The Global Dream Lab

Listen to This Article

Audio Recording by Audm

To hear more audio stories from publications like The New York Times, download Audm for iPhone or Android.

Deirdre Barrett’s body was in bed, but her mind was in a library.The library was inside a very old house, with glowing oil lamps and shelves of beautiful leatherbound books. At first it felt snug and secure and timeless, exactly the sort of place an academic like Barrett, who teaches in the psychiatry department at Harvard Medical School and edits the scientific journal Dreaming, might find inviting. But as the dream went on, she remembered later, “I became less able to focus on the library and more overwhelmed by the unseen horror outside.” Beyond the windows of the softly lit library, “a terrible plague was ravaging the world.”

When Barrett woke up, it was mid-March of 2020. She had been reading about the novel coronavirus in Wuhan since it began to make headlines, and she wondered, as she often did when she read about events in the news, how this one might be showing up in the dreams of the people who were experiencing it: residents on lockdown in China, overwhelmed doctors and nurses in Italy. The dreamlife of collective catastrophe was something she had studied repeatedly during her academic career — analyzing, for example, the dreams of Kuwaitis after the Iraqi invasion and those of British officers held prisoner by the Nazis during World War II, to see how the dreams compared with one another and with dreams from calmer times.

As a child, Barrett was fascinated by her own dreams, which were often vivid. They tended to stay with her well after she woke up, making nights feel like a time for slipping in and out of new worlds and adventures, often ones she’d read about but was now able to interact with and inhabit fully. When she grew up, she decided, she would become a writer of fiction; many of the early stories she wrote were set not just in worlds that she imagined, but also in and out of the various dream worlds of her characters. She was deeply curious about the dream lives of other people: When she started writing for her high school newspaper, she occasionally asked her sources if they’d had dreams related to whatever she was interviewing them about. Dreams were a window, albeit a very strange one, into the way that other people and their minds worked. In college Barrett decided that fiction was not her future (though she did develop a practice of making visual art about what she saw and felt while sleeping). What she wanted was to be a scientist who studied what happened inside dreams.

Perhaps this sounds like an oxymoron. Science, after all, is about what is observable, quantifiable, testable, predictable, explicable — and dreams are none of these things. They happen inside someone else’s head, quite invisibly to observers, and can be accessed, at best, through blurry and fragmented bits of fast-fading memory. Their bizarre, arbitrary-seeming contents seem to defy all narrative logic (“I was in my grandmother’s dining room, except it looked like my middle school cafeteria, and then suddenly my old orthodontist and this character from a book I’m reading were there”). As Barrett worked her way through a Ph.D. in psychology, she learned that many experts in the field believed that dreams were fundamentally meaningless — that they had no evolutionary purpose of their own and were merely a side effect of random neural firings as the sleeping brain went about more important business. It was silly, the thinking went, to pay too much attention to the results of dozy neurons making odd little stories out of loose bits and pieces rattling about in our brains.

Barrett, however, never lost her conviction that dreams mattered. Her first book was an edited collection that took seriously the dreams of trauma survivors: “Dreams,” she wrote, “can give voice to the unspeakable and begin to restore the savaged.” A subsequent book, “The Committee of Sleep,” examined the role of dreams in creativity, noting that dreams were credited as the direct origin of, to name a few examples, Jasper Johns’s “Flag,” the character Stuart Little and the plot of “Frankenstein,” the Beatles’ song “Yesterday,” the first ironclad battleship, the scientific breakthrough that earned researchers the 1936 Nobel Prize in Medicine and — though this one may be apocryphal — the structure of the periodic table. Stephen King, who struggled with the conclusion of “It” before dreaming the ending exactly as he published it, once explained that he uses dreams as a purposeful part of the creative process, “the way you’d use mirrors to look at something you couldn’t see head-on.” Barrett was also drawn — like researchers who study the dreams that follow hurricanes and fires and wars — to large, collective events, things that lots of different people experienced and then dreamed about. One person’s dreams might seem idiosyncratic and incoherent, but when you looked at many people’s dreams, all affected by the same experience, you could find patterns. Within patterns, you might find meaning.

