Every night, in a state the brain enters without being asked, we become the authors of experiences we did not choose and cannot fully control. We fly. We sit exams for subjects we never studied. We speak with people who have been dead for years. Dreaming is so ordinary that it can feel like it needs no explanation. But for neuroscientists, it remains one of the most genuinely open questions in the field.

The problem with obvious answers

The most commonly repeated explanations for why we dream fall into two categories: memory consolidation and emotional processing. Sleep, and REM sleep in particular, does appear to help the brain sort and store the experiences of the day. Studies going back decades show that people who sleep after learning something remember it better than people who stay awake. And there is good evidence that REM sleep plays a specific role in processing emotionally loaded experiences, which may be part of why disrupted sleep and mood disorders so often travel together.

But neither explanation fully accounts for the specific phenomenology of dreaming: the narrative quality, the bizarre juxtapositions, the visual vividness, the near-complete loss of self-awareness that allows us to mistake a dream for waking life. Memory consolidation happens during non-REM sleep too. Emotional processing happens during waking life. Neither tells us why any of this needs to feel like something.

Defending the visual cortex

One of the more striking recent theories comes from Stanford neuroscientist David Eagleman and his colleague Don Vaughn. Published in Frontiers in Neuroscience and developed in subsequent years, their Defensive Activation Theory proposes that dreams exist primarily to protect the visual cortex from being taken over by neighboring sensory systems during the hours of darkness.

The brain is radically adaptive. When a brain region goes unused, neighboring regions begin to colonize it through synaptic competition. Research on people wearing blindfolds shows that the visual cortex begins generating touch-related activity within roughly 60 to 90 minutes of sensory deprivation. Blind people who lost their sight in adulthood show cross-modal rewiring within months. The brain does not hold territory it is not using.

Eagleman and Vaughn’s argument is that the visual cortex is especially vulnerable during the long stretch of darkness that constitutes the ancestral night. With no visual input for eight hours, the cortex risks being colonized by adjacent auditory or somatosensory regions. REM sleep, with its bursts of intense visual activity, keeps the territory occupied. Dreams, on this view, are not the point. They are the side effect of the brain running a maintenance program.

“The brain is in a use-it-or-lose-it competition with itself every night. Dreaming may be how the visual system fights back.”

Evidence from across species

The theory makes several testable predictions. One is that animals that are more visually dependent should spend more time in REM sleep. Predators, who rely on vision for hunting, do tend to have more REM than prey animals, which often have eyes positioned for wide peripheral awareness rather than acute forward vision. Newborn humans, whose visual systems are the most immature and therefore the most vulnerable to competitive takeover, spend roughly half their sleep time in REM, far more than adults. As visual plasticity declines with age, REM percentage drops in parallel.

A 2023 study on lucid dreamers added another angle. Researchers found that the frontal regions associated with self-awareness show unusually strong connectivity to visual areas during lucid REM sleep, suggesting that the balance between self-monitoring and visual generation is actively managed during the dream state. When self-awareness comes back online, as in lucid dreaming, something in the visual experience changes too.

What remains unknown

The Defensive Activation Theory is compelling and falsifiable, which is more than can be said for many dream theories. But it is still contested. It does not explain non-visual dream content, the emotional intensity of many dreams, or the fact that blind people who have never had visual experience still enter REM sleep and report dream experiences involving the senses they do use.

What the current generation of research agrees on is that dreaming is not random noise. It is correlated with specific brain states, specific neurochemical environments, and specific life circumstances in ways that point toward function. Whether that function is visual maintenance, memory processing, emotional regulation, threat simulation, or something yet unnamed, the brain goes to considerable metabolic expense to make dreams happen every night. That alone suggests they are not nothing.

Sources

  • Eagleman, D.M., & Vaughn, D.A. (2021). The Defensive Activation Theory: REM sleep as a mechanism to prevent takeover of the visual cortex. Frontiers in Neuroscience.
  • Baird, B., et al. (2023). Prefrontal-visual connectivity in lucid dreaming. Nature Neuroscience.
  • Walker, M. (2017). Why We Sleep. Scribner. (contextual reference)