How the Brain Perceives and Distorts Time

Your brain is not recording reality. It is constructing a story about reality, using incomplete sensory data, ancient circuitry, and a memory system that is considerably less reliable than most people assume. Neuroscientist David Eagleman spent an hour with Neil deGrasse Tyson, Chuck Nice, and Gary O'Reilly on StarTalk making this case across nearly every domain of human experience, from why you cannot see a gorilla standing in front of you to why a terrifying fall from a 150-foot tower felt like it lasted four or five seconds when it lasted about two. The full conversation is available on YouTube.

The range of territory covered might look, at first glance, like a list of interesting brain facts. It is not. Every example Eagleman reaches for is a variation on one core claim: perception is not passive reception. It is active, competitive, and reconstructive all the way down. The breadth earns its keep because each example closes a potential escape route. You might accept that blind people repurpose their visual cortex for touch while telling yourself this says nothing about your own ordinary experience. Then Eagleman describes the invisible gorilla experiment, and the escape route closes.

The Brain Does Not Care Where Information Comes From

Eagleman opens with sensory substitution, and the logic is worth following carefully. The brain, he explains, receives not raw sensory data but electrical signals: photons converted to spikes, sound waves converted to spikes, pressure on skin converted to spikes. It is all the same currency. So a 1969 paper in Nature demonstrated that blind people could learn to perceive visual information delivered through patterns of stimulation against the skin of their backs, fed by a camera. The brain figured it out, not because it was told to, but because territory that goes unused gets colonized by competing signals.

This is the mechanism behind Braille reading activating what would otherwise be the visual cortex in blind individuals, behind deaf people developing heightened sensitivity to lip movement, behind thousands of blind people using echolocation clicks. Eagleman is careful to note that echolocation is not exotic: "Tens of thousands of blind people use echolocation," he points out, and the phenomenon was named in a 1930 paper in Science titled "Echolocation in Bats and the Blind." The documentary claim that one particular boy was the only human capable of it was simply wrong.

The competitive logic extends further. Eagleman and his colleagues at Harvard studied what happens when sighted people are tightly blindfolded and placed in a scanner. Eagleman cited their findings showing that within a surprisingly short window, the visual cortex begins responding to touch and sound rather than lying dormant. His takeaway was direct: cortical territory is not reserved. It is contested, continuously, and the contests happen faster than anyone expected.

Dreams as Territorial Defense

That observation about rapid cortical reorganization led Eagleman and his student Don Vaughn to a hypothesis about dreaming that is genuinely provocative. We live on a planet that rotates into darkness roughly half the time. For most of evolutionary history, darkness meant one sense, vision, was suddenly losing its input while every other sense continued operating. If cortical takeovers can begin within hours of deprivation, the visual cortex faces a recurring nightly threat from adjacent sensory regions.

Dreams, on this account, are not psychological processing. They are a defensive firing: every 90 minutes or so, a burst of random activity blasts into the primary visual cortex to keep it claimed. The brain, as Eagleman puts it, "is smart enough to know that it's not really that useful to remember" this random content, which is why dream memory degrades so rapidly after waking.

To test whether this is correlational or something more structural, Eagleman's lab examined dreaming behavior across multiple primate species with varying degrees of neuroplasticity. The more plastic the brain, and therefore the more vulnerable to cortical takeover, the more dreaming occurred. The correlation held. Eagleman notes that the underlying circuitry, called PGO waves (pontine-geniculate-occipital), appears documented across a wide range of species, suggesting it is ancient and fundamental rather than a quirk of primate neurology.

The hypothesis has one obvious implication Tyson surfaces immediately: if the theory is correct, plugging your ears for extended periods should produce auditory hallucinations. Eagleman confirms this is exactly what happens in solitary confinement. He also extends it to tinnitus, proposing that the chronic ringing many people experience after inner ear cell death is the auditory cortex generating its own signal because the expected input has gone quiet. He calls this the "defensive activation theory."

The Time Problem

Memory is not a recorder. Eagleman is insistent on this, and the September 11th research he describes makes the stakes concrete. Colleagues of his tracked New Yorkers who witnessed the towers collapse, quizzing them on what they saw on September 11th and on the banal events of September 10th, then re-interviewing them at intervals over a decade. Both memories drifted equally. The emotionally intense memories were not preserved with greater fidelity; they just felt more certain. The only details that stayed consistent were those repeatedly reinforced by news coverage, external references that pinned the reconstruction in place.