Analysis of the dreams that followed the Sept. 11 attacks — including one study that followed subjects who had already been recording their dreams when the attacks happened — found that the aftermath affected dreamers differently. First responders and survivors often dreamed realistic versions of the trauma they had experienced; some nightmares stuck, recurring night after night, while others introduced new elements and changed over time. In her sample, Barrett interviewed two witnesses who kept dreaming about people jumping from buildings, until finally their dreams equipped the jumpers with umbrellas or parachutes to float safely down to the ground. But even people who had merely watched the attacks on television experienced a surge of anxiety dreams and nightmares. Their dreams often replaced or mixed what really happened with a variety of other disasters, like being swept away by a tsunami or shaken by an earthquake, which researchers understood as metaphoric placeholders for the emotions they felt.

After 9/11, Barrett believed that she’d never encounter another event that would have such a profound and widespread influence on dreams. Now, as the new virus spread and the world began to shut down, she realized she had been wrong. Like her dream-self in the library, she was living through a storm — what she would later call “the biggest crisis to be reflected in our dreams in my lifetime” and also “an extraordinary time in the history of dreaming.” Even as she, too, was now locked down at home, she felt she couldn’t miss the chance to try to study what it all meant.

The study of dreams has sometimes been dismissed as unrigorous at best, woo-woo at worst. When I attended a virtual conference of the International Association for the Study of Dreams in June, the sorts of research-paper presentations I was used to from other scientific conferences were punctuated by sessions dedicated to premonitions, writing dream haikus and something called a “dream telepathy contest.” But researchers have increasingly brought the tools of the social and biomedical sciences to the worlds inside our dreams. There are sleep labs attached to universities around the world, studying dreamers with fM.R.I.s and EEGs and publishing empirical research in respected journals. Scientists are studying the use of virtual reality tools to “engineer” dreams and sleep experiences, and have even begun to use brain scans to create algorithms that allow them to predict, though still with limited accuracy, what images dreamers are seeing. Their goal is to breach the walls that surround our understanding of dreams. But even with sleep labs and brain scans, laments Tore Nielsen, the director of the Dream and Nightmare Laboratory at the University of Montreal, dreaming just “doesn’t allow a lot of doorways in.” That’s what the pandemic offered: not an answer, but another access point, a natural experiment built of collective experience.

As the novel coronavirus spread and much of the world moved toward isolation, dream researchers began rushing to design studies and set up surveys that might allow them to access some of the most isolated places of all, the dreamscapes unfolding inside individual brains. The first thing almost everyone noticed was that for many people, their dream worlds seemed suddenly larger and more intense. One study of more than 1,000 Italians living through strict lockdown found that some 60 percent were sleeping badly — prepandemic, only a third of Italians reported trouble sleeping — and they were also remembering more of their dreams than during normal times and reporting that those dreams felt unusually real and emotional and bizarre. In Wuhan, a study of 100 nurses conscripted to work on the front lines found that 45 percent of them were having nightmares — a rate, Nielsen notes, that is “twice the lifetime rate among Chinese psychiatric outpatients and many times higher than that among the 5 percent of the general population who have nightmare disorder.” In France, the Lyon Neuroscience Research Center found that dream recall increased by 35 percent the month after lockdowns began, while a survey of 3,000 Americans found that nearly a third noticed themselves suddenly remembering more dreams. Even social media sites, researchers found, were full of people surprised at how much more active and vivid their dream lives had become. “Is it just me?” many of them asked. It was not.

