Why Do Your Eyes Move While You’re Sleeping?

Your eyes move during sleep because circuits deep in the brainstem periodically activate the same neural machinery that drives eye movements when you are awake. The most dramatic bursts happen during rapid eye movement (REM) sleep, a recurring stage of the sleep cycle first documented in the 1950s, where the brain becomes highly active while the rest of the body stays largely paralyzed. But these movements are not random twitches. A growing body of research suggests they reflect something remarkable: your sleeping brain is looking around inside its own self-generated world.

The Brainstem Engine Behind the Movements

The eye movements of REM sleep originate in a region called the dorsolateral pontine reticular formation, a cluster of neurons in the brainstem that interacts with other brainstem and forebrain areas to generate the core features of REM sleep.¹ This area acts as a kind of switch. When these neurons become active, they set off a cascade: the cortex wakes up into a pattern resembling alert consciousness, muscles throughout the body lose their tone (a state called atonia), and the eyes begin darting beneath closed lids. Pioneering work by researchers Eugene Aserinsky, Nathaniel Kleitman, William Dement, and Michel Jouvet in the 1950s and 1960s established these foundational features and showed that REM sleep recurs in roughly 90-minute cycles throughout the night.²

The brainstem does not simply flip a switch and walk away. It sends bursts of electrical activity called ponto-geniculo-occipital (PGO) waves that travel from the pons through the visual relay station of the thalamus and into the visual cortex. These waves arrive in parallel with the rapid eye movements themselves, stimulating the visual processing system even though no light is entering the eyes. In a very real sense, the brain is running its visual hardware without any external input.

Are Your Eyes Actually Tracking Your Dreams?

This is the question researchers have debated for decades, and the answer turns out to be nuanced: sometimes yes, sometimes no, and the “yes” part is more interesting than it sounds.

The so-called scanning hypothesis proposes that rapid eye movements correspond to where a dreamer is looking in their dream scene. Early studies tried to test this by waking sleepers mid-REM, asking them to describe what they had just been dreaming, and then checking whether the direction of recorded eye movements matched the reported dream content. Results were wildly inconsistent, with concordance rates ranging from about 9% to 80%.³ The problem was partly methodological: people are not great at accurately recalling dream details moments after waking, and the muscle paralysis of REM sleep makes it hard to determine exactly where the eyes were pointing relative to the head.

Several observations complicate the simple scanning story. Fetuses, newborns, and people who have been blind since birth all show rapid eye movements during sleep, even though none of them are scanning visual dream imagery in the conventional sense. Some REM-sleep eye movements are not even conjugated, meaning the two eyes do not move together toward a single point, which is what you would expect if someone were actually focusing on an object.³

But more recent evidence has tipped the balance. Researchers studying patients with REM sleep behavior disorder, a condition where people physically act out their dreams because the normal muscle paralysis fails, found something striking. When patients performed goal-directed actions in their dreams, like reaching for something or climbing, 90% of their eye movements were directed toward the action they were performing.³ Because these patients could move their heads and limbs freely, researchers could directly compare the direction of eye, head, and limb movements and found strong coherence. The eyes were looking where the action was.

A study using lucid dreamers, people who become aware that they are dreaming while still asleep, took the evidence further. Lucid dreamers were asked to trace specific shapes with their eyes inside a dream, like circles or infinity signs, by following their dream-image of a thumb. The resulting recordings showed recognizable versions of those shapes in the electrical signal from the eyes.⁴ Intentional shifts in dream gaze produced corresponding physical eye movements.

Perhaps the most elegant evidence came from a mouse study that tracked the head-direction system in the thalamus, a population of neurons that essentially reports which way the animal is “facing” in its environment. During REM sleep, the direction and size of the mice’s eye movements matched the direction and size of the changes in their virtual head direction.⁵ The rapid eye movements were not random brainstem noise. They were tracking gaze shifts in whatever internal world the sleeping brain was generating.

The current picture, then, is that at least a subset of rapid eye movements genuinely correspond to gaze direction in dreams. But not all of them do, and some clearly arise from brainstem activity that has nothing to do with visual content. The eyes are partly watching the dream and partly just doing what activated brainstem circuits tell them to do.

Not All REM Sleep Is the Same

Even within REM sleep, the eyes are not moving constantly. Researchers distinguish between phasic REM, the moments when rapid eye movements are actually occurring, and tonic REM, the quieter stretches of REM sleep between bursts of eye movement. These two sub-states are neurologically quite different.

