revonsuo) proposes dreams rehearse threat scenarios. Evidence is suggestive but not conclusive.
- There's no scientific evidence for the symbolic interpretation of dreams (Freudian or Jungian). Dream content reflects recent waking experience and emotional state, not hidden meaning.
What dreams actually are
Dreams are experiences that happen during sleep — sequences of images, narratives, emotions, and sensory content. They occur in all stages of sleep but are most vivid, narrative, and emotionally charged during REM. Reports from N3 (deep sleep) are typically fragmented and abstract.
Everyone dreams every night, multiple times. Whether you remember depends on whether you woke during or near a REM period.
The memory consolidation hypothesis
The most well-evidenced theory is that REM-stage dreaming is the conscious experience of the brain processing and consolidating recent emotional and procedural memories.
Two specific findings support this:
- Targeted memory reactivation studies show that replaying cues associated with learned material during sleep improves recall — and the effect is strongest during REM.
- Selective REM deprivation(waking participants every time they enter REM) impairs the next day's performance on procedural and emotional tasks specifically.
The NEXTUP theory
Antonio Zadra and Robert Stickgold (2021) propose that REM dreaming functions to discover weak associative connectionsbetween memories — connections too tenuous for the waking brain to make. The bizarre, non-sequitur quality of dreams is the feature, not a bug: dreams are the brain exploring “what if” adjacent possibilities.
This dovetails with the well-documented but informal observation that sleep aids creative problem-solving. The brain finds connections it can't find awake.
Threat simulation theory
Antti Revonsuo proposed that dreams evolved to rehearse threat scenarios in a safe environment, which is why anxiety dreams, chase dreams, and fight dreams are so common across cultures.
The evidence is suggestive but not strong. Cross-cultural dream studies do find a disproportionate representation of threat content. People in genuinely threatening environments (war zones) dream more about threats. But there's no clean experimental test that would distinguish threat simulation from a general bias toward emotionally salient content.
What dreams are not
Dream content reflects recent waking experience, current emotional state, and physiological state during the dream (e.g. needing the toilet, feeling cold). It does not encode hidden symbolic meaning in any way the scientific literature has been able to validate.
Freudian and Jungian dream interpretation traditions are cultural artifacts, not evidence-based practice. This doesn't mean dreams are meaningless — but their meaning is mostly what they obviously are: your brain processing what happened.
Nightmares and bad dreams
Occasional nightmares are normal. Chronic nightmares (more than once a week, causing distress) may indicate PTSD, anxiety disorders, certain medications (especially beta-blockers and SSRIs), or sleep apnoea. Image Rehearsal Therapy is the evidence-based treatment — patients rehearse rewritten endings to recurring nightmares, and frequency typically drops within weeks.