This is the necessary foundation for understanding why time distorts. Eagleman collected hundreds of accounts from people who reported experiencing slow-motion time during accidents, gunfights, and other life-threatening events. He then ran what he describes as the first controlled experiment on the phenomenon: dropping volunteer subjects from a 150-foot tower, catching them in a net at roughly 70 miles per hour. Subjects wore a wrist device that flashed visual information at rates calibrated to test whether perception actually slowed.

It did not. What changed was not the speed of perception but the density of memory encoding. The amygdala, functioning as an emergency priority system, effectively orders the brain to stop everything else and record everything now. When you reconstruct that experience afterward, you have an unusually rich archive to draw on, and you infer from the richness that more time must have elapsed. It is a trick of retrieval, not a change in processing speed.

The same mechanism explains the compression of adult time. "When you're a kid, everything is novel and you're writing down tons of memories," Eagleman tells Tyson. "But once you're an adult, your internal model of the world has developed pretty well and there's not that much that's really novel." Less novel experience means thinner memory encoding means a shorter subjective reconstruction of any given period. This is why childhood summers felt endless and why the last three years of your forties seem to have passed during a long weekend.

A Mirror Experiment Worth Trying

Eagleman describes one experiment any reader can run without a tower or a wrist device. Stand close to a mirror. Alternate looking at your left eye, then your right eye, then back. When you watch someone else do this, you can clearly see their eyes making rapid lateral jumps, called saccades. When you do it yourself, you experience no movement at all: just instantaneous switching. Your brain edits out the transit entirely, filling the gap with a seamless cut.

The experiment sounds trivial. Its implication is not. If the brain edits out the time occupied by your own eye movements without your awareness, what else is it editing? The answer, based on Eagleman's broader body of work, is: quite a lot. In a separate lab experiment he describes, subjects trained to expect a 200-millisecond delay between pressing a button and seeing a flash of light eventually stopped perceiving the delay at all. When the delay was then removed, they reported that the light flashed before they pressed the button. Cause appeared to precede effect because the brain's recalibration had overshot.

From Lab Curiosity to Clinical Hypothesis

Eagleman's reaction to that reversal experiment was not just fascination. He recognized it as a laboratory model of something clinically significant. In schizophrenia, a symptom pattern called credit misattribution causes patients to experience their own actions and internally generated voices as external events. "You're always talking to yourself," Eagleman explains. "You're always generating an internal voice and listening to it. But if you get the timing of that just slightly wrong so that you think you heard the voice before you generated it, that's an auditory hallucination."

He has been pursuing the hypothesis that schizophrenia is, at its core, a disorder of temporal processing, and that recalibrating timing might reduce or eliminate symptoms. He acknowledges the treatment application is speculative and years away: "I can't say that works, but I hope so." The clinical ambition is real but explicitly unproven, which is worth noting in a field where preliminary findings frequently acquire the certainty of established treatments long before evidence warrants it.

The Hardest Question

The episode closes where neuroscience always eventually arrives: consciousness. Tyson observes, with some acuity, that a field that has explained something tends to stop writing books about it. Neuroscience has not stopped writing books about consciousness. Eagleman agrees that the "why does it feel like something" question remains genuinely open. He frames it cleanly: a laptop and a brain are both sophisticated collections of parts, but we do not assume the laptop is having an experience. Why should neurons, chemicals, and electrical signals produce subjective experience at all?

His working answer is that consciousness functions as a high-level operating system, allowing 30 trillion cells to coordinate as a single decision-making entity across time. This makes consciousness useful rather than merely epiphenomenal, which is an important distinction. But usefulness does not explain the underlying mechanism, and Eagleman does not pretend it does.

The AI question he raises, how would we recognize consciousness in a language model or an advanced AI system, is not rhetorical decoration. If we cannot explain what generates subjective experience in biological systems where we are confident it exists, we have no principled basis for detecting its presence or absence in artificial ones. That is a problem with consequences that extend well beyond the laboratory, and the neuroscience has not caught up with it yet.

By Priya Sharma, Science and Health Correspondent