Some of this surge in dreaming probably reflected the anxiety of the moment; research has shown that the emotions of waking life are reflected in dreams more often than actual events are, while our most emotional dreams are the ones we are most likely to remember. Another factor in the surge of dreaming was simple timing. When humans sleep, we enter REM sleep — the stage of sleep when most of our emotional, richly narrative dreams occur — about once every 90 minutes. (It’s a myth that we dream only during REM sleep.) As the night progresses, these REM periods become longer and longer: The first might last just five minutes, compared with 40 for a later one. Because REM sleep is so backloaded toward the morning, a shortened night of sleep can be catastrophic for dreaming. “When you sleep six instead of eight hours, you don’t lose one-fourth of your dream life,” Barrett noted in “Pandemic Dreams,” the short book she would later assemble about the results of her own survey. “You lose almost half — and exactly the dreams that will be the most vivid, bizarre and memorable.”

Access to enough sleep has become a privilege: Researchers now study what they call “the great sleep divide,” which cuts along racial and socioeconomic lines and helps consign poorer people to worse sleep and worse health. In 2017, the psychologist Rubin Naiman argued in The Annals of the New York Academy of Sciences that the loss of dream life was an additional and underappreciated “silent epidemic,” contributing to both physical and mental problems. “We are at least as dream-deprived as we are sleep-deprived,” he wrote.

Antonio Zadra, a psychology professor at the University of Montreal and researcher at the Center for Advanced Research in Sleep Medicine, told me that the usual state of the world — in which busy, sleep-deprived people fail to have, or fail to notice, dreams — is a bit as if we all lived in a bright city, unable to see beyond the artificial lights that surround us. Then came the arrival of lockdowns, and suddenly many people weren’t setting alarms. They were sleeping longer and sleeping later, rather than rushing to the coffeepot or shower. (Studies also showed that they were more likely to wake up during the night from stress, which can increase recall of the dreams we have before morning.) When people on lockdown noticed a sudden surge in dreaming, Zadra says, it was “as if a catastrophic event put out all the outdoor lights, and people were amazed to see so many stars.”

It’s a lovely analogy. But also: So what? Did it matter to anyone’s waking life that dream life, or at least awareness of it, had expanded? In his book “Why We Sleep,” the neuroscientist Matthew Walker half-jokes that dreams are a time when everyone on Earth becomes “flagrantly psychotic,” experiencing hallucinations, delusions, disorientation, emotional lability and amnesia. We see things that aren’t there, believe things that simply can’t be true, don’t know where or with whom we are, have wild mood swings and then promptly forget almost everything we experienced. My partner’s most vivid dream in the early, scary days of the pandemic involved our hosting a party to which his former roommate brought a dish called “carrot balls”: carrots scooped up like melon balls and served, floating in water, inside an old shoe. What could this possibly have gained him?

For the bulk of human history, dreams were understood to be full of messages and portents. In many cultures and texts, from the Epic of Gilgamesh to the Old Testament to the Tibetan Book of the Dead, they were seen as omens or communications from gods, ancestors or spirits, telling us things we couldn’t otherwise know. Dreams predicted the birth of the Buddha; guided Jacob, Daniel and Joseph; and provided sacred knowledge to members of Indigenous communities around the world.

In the mid-19th century, dreams started to be understood as creations of the brain, something that might be studied scientifically. Alfred Maury tested whether sensory stimuli like smells and sounds, experienced while sleeping, could affect his dreams (they could). Mary Whiton Calkins pioneered practices like collecting and statistically analyzing dream reports, or waking research subjects systematically at different times of the night to understand how their dreams changed. Sante de Sanctis compared the dreams of different groups, like children and the elderly or the mentally ill and the healthy. Some psychologists now frame the Freudian approach to dream analysis — Freud’s theory that all dreams are elaborately coded versions of wish fulfillment, and the subsequent wave of interest in so-called dream interpretation, with certain elements having fixed, identifiable meanings — as an interlude of sorts, a departure and distraction from a century of scientific advancement in dream studies.

