Can Dreams Predict the Future, or Does Memory Rewrite It?

Science

Can Dreams Predict the Future, or Does Memory Rewrite It?

alok

September 25, 2026

11–17 minutes

Table of Contents

There is a peculiar sentence people sometimes say only after something astonishing has happened:

“I knew this was going to happen.”

Perhaps someone dreams of a car accident, and the next day a relative has one. Someone walks into a building they have never visited and suddenly knows where the corridor will turn. A person thinks intensely about an old friend and that friend calls minutes later.

The experience can be so convincing that dismissing it as “just coincidence” feels almost insulting.

But there is a problem.

The human brain is not merely a machine for recording events. It is a machine for finding patterns, reconstructing memories and assigning meaning. Sometimes those abilities help us navigate the world. Sometimes they create an astonishing illusion: the feeling that yesterday contained information about tomorrow.

The scientific question, therefore, is not simply “Can people have premonitions?”

It is more precise:

When someone says, “I knew this before it happened,” did the information genuinely exist beforehand or did the brain create the certainty afterward?

That distinction changes everything.

The Strange Experience of Knowing Something Before It Happens

Consider two statements.

“I felt something was going to happen.”

and

“I knew exactly what was going to happen.”

They sound similar, but scientifically they are radically different.

A vague feeling is easy to match with thousands of possible future events. A precise prediction recorded before the event is much harder to explain away.

Imagine writing:

“Tomorrow at 3:15 p.m., my brother will call me from Delhi about a broken red car.”

If that happens, we have something interesting.

But saying afterward, “I had a strange feeling something was wrong with my brother,” is different. The statement contains enormous room for interpretation.

This is where coincidence, memory and probability begin working together.

Déjà vu is particularly deceptive

Déjà vu is the strange sensation that a present situation has happened before, even though we know it has not.

Modern research increasingly treats déjà vu as a memory phenomenon rather than evidence of foreknowledge. Experiments by Anne Cleary and colleagues found that people experiencing déjà vu can develop a strong feeling that they know what will happen next but their actual ability to predict the next event does not rise above chance.

In other words, the brain can generate:

“I remember this.”

and then quietly transform it into:

“Therefore, I know what happens next.”

That second step is the fascinating one.

The feeling of prediction can be real even when the prediction itself is not.

Why Hindsight Makes Coincidences Feel Extraordinary

Suppose you have 10,000 dreams over several years.

Most disappear.

A few involve relatives. Some contain hospitals. Others contain flights, storms, arguments, deaths, weddings, accidents, strange buildings, old friends and countless other fragments.

Then one day something dramatic happens.

Suddenly, one old dream appears to resemble it.

The coincidence feels extraordinary because the unsuccessful dreams have vanished from the stage.

This is selective recall.

Research specifically examining precognitive-dream experiences has found that people are more likely to remember dream-event pairs that were confirmed than pairs that were disconfirmed. In one study, participants showed significantly greater recall for confirmed than disconfirmed dream-event combinations.

That produces a powerful psychological illusion:

Dream → event → remembered match

while thousands of:

Dream → nothing

are barely noticed.

The brain is effectively keeping a highlight reel.

And nobody watches the deleted footage.

The Mathematics of Coincidence Is More Strange Than It Looks

Coincidence does not require something supernatural.

It requires enough opportunities.

Imagine you think about an old friend once every few days. That friend also communicates with you occasionally. Eventually, the two events will happen close together.

Now multiply this across:

thousands of thoughts,

thousands of dreams,

hundreds of people,

years of memories,

millions of daily events.

Rare coincidences become inevitable.

This is why probability must be considered before calling something a prediction.

A dream containing “a disaster” has a huge number of possible matches.

A dream containing:

water,

crowds,

a vehicle,

someone becoming ill,

an argument,

a familiar person,

can potentially be matched to countless real-world events.

The narrower the prediction, however, the more scientifically interesting it becomes.

A prediction such as “something bad will happen” has very little evidential power.

A prediction such as “on Tuesday, my sister will break her left wrist while cycling”, recorded before Tuesday and independently timestamped, is dramatically different.

That is why controlled experiments matter.

The Dream That People Claimed Was a Warning

The most haunting stories usually involve dreams.

Someone dreams that a relative dies.

Days later, the relative dies.

Someone dreams of a plane crash.

The following week, news reports an aviation disaster.

Someone dreams about an unfamiliar place.

Months later, they visit somewhere that strangely resembles it.

The emotional impact can be enormous.

But emotional intensity is not the same thing as evidential strength.

In fact, emotional intensity may make the memory stronger and more resistant to correction.

Why Dreams Are Perfect Machines for Producing Coincidences

Scientists still do not have a complete explanation for why humans dream. The National Institutes of Health notes that dreaming is not restricted to REM sleep and that scientists are still investigating its purpose.

One important point is often forgotten:

Dreams are enormously complicated.