During phasic REM, the brain appears to become more internally focused. Studies using auditory stimulation have found that the brain’s responses to outside sounds are diminished when the eyes are actively moving, suggesting a kind of sensory gating.⁶ In tonic REM, by contrast, the brain responds to external sounds in a way that looks more like wakefulness. The pattern resembles something familiar from waking life: when you make a rapid eye movement (a saccade) while awake, your visual perception briefly dips in what is known as saccadic suppression. Recordings from single neurons in people undergoing brain monitoring have shown a transient drop in neural activity just before each REM onset during sleep, hinting at a similar suppression mechanism operating without any actual visual input.⁷

The electrical brain activity during these two sub-states also differs in measurable ways. Analysis of the “aperiodic” component of brain signals, essentially the background neural noise pattern, shows distinct signatures between phasic and tonic REM, with medium-to-large differences in certain frequency ranges.⁸ The upshot for the dreamer: the moments when your eyes are actively darting around may be the moments when you are most deeply immersed in the dream and least aware of the outside world.

Eye Movements That Happen Outside REM Sleep

Rapid eye movements get the headlines, but your eyes also move during the transition into sleep and during lighter sleep stages. These are slow eye movements (SEMs), gentle rolling or drifting motions that are fundamentally different from the sharp, fast saccades of REM sleep. They tend to appear as you start to feel drowsy and persist through the earliest stages of non-REM sleep.⁹ Sleep researchers use them as a reliable marker of sleepiness and the onset of sleep.

These slow eye movements show an interesting relationship with breathing. Research has found that the timing of slow eye movements and respiratory rhythms are coordinated, not because the eyes are literally breathing but because both are driven by overlapping circuits in the brainstem’s reticular formation.¹⁰ The brainstem is essentially idling on shared circuitry as wakefulness fades.

Eyelid movements during sleep tell their own story. In one study, when sleepers were woken from stage 2 (light non-REM) sleep during a period of spontaneous eyelid movements, they were significantly more likely to report mental imagery than when woken from eyelid-quiet periods of the same sleep stage. This did not hold during REM sleep, where dream imagery was common regardless of eyelid activity, but it suggests that even outside REM, certain eye and eyelid movements may accompany some form of internal visual experience.

What Eye Movements Do for Memory

The eye movements of REM sleep are not just a byproduct of dreaming. They appear to play a functional role in how the sleeping brain consolidates what you learned during the day. Research on procedural memory, the kind of learning involved in mastering a new skill, found that the brain waves immediately following eye movements during post-learning REM sleep were different from those on a control night. The oscillatory changes that accompanied phasic REM eye movements correlated with overnight improvement in task performance, while changes during tonic REM periods did not show the same predictive relationship.¹¹

This suggests that the bursts of eye movement during REM sleep are not just indicators that dreams are happening. They mark moments when the cortex is actively engaged in processing and storing recently acquired information. The eye movements are tied to the cognitive work, not incidental to it.

Why Newborns Spend So Much Time in REM

Newborn humans spend roughly half their total sleep time in REM sleep, compared to about 20-25% in adults. Premature infants spend even more. This pattern holds across many species: the younger the animal and the more rapidly its brain is developing, the more REM sleep it gets.¹²

The leading explanation is that REM sleep provides a kind of internal stimulation that supports brain maturation. During a period when the visual system, motor pathways, and neural connections are rapidly forming, the brainstem-driven activity of REM sleep, including eye movements, may help optimize neural development, from the formation of new connections to the pruning of unnecessary ones.¹² Newborns are not scanning elaborate dream scenes. Their REM-sleep eye movements are more likely a sign that developing brain circuits are being exercised and tested.

What Happens in People Who Cannot See

If rapid eye movements were purely about scanning dream imagery, you might expect them to vanish in people who have never had vision. They do not, though they are reduced. In a study comparing blind and sighted individuals, both congenitally blind and late-blind subjects showed less eye movement activity during REM sleep than sighted controls.¹³ The reduction was not limited to REM sleep; the blind groups also showed less eye movement during wakefulness and lighter sleep stages. An earlier study found that seven out of eight people with lifelong blindness still had rapid eye movements during REM sleep, though the overall pattern differed from sighted sleepers in frequency, amplitude, and how well the two eyes coordinated.¹⁴

Brain imaging research has shed some light on why. In congenitally blind individuals, eye movements still correlate with activity in the motor cortex, which makes sense since the motor commands for moving the eyes are intact. But the connection between eye movements and the visual cortex is weaker, and the temporal pattern of these signals differs from what is seen in sighted people.¹⁵ Some blind individuals even showed uncoordinated movements between the two eyes, with each eye’s movements linked to different brain networks. The brainstem circuitry that generates the movements persists, but without visual experience shaping the system, the movements are less organized and less connected to the brain’s visual processing areas.

When the Paralysis Fails

During normal REM sleep, your brain is busy sending motor commands, not just to the eyes but to the limbs, the torso, everywhere. You do not act on those commands because a separate brainstem mechanism actively suppresses muscle tone throughout the body. The eyes are spared from this paralysis, which is why they can move freely while everything else stays still.