The field expanded further with a discovery in the mid-20th century. A graduate student named Eugene Aserinsky, who was monitoring his sleeping son with electrodes, was surprised to find that the boy’s eyes began to move as if he were awake. Aserinsky had discovered REM sleep, which researchers soon went on to link to periods of vivid, narrative dreams. As Zadra and Robert Stickgold, a professor of psychiatry at Harvard Medical School and director of the Center for Sleep and Cognition in Boston, wrote in their book “When Brains Dream”: “No longer was dreaming just a mystical mental phenomenon that seemed to come from nowhere except maybe the hidden resources of our psyche. Suddenly, there was a biology of dreaming.”

By the late 1970s, two Harvard psychiatrists, John Allan Hobson and Robert McCarley, were arguing that it was foolish to begin a theory of dreaming in the way that humans always had: by starting with actual dreams and their content. Dreams, they argued, were so bizarre and so subjective, with their frustrating ephemerality and their weird story lines and their odd non sequiturs, that they defied logical study. Instead, researchers should consider physiological brain changes during sleep, and then work backward to try to understand the psychology of dreaming humans — which, they posited, was probably just the forebrain “making the best of a bad job in producing even partially coherent dream imagery from the relatively noisy signals sent up to it from the brainstem.”

Dreams, in other words, were just a neurological epiphenomenon left over from the meaningful parts of sleep — a side effect, an accident of biology, much like the way an incandescent light bulb designed to create light also happens to put out heat. They carried vestigial echoes of what happened during waking life, but they had no lasting impact, no purpose of their own. If the dream-heeders of old were like Romeo, primed to believe in portents, then the new theory was Mercutio, cautioning him not to listen to “the children of an idle brain,/Begot of nothing but vain fantasy,/Which is as thin of substance as the air/And more inconstant than the wind.”

Hobson and McCarley’s theory, known as the activation-synthesis hypothesis, has not been disproved, and its adherents still expand and refine it regularly; it may very well be correct in asserting that dreams are all heat and no light. But it is also balanced by a countertrend in dream studies, one that’s characterized by a rush of theories and models that rely on our expanding knowledge of the dreaming brain to make different versions of the opposite argument: that dreams have a genuine, biological and deeply important purpose to play in human life. Every year, more and more theories of this sort (so many that several dream researchers I spoke to sighed or bristled at their quantity or quality) are added to the pile. This fervent debate — do dreams themselves even matter, and if so, why? — is, the British research psychologist Mark Blagrove told me mildly, “the really big issue in dream research.”

One thing that everyone agrees on is that sleep, and especially REM sleep, does matter. For one thing, evolution wouldn’t have favored such a dangerous activity — in which we are disconnected from reality, sitting ducks for accidents or predators — if it weren’t deeply helpful for survival. It can’t be an accident that so many animals, including humans, devote enormous chunks of their lives to sleeping. In fact, science has yet to discover an animal that doesn’t sleep at all. (One outlier is a 1967 study that suggested that bullfrogs don’t sleep; it is now considered to have been flawed.) Migrating birds and swimming dolphins manage to sleep while on the move by resting one hemisphere of their brains at a time. Sitting ducks do this, too — they take turns on guard duty. There’s also a less successful version of the phenomenon in humans, known as the “first-night effect,” which occurs when the left hemisphere of our brains refuses to fully rest when we’re sleeping in a new, uncertain environment for the first time, causing us to wake up tired. Even jellyfish sleep, despite not having brains, and earthworms that don’t get a chance to sleep for several hours after experiencing a stressful event, like extreme heat, cold or exposure to toxins, are less likely to survive. One study, using a magnetic device called the insominator, tested the effects of sleep deprivation on honeybees and found that it made them bad at communicating with the rest of their hive. Another found that rats deprived of all sleep will be dead within a month.

In humans, shorter sleep is associated with heart disease, obesity, stroke and Alzheimer’s, and various studies have suggested why: Sleep is when the brain does much of its “housekeeping,” allowing our bodies to secrete growth hormone, to produce antibodies and regulate insulin levels and to repair neural cells and remove waste proteins that build up in our brains. It’s also critical to lots of intellectual and emotional processing; without enough sleep, it’s harder for us to learn new things, evaluate threats, deal with change and generally control our emotions and behavior.

