Do Animals Dream? What Rats, Birds, and Octopuses Reveal While They Sleep

Dreaming feels like it should be a human thing, mostly because the only evidence any of us ever has for it is a sentence: "I had the weirdest dream last night." Animals can't file that report, so it's easy to assume nothing much is happening in there. The actual sleep research points somewhere stranger. A lot of animals show every measurable sign of dreaming. We just can't ask them what it's about.

The rat study that started the modern version of this question

In 2001, Kenway Louie and Matthew Wilson at MIT ran rats through a maze while recording the firing patterns of neurons in the hippocampus, the brain region that handles spatial memory. Then they recorded the same neurons while the rats slept.

During REM sleep, the same neurons fired in the same sequence they had during the maze run, stretched out over the same rough timescale. The match was tight enough that the researchers could tell, watching a sleeping rat's brain activity alone, roughly where in the maze it would have been if it were awake. Published in Neuron, this is usually the first citation in any serious discussion of animal dreaming, because it's not a twitch or a guess. It's a replay.

That doesn't prove the rat was experiencing anything like a human dream. What it proves is that the sleeping brain was actively re-running yesterday, in order, using the exact same circuitry it used the first time. Whatever you want to call that, it isn't nothing.

The cats that acted theirs out

The oldest and strangest evidence comes from a much rougher kind of experiment. In 1965, French researcher Michel Jouvet lesioned a small region of the brainstem in cats, disabling the mechanism that paralyzes the body during REM sleep. Normally that paralysis is what keeps a dreaming brain from moving a sleeping body. Without it, the cats stood up mid-REM, stalked, arched, swiped at things that weren't there.

That circuitry turns out to be the same one that fails in people with REM sleep behavior disorder, a real human condition where people physically act out their dreams, sometimes violently, discovered decades later and matched straight back to Jouvet's cats. Dog owners have their own informal version of this. Psychologist Stanley Coren, who has written extensively on canine cognition, has described watching dogs with a similarly disabled pons act out what looked like breed-specific behavior mid-sleep: pointers holding a point at nothing, dogs bred to flush birds running through the motion. None of that is a controlled clinical trial. It's an observation from someone who studies dog behavior for a living, and it lines up unsettlingly well with what the cat research already showed.

Birds that practice a song they haven't learned yet

The bird research is the one I find hardest to explain away. Daniel Margoliash's lab at the University of Chicago recorded from the forebrain neurons zebra finches use to produce song, both while the birds were singing and while they slept. The same neurons fired the same complex patterns during sleep, with small variations from one repetition to the next, as if the bird were running through a few slightly different versions of the same song.

Margoliash's read is that this is rehearsal, not memory replay in the rat sense. Young birds that haven't finished learning their adult song show the clearest version: exposure to an adult tutor changes their overnight neural firing pattern, and that change shows up in how the song sounds the next day. A bird that has never successfully sung its full song correctly appears to be practicing it, silently, at night.

Even octopuses seem to have two kinds of sleep

The most visually strange evidence isn't from a mammal at all. In 2021, Sylvia Medeiros and a team in Brazil filmed sleeping Octopus insularis and found the animals alternate between two distinct states roughly every half hour. One is quiet: pale, still, eyes closed. The other is active, lasting less than a minute, and during it the octopus's skin flickers rapidly through colors and textures while its eyes move behind closed lids, the same patterns it uses awake to camouflage, threaten, or hunt.

An octopus's skin color is controlled directly by its brain. There's no obvious reason for it to be doing that while asleep and alone in a tank unless something upstream is generating the same signal it would generate while awake. Cuttlefish research has turned up something similar. Whatever this is, it evolved on a branch of the tree of life that split from ours something like 500 million years ago, long before anything resembling a hippocampus existed.

It goes back further than mammals, too

For a long time REM sleep looked like something birds and mammals had and reptiles didn't, which fit a tidy story about it evolving once, late, in warm-blooded animals. That story broke in 2016, when a team at the Max Planck Institute recorded brain activity in sleeping bearded dragons and found clear REM and slow-wave sleep cycling, just much faster: around 80 seconds per cycle, compared to roughly 90 minutes in a human. That pushes the origin of this kind of sleep back to a common ancestor of reptiles, birds, and mammals that lived over 300 million years ago, near the start of animals living on land at all.

What none of this actually proves

Here's the honest caveat, and it's a real one, not a hedge. Every example above is measurable: neuron firing, skin color, muscle movement, EEG waves. None of it is a report of subjective experience, because animals can't give one. A rat's hippocampus replaying a maze is strong evidence of memory processing during sleep. It is not the same as evidence that the rat is having an experience while it happens, the way you have an experience during your own dreams. Researchers are careful about that gap for good reason. It's the actual crux of the question, and no amount of neuron data closes it by itself.

What the research does rule out is the easy version of the myth, that dreaming is a uniquely human overlay on top of otherwise blank animal sleep. The circuitry involved in human dreaming, memory replay, sensorimotor rehearsal, active sleep with REM-like signatures, shows up in species as distant from us as octopuses and bearded dragons. Something is running during their sleep that looks a lot like what runs during yours.

Why this is worth thinking about before bed

The rat and bird research both point the same direction: whatever sleep is doing, it's processing what happened during the day and rehearsing it. That's a large part of why keeping a dream journal is worth the two minutes it takes. Your brain is already doing this work every night, sorting the day's experience and running it back. A journal is just the closest thing you have to reading that replay out.

You won't get raw neuron traces the way Louie and Wilson did with their rats. But write down what surfaces for a few weeks and you'll start noticing your own version of the pattern: which parts of your day actually made it into the rerun, and which got left on the floor.