In REM sleep behavior disorder (RBD), that suppression breaks down. The brainstem circuits responsible for maintaining muscle atonia during REM sleep malfunction, allowing people to physically act out their dreams.¹⁶ This can mean anything from mild twitching to violent kicking, punching, or leaping out of bed. The disorder is characterized by loss of the normal skeletal muscle atonia that should accompany REM sleep.¹⁷

RBD is clinically significant for a concerning reason: it is strongly associated with later development of neurodegenerative diseases, particularly those involving synuclein protein buildup in the brain. The brainstem region that controls REM atonia is among the earliest affected by these diseases, so the loss of sleep paralysis can precede other symptoms by years or even decades. The condition is diagnosed through overnight polysomnography that demonstrates the absence of normal muscle atonia during REM periods.

Medications That Change Sleep Eye Movements

Certain medications, particularly selective serotonin reuptake inhibitors (SSRIs), can cause eye movements to appear during stages of sleep where they normally would not. In a study of polysomnography recordings, prominent eye movements during non-REM sleep were detected in a group of patients, and 78% of those patients were taking an SSRI at the time. Among all patients taking an SSRI, about 36% showed these abnormal non-REM eye movements.¹⁸ Other classes of antidepressants and medications like benzodiazepines had a much lower incidence of the same effect. SSRIs are also well known to suppress REM sleep overall, reducing the total time spent in that stage, which makes the appearance of eye movements in non-REM sleep all the more striking. It is as though the medication displaces some REM-like phenomena into sleep stages where they do not belong.

Eye Movements in the Sleep of Other Species

REM sleep with its characteristic eye movements is not unique to humans. It is shared broadly across mammals and birds, though the specific expression varies considerably from one species to another.¹⁹ Cats, dogs, rats, and primates all show recognizable REM periods with rapid eye movements, muscle atonia, and brain activation patterns that parallel the human version.

The more surprising finding is how far down the evolutionary tree something resembling REM sleep may extend. Most insects and arthropods cannot move their eyes at all, which makes studying the phenomenon in invertebrates difficult. Jumping spiders, however, are an exception. They have movable retinal tubes that they use to redirect their gaze, and in young spiderlings, these tubes are visible through the temporarily translucent exoskeleton. Researchers observed that during nocturnal rest, jumping spiders showed periodic bouts of retinal movements coupled with limb twitching and stereotyped leg-curling behaviors, a pattern that looks strikingly like REM sleep.²⁰ Whether this is truly homologous to mammalian REM sleep or a case of convergent evolution remains an open question, but it hints that the coupling of sleep with periodic eye movements may be an ancient and widely useful arrangement.

A Hypothesis About Eye Health

One lesser-known idea about REM-sleep eye movements has nothing to do with dreams, memory, or brain development. A hypothesis proposed in the 1990s suggests that the physical churning of the eyes during REM sleep serves a mundane but essential purpose: stirring the aqueous humor, the fluid inside the eye, behind closed lids.²¹ When the eyelids are shut for hours, the cornea loses its exposure to atmospheric oxygen. The aqueous humor becomes the cornea’s main oxygen source, but if the fluid sits still, local oxygen levels near the cornea can drop dangerously. According to this idea, the rapid movements of the eyes stir the fluid enough to prevent stagnation and keep the cornea oxygenated through the night.

This hypothesis has not been widely adopted as the primary explanation for REM-sleep eye movements, and the brainstem-driven and dream-related functions have far more supporting evidence. But it is a reminder that biological systems often serve multiple purposes simultaneously, and the mechanical act of moving the eyeball may have consequences beyond whatever the brain is up to.

Eye Movement Patterns in Psychiatric Conditions

For decades, researchers explored whether the density of eye movements during REM sleep could serve as a biological marker for psychiatric conditions like depression and schizophrenia. The intuition was appealing: if depression involves disturbances in REM sleep, maybe the specific pattern of eye movements could help with diagnosis. Some early work found a positive relationship between self-rated depression scores and REM eye movement density in a small sample of young men.²²

But larger and more systematic studies showed that the range of eye movement densities in people with schizophrenia and depression overlaps substantially with what is seen in healthy controls. The conclusion from that work was straightforward: abnormalities of REM-sleep eye movement activity should not be considered a biological marker for affective illness.²³ The overlap is simply too large to be diagnostically useful.

Post-traumatic stress disorder may be a different story. A study comparing sleep in PTSD and major depression found that REM density was significantly higher in the PTSD group, and this was correlated with nightmare complaints.²⁴ Whether this reflects the brain replaying traumatic imagery (consistent with the scanning hypothesis) or simply heightened arousal of the brainstem systems that generate REM sleep is still being worked out. Either way, the eyes of someone with PTSD appear to be busier during REM sleep than those of someone with depression alone.