Still, none of that means that the dreams that happen during sleep — their content or even their existence — are meaningful in their own right. As Zadra explained to me, “Sleep could do all its stuff without us having these virtual simulations,” these elaborate narratives unfolding inside our heads every night. Anyone making the case that dreams matter, therefore, has to grapple with that fundamental question of content. Is there a point to spending our nights inside strange, phantasmagoric stories that we rarely even remember the next day?

Within a week of her library dream, Barrett posted a survey online. Along with basic information about the dreamers who filled it out — where they lived, whether they worked in health care, if they had been sick — she gave people the space to describe any recent dreams they believed to be about the pandemic. In many, the connection was obvious: dreams of working in an I.C.U. or getting a positive Covid test or hiding from disease. (Barrett was collecting dreams in English, which, she acknowledges, created biases in the data, as did self-selection by participants who — presumably — cared about the pandemic, had an interest in dreams and consumed the sorts of news media that might point them toward her work.) Other dreams were more metaphorical but still offered intuitive connections, the kind of transference of emotions that dream researchers are used to identifying. A common dream of this type involved monsters lurking just out of sight, or invisibly attacking the people around them; in one dream, the invisible monster could kill only people who were within six feet of its most recent victim. Barrett also noticed a surge in bug imagery, often scary swarms of insects, which she chalked up to the dreaming mind searching for visual representations to match the fear it felt, and landing on a pun — a virus, after all, is known as a bug.

Still other supposed connections to the pandemic, though intuited by the dreamer, were not clear to Barrett. (For example: a dream in which Oprah Winfrey threatened a gymnasium full of people with a hand-held circular saw.) But many people took pains to explain the connections that they saw in their own dreams, like when a bat entered a dreamer’s house and the dreamer used a thick copy of The Washington Post to swat it. The fear, during the dream, was of rabies, but waking up brought instant recognition that bats were also a possible source of the virus that causes Covid-19. The dreamer speculated that the dream “perhaps symbolizes the need to arm oneself with information, data and knowledge to protect against an invisible virus quickly circulating way too close to home.”

Some days dreams arrived by the hundreds, and it took Barrett hours just to read through them all. She began to note themes and similarities, which she later explored through statistical and linguistic analysis. Women, who according to other studies experienced more job loss and more pandemic stress than men, also saw their dreams change more: Their levels of anxiety, sadness and anger were much higher than the prepandemic dreams with which Barrett compared her new sample. (Women also had most of the anxiety dreams about home-schooling.) And the dreams of the sick, as is common when the body is fighting a fever, were the most bizarre and yet the most verisimilar of all — vivid-but-strange hallucinations that made it difficult to separate sleep from waking life. A Covid patient named Peter Fisk described feeling wide awake, curled up in bed and thinking fondly back to his days of living in a cozy den in a riverbank. “But then,” he wrote, “it occurred to me that I had never actually done that. I was having false memories of being an otter.”

As was the case with post-9/11 dreams, the most affected dreamers were those living closest to trauma. More than 600 health care workers sent in dreams, which Barrett recognized as often the same story, told with small variations: “There’s a critically ill patient in their care, something is not working and the patient is dying. They feel desperately responsible and yet have no control over death.” Research has shown that the dreams of trauma victims often start by replaying the traumatic event in great detail, but over time they often incorporate more and more new elements and story lines, blunting the emotion of the original dream. (Some therapists encourage this evolution, coaching patients to imagine, and then to try to dream, more empowering endings to their traumas.) In cases of post-traumatic stress disorder, however, this process seems to break down; the classic PTSD nightmare is a realistic, flashbacklike trauma that repeats again and again with few alterations.

