Dreams & Sleep: While You Sleep, Your Brain Does Its Most Important Work

On "why dreams exist" there is still no consensus — but it hasn't been Freud's "disguised repressed wishes" for a long time. Three serious contemporary theories are not mutually exclusive: dreams may simultaneously consolidate memory, regulate emotion, and rehearse threat. We know sleep has functions; whether the dream itself has an independent one is still debated.

Most vivid dreams occur in REM (rapid eye movement): brain activity approaches waking, yet the prefrontal cortex (logic, self-monitoring) is relatively offline while the limbic system (emotion) runs hot — which is why dreams feel emotionally intense but logically absurd. The main theories are in the table below: Revonsuo's threat simulation, Cartwright's emotional digestion, and Hoel's (2021) analogy of dreams as "noise" injected into a neural network to prevent overfitting — each with supporting evidence.

Sleep alternates roughly every 90 minutes between deep (slow-wave) sleep and REM. Deep sleep is like moving memories from temporary cache to hard disk (hippocampus → cortex); REM is more like sorting and linking — weaving new memories into your old knowledge web, often producing creative connections ("sleep on it" has a neural basis). A dream is likely the by-product of this offline processing being watched by consciousness.

Lucid dreaming isn't mysticism — it's a lab-verified state of consciousness: mid-REM you suddenly realize "I'm dreaming," and can even partly steer the dream. It offers consciousness research a rare window: a brain that is simultaneously asleep, self-aware, and able to signal the outside world.

Stephen LaBerge (Stanford, 1980s) devised an elegant proof: in a lucid dream the muscles are paralyzed but the eyes can still move. He had subjects send a pre-agreed eye-movement signal the moment they became lucid — and the recorder did capture that signal during REM, proving physiologically that one can hold waking self-awareness inside objective sleep. Imaging shows that the dorsolateral prefrontal cortex (seat of metacognition), normally offline in dreams, is reactivated.

An ordinary dream lacks the "I'm dreaming" knowing precisely because the prefrontal cortex is offline — you take the absurdity at face value. Lucid dreaming is metacognition relit during sleep: you can experience the dream and know it's a dream at once. This is the same capacity as waking mindful awareness (observing thoughts without being swept up) — Tibetan Buddhism's "dream yoga" trains exactly this.

Healthy REM is like a nightly "overnight emotional therapy": it keeps the facts of a painful memory while dialing down its emotional sting. Trauma (PTSD) interrupts exactly this — the memory replays at original intensity in nightmares, hurting more each time. That is overnight therapy failing.

The "sleep as overnight therapy" framework of van der Helm and Matthew Walker proposes that REM is the one brain state in which noradrenaline (the stress chemical) drops nearly to zero, letting emotional memory be reprocessed safely in "chemical calm." In PTSD, REM is repeatedly broken by nightmares and the memory fails to "desensitize." For trauma nightmares, Imagery Rehearsal Therapy (IRT, Barry Krakow) is an evidence-based first line: while awake, rewrite the nightmare's script and rehearse the new ending — significantly reducing nightmare frequency.

Normally, after several nights of REM processing, you can recall the event without being flooded by the same intensity of emotion ("remember it, but it no longer hurts as much"). Trauma breaks the loop: the nightmare jolts you awake at peak emotion, processing stops unfinished, and it restarts the next night. IRT doesn't erase the memory — while awake, it hands the brain a replaceable new script, clearing space for overnight processing. This is another corner of the same neural puzzle as Day 8 ("trauma is stored in the body") and Day 44 ("memory can be reconstructed").

"Sleep badly, feel worse" isn't weakness — it's a neural mechanism. Sleep deprivation makes the emotional brain (amygdala) overreact while cutting the braking action of the rational brain (prefrontal cortex). The result: you respond harder to negative cues and recover more slowly — reason isn't unwilling, it's physically disconnected.

Yoo, Walker et al.'s (2007) classic fMRI experiment: after one sleepless night, the amygdala's response to negative images amplified by roughly 60%, and its functional connection with the regulating medial prefrontal cortex weakened markedly — the emotional accelerator got touchier while the brake loosened. Later work repeatedly shows poor sleep (especially REM) is both a risk factor and a maintaining factor for anxiety and depression, and the link is bidirectional (bad mood also wrecks sleep, forming a loop).

Sleep, especially REM, nightly maintains the prefrontal-over-amygdala braking circuit. When you're short on sleep the circuit loosens and you revert to a "threat-first" mode where small things balloon into crises. More subtly, sleep deprivation also dampens positive emotion — not only "more easily irritated" but "harder to feel joy."