Octopus Dreams and What They Reveal About the Origins of Consciousness in Animals

In 2021, researchers at the Okinawa Institute of Science and Technology filmed sleeping octopuses and watched something that had no clean explanation: the animals' skin began to pulse with colour. Waves of white, orange, and dark brown rolled across their bodies in rhythmic bursts, lasting anywhere from a few seconds to a minute, before the octopus settled back into stillness. The octopus was not awake. Its eyes were closed. Nothing in its environment had changed.
This is not a quirk of one species. Octopus insularis, studied off the coast of Brazil by researcher Sylvia Medeiros and her team at the Brain Institute of the Federal University of Rio Grande do Norte, showed the same pattern, active skin displays during sleep, cycling roughly every 30 to 90 minutes. The researchers published their findings in iScience in 2021, describing what they called "active sleep" phases that looked strikingly similar to REM sleep in vertebrates.
The biological fact underneath this is startling: octopuses have no centralized brain in the vertebrate sense. About two-thirds of their 500 million neurons are distributed across their eight arms. Each arm can process information and act independently. Yet somehow, during sleep, the whole animal enters a coordinated state of neural activity. Something is being processed. The question is what.REM sleep in mammals is the phase most associated with dreaming, memory consolidation, and emotional processing. It is marked by rapid eye movement, muscle paralysis, and intense brain activity. In humans, REM deprivation over several nights produces cognitive deterioration faster than total sleep deprivation does.
The octopus version is not identical. But the structural parallels are hard to dismiss. During active sleep phases, octopus eyes move beneath their closed lids. Their suckers twitch. Their skin produces complex, rapidly changing chromatic patterns, the same chromatophore system they use when hunting, hiding, or communicating while awake. The brain, or whatever distributed version of it coordinates these animals, is running something that looks like a replay.

Neuroscientist Marcos Frank at Washington State University, who studies sleep function across species, has pointed out that when distantly related animals independently evolve similar sleep architectures, it suggests the function those sleep phases serve is under strong evolutionary pressure. REM-like sleep appears in birds, reptiles, and now cephalopods. These lineages diverged from a common ancestor roughly 550 million years ago. If active sleep evolved separately in each, the pressure to develop it must be enormous, which means whatever it does for cognition, the brain cannot easily do without it.Consciousness research has long been hampered by a definitional problem: we cannot agree on what consciousness is, so we cannot agree on how to measure it. The dominant frameworks, Global Workspace Theory, Integrated Information Theory, were built primarily on human and primate data. Octopuses don't fit those models cleanly, and that is precisely what makes them useful.
The cephalopod nervous system evolved entirely separately from the vertebrate line. An octopus brain shares almost no architectural features with a mammalian cortex. Yet octopuses solve puzzles, recognise individual human faces, use tools, and now appear to dream. If awareness or something functionally equivalent to it can arise in a brain built on a completely different plan, then consciousness is not a product of any specific neural structure. It may be a product of complexity, integration, and, the sleep research suggests, the need to process experience offline.

This has direct implications for how scientists think about animal cognition. For decades, the assumption was that sentience required something like a mammalian cortex. The octopus data pushes that boundary back toward a more basic question: what is the minimum architecture required for a system to have something it is like to be?The honest answer is that nobody has proven octopuses dream. What the research has shown is that they enter sleep states with behavioural and physiological signatures that parallel dreaming sleep in other animals. Whether there is any subjective experience accompanying those states, whether there is anything it is like to be an octopus cycling through chromatic patterns in the dark, remains genuinely unknown.
The philosopher Thomas Nagel asked this question about bats in his 1974 essay "What Is It Like to Be a Bat?" His point was that echolocation is so alien to human experience that we cannot imagine the bat's inner world, even if we accept the bat has one. The octopus problem is harder. Its sensory world, its distributed cognition, its colour displays produced by an animal that is likely colourblind, these are further from human experience than a bat's sonar.

What the research does establish is that the capacity for complex sleep architecture is far older and more widespread than the vertebrate-centric model assumed. That alone demands a revision of where scientists draw the line when thinking about which animals have morally relevant inner lives.
The octopus has been carrying 500 million neurons for roughly 300 million years, solving problems and apparently processing its days in sleep, without any of the neural hardware scientists once considered necessary for awareness. The dreaming skin is not proof of consciousness. It is proof that the question of consciousness is much older, and much stranger, than the human brain that is asking it.