They can contain people, locations, conversations, emotions, objects, colours, movements and bizarre combinations that would never occur in ordinary life.

That enormous combinatorial space creates opportunities for accidental matches.

Imagine a dream containing 20 different elements.

Even if each individual element is ordinary, the possible combinations become enormous.

Now imagine having dreams repeatedly for decades.

Eventually, reality will occasionally produce something that looks uncannily similar.

That does not prove that the dream predicted reality.

It demonstrates something equally important:

A sufficiently large number of opportunities can manufacture apparently impossible coincidences.

Why Tragic Dreams Become Especially Memorable

Suppose you dream about a stranger falling from a building.

Nothing happens.

You forget the dream.

A month later, you dream about a family member visiting a hospital.

Nothing happens.

You forget that too.

But imagine dreaming that your mother dies—and three days later she becomes seriously ill.

Even if the dream did not accurately predict the illness, the emotional shock makes the memory extremely difficult to forget.

Research increasingly supports the importance of emotion in dreaming and memory. A 2024 Scientific Reports study involving 125 participants examined dreaming and emotional memory processing. Among the 122 participants who supplied usable dream reports, 51 were classified as dream recallers and 71 as non-recallers. The researchers found differences in emotional-memory processing between these groups and reported evidence consistent with an active role for dreaming in overnight emotional processing.

This matters because emotionally important events are not remembered like ordinary events.

The brain effectively puts a red circle around them.

So when a dream seems to predict something emotionally shocking, two forces combine:

coincidence + enhanced memory

The result can feel supernatural.

Could Dreams Actually Predict Ordinary Future Events?

There is a more interesting possibility than supernatural prophecy.

Perhaps some apparent “predictions” are actually unconscious predictions.

Your brain constantly absorbs information that you do not consciously notice.

A relative may look unusually tired.

A relationship may be deteriorating.

A colleague may behave differently.

A storm may be approaching.

A company may be showing signs of trouble.

You may not consciously connect these clues.

But your brain has already received them.

During sleep, those fragments can be recombined into a dream.

If the expected event later happens, the dream can appear prophetic.

Researchers have explicitly investigated this implicit-processing hypothesis. One proposal is that people may unconsciously detect subtle environmental information and incorporate it into dreams; when the predicted event subsequently occurs, the dream is interpreted retrospectively as precognition.

This would not be paranormal.

It would be an extreme example of ordinary prediction occurring below conscious awareness.

The Most Important Experiment: Write the Dream Down First

Here is where the debate becomes scientifically useful.

If someone believes their dreams predict the future, there is a remarkably simple way to test the claim:

Record the dream before the event occurs.

Not later.

Not from memory.

Not after reading the news.

Immediately.

Ideally:

Write the dream immediately after waking.

Add the exact date and time.

Do not edit it later.

Store the original record somewhere that preserves its timestamp.

Define beforehand what would count as a “hit.”

Define what would count as a “miss.”

Compare the dream with future events using predefined rules.

Include unrelated control events.

Have independent judges evaluate similarities.

Repeat the experiment many times.

This is much stronger than collecting remarkable stories after the fact.

Why?

Because hindsight has been locked outside the laboratory.

What Happens When Researchers Try This?

The evidence is mixed at the level of individual studies, but the major methodological problem is consistency.

Some parapsychological experiments have reported statistical anomalies interpreted as evidence for precognition.

The famous 2011 work by Daryl Bem reported nine experiments claiming that future events could influence earlier psychological responses. The results attracted enormous attention precisely because they challenged conventional assumptions about causality.

But replication became the critical test.

Three preregistered attempts to replicate one of Bem’s effects, involving a combined sample of 150 participants, failed to reproduce the statistically significant result.

Another replication programme involving seven experiments and 3,289 participants also failed to reproduce the relevant finding, reporting an average effect size of approximately d = 0.04, statistically indistinguishable from zero.

More recently, an unusually large metascientific project tested a Bem precognition experiment using 26,483 participants and 420,472 critical trials across three studies.

Study 1 failed to replicate the original effect.

Study 2 produced a statistically significant result in the opposite direction.

Study 3, however, failed to reproduce Study 2’s result.

The researchers concluded that, under conventional scientific standards, the lack of replicability and absence of an accepted mechanism make a methodological or false-positive explanation more plausible than a demonstrated violation of established physical laws.

That is an important distinction.

A strange statistical result is not automatically evidence of precognition.

It is evidence that something happened in an experiment that requires explanation.

What About Actual Precognitive Dream Studies?

Dream research presents an even harder problem because dreams are flexible.

A researcher can compare a dream with many possible future events and discover similarities.

The question becomes:

How many similarities were possible before the comparison began?

A controlled study of precognitive dreaming involving 50 participants asked people to dream about video clips they would later view. Independent judges evaluated dream-target correspondence. The study found no support for the hypothesis that people who were less tolerant of ambiguity would produce greater correspondence with future targets. Participants’ prior confidence was related to perceived similarity, but perceived similarity did not correspond to judges’ actual hit rates or similarity ratings.

That finding is revealing.

People can feel that their dream matches the future without independent judges finding the same level of correspondence.

In other words:

subjective certainty and objective accuracy can separate.

Why Dream Diaries Are Better Than Memory

A dream remembered six months later is already contaminated by six months of experience.

A diary entry written at 7:00 a.m. on the morning after the dream is different.

The diary freezes the evidence before the future happens.

Dream research has repeatedly demonstrated that retrospective estimates and diary measurements can differ. In one study of lucid dreaming involving 425 people and 1,612 dreams, diary-based estimates were lower than retrospective estimates of lucid-dream frequency.

That is a small but important lesson:

What people remember about their experiences is not always identical to what they actually recorded while experiencing them.

The same principle applies to alleged premonitions.

Déjà Vu, Dreams and Premonition Are Not the Same Thing

Experience	What the person feels	What research suggests
Déjà vu	“I have experienced this before”	Strongly associated with memory/familiarity processes
Déjà vu + prediction	“I know what happens next”	Feeling can increase without actual predictive accuracy
Ordinary intuition	“Something feels wrong/right”	Can arise from unconscious processing of subtle cues
Precognitive dream claim	“I dreamed this before it happened”	Evidence remains controversial and difficult to reproduce
Retrospective match	“My old dream predicted this”	Highly vulnerable to selective recall and hindsight
Controlled prediction	“This exact event was recorded beforehand”	Much stronger evidence because hindsight is constrained

Research on déjà vu is particularly instructive. Participants can experience a powerful sense of prediction during déjà vu without demonstrating above-chance predictive performance.

The brain can therefore manufacture a very convincing feeling of knowing.

That does not make the experience fake.

It means the experience and its interpretation are two different things.

What We Know, What We Think, What We Don’t Know

What We Know

Human memory is reconstructive.

Déjà vu can create powerful feelings of familiarity and prediction without demonstrated predictive accuracy.

Dreaming is closely connected with memory and emotional processing, although its precise function remains incompletely understood.

Selective recall can make confirmed dream-event matches more memorable than disconfirmed ones.

Large replication efforts have generally failed to establish a robust, reproducible precognitive effect.

What We Think

Many apparent premonitions can plausibly arise from a combination of:

unconscious prediction + coincidence + selective attention + memory reconstruction.

Some dreams may incorporate subtle information that the sleeper encountered while awake without consciously recognizing it.

What We Don’t Know

Science cannot simply declare that every strange subjective experience has been completely explained.

Dreaming itself remains incompletely understood.

Human intuition remains an active research area.

And unusual experiences can be psychologically meaningful even when they do not demonstrate paranormal causation.

The scientifically responsible position is therefore neither:

“Every premonition is supernatural.”

nor:

“Every strange experience is meaningless.”

It is:

“Show us what was recorded before the event, define the prediction precisely, control the alternatives, and see whether the effect survives repetition.”

The Real Mystery May Not Be the Future

Perhaps the strangest part of a premonition is not that someone apparently saw tomorrow.

It is that the human brain can make yesterday look as though it already contained tomorrow.

A dream occurs.

An event happens.

Memory connects them.

Emotion strengthens the connection.

Probability supplies the coincidence.

And hindsight quietly removes the failures.

By the time we tell the story, the original uncertainty may have disappeared completely.

That is why the sentence “I knew it” deserves investigation.

Sometimes it means:

“I genuinely noticed something before I consciously understood it.”

Sometimes it means:

“I had a vague feeling that happened to match reality.”

And sometimes it means:

“I remembered the past differently after learning what happened.”

Those are three very different psychological events.

The future may be mysterious.

But memory is already strange enough.

Frequently Asked Questions

Can déjà vu predict the future?

Current experimental evidence does not show that déjà vu gives people above-chance predictive ability. Research instead suggests that the feeling of prediction can arise from familiarity and memory processes.

Can dreams predict future events?

There is no established scientific evidence that dreams reliably predict future events. Some studies have reported anomalous findings, but replication and methodological problems remain major issues.

Why do some dreams seem incredibly accurate?

Possible explanations include coincidence, selective memory, emotional salience, broad dream content and unconscious processing of information encountered while awake.

Is a dream diary useful?

Yes. A diary made before the predicted event occurs is much stronger evidence than a memory reconstructed afterward.

What would count as strong evidence for precognition?

A specific prediction recorded and timestamped before the event, evaluated using rules established in advance, with independent judges and repeated successful trials.

Why do people remember failed predictions less?

Because attention and memory are selective. Research on precognitive-dream experiences has specifically found greater recall for confirmed than disconfirmed dream-event pairs.

Does science prove that premonitions are impossible?

Science does not normally prove universal impossibility. The more precise conclusion is that reliable precognition has not been established by reproducible evidence sufficient to overturn conventional explanations.