As the pandemic progressed, dreams about the illness itself began to be mixed with dreams about its secondary effects, especially life under lockdown. These dreams, too, were broadly similar across countries, except that they came at different times, as various countries imposed and lifted restrictions on movement. What divided such dreams, Barrett found, was less national than situational. People quarantining alone often dreamed what she calls “exaggerated scenarios of isolation, of abandonment”: being held in prison or marooned on a spaceship. One person was sent on a solo trip to Mars for which she had decidedly not volunteered. In contrast, people who found themselves stuck inside with what suddenly felt like too many people had the opposite dream: losing control of their homes to crowds.

For Barrett, the reports offered a strange yet deeply intimate window into a varied and evolving pandemic. While health care workers in Italy were suffering through realistic nightmares about intubations gone wrong, those in the United States were often dreaming about creepy threats that were approaching but had not yet arrived. While people in China dreamed about masks right away, it took masking a long time to show up in American dreams. When it finally did, the dreams were often about sudden realizations that the dreamers had forgotten to wear a mask, or were wearing it improperly, followed by fear that they had exposed themselves or others to the virus. By the fall, though, the reaction was more likely to be embarrassment: What would people think of the dreamer? “It’s sort of starting to replace the classic ‘naked in public’ dream,” Barrett says.

As time passed and the pandemic dragged on, Barrett noticed more dreams that she described as post-apocalyptic, often involving small groups of survivors living in changed, dangerous worlds. “I dreamed of never returning to life-as-normal, of being old and quarantined with my future grandchildren,” one dreamer wrote.

“I dreamed that all of humanity was slowly blinking out of existence,” another reported, “preserved only by me forcefully willing everyone to stay by remembering them.”

One of the first known dream dictionaries was written 3,100 years ago, in ancient China. Its author, was a member of the Zhou dynasty; its contents, which are organized around thematic elements in dreams (dreams about the sun, moon and stars come before those about shoes, socks and clothes), would be recognizable to anyone who bought a modern dream guide in which dream motifs stand in for a supposed deeper meaning. If wind blows your clothes in a dream, the book advises, it means a disease is coming for you.

Erin Wamsley, a research psychologist who studies dreams at the Sleep Laboratory at Furman University in South Carolina, is quick to tell strangers who ask about her work that no, she will not interpret their dreams for them, because no, that is not how dreams work. When I asked Wamsley why this view of dreams as code has been so entrenched for so long, she pointed to the tendency to see dreams as cryptic messages from the supernatural, which arose independently in many cultures, long before Freud’s theories cast their long shadow. Dreams feel so strange and nonsequitous and generally beyond our own control, she said, that “it can seem that whatever happens inside them must come from somewhere else.”

To the contrary, though, dream studies have shown that a great deal of what we dream about derives from our waking lives. We dream about specific experiences we’ve had, our own ongoing concerns and anxieties, real people and places that we actually know. We’ve been the god in the machine all along.

It’s no surprise, then, that the fears and novel experiences of the pandemic would emerge in our dreams. And they did, in study after study. While Barrett’s study was restricted to dreamers’ self-selected pandemic dreams — she didn’t have access to each dreamer’s full dream life, let alone what they were doing and thinking during waking life — other studies tried to take a broader view. A study of 796 Italians in April and May of 2020 found that 20 percent of the time, their most recent dreams explicitly referenced Covid-19. Dream reports from Brazilian adults in lockdown were unusually full of words related to anger and sadness, contamination and cleanliness. Nielsen, who with his lab in Montreal collected and analyzed a large sample of pandemic dreams, including descriptions of dreams that people posted on Twitter, found that dreams about navigating new relationships with crowds and social distancing — often with a sudden and alarming realization that what started out seeming fun and normal was no longer safe — were extremely common. When Nielsen compared pandemic dreams with a prepandemic sample, he found a marked increase in people being so startled or unmoored by the dissonance within their dreams that they woke themselves up.

A study in China found that those who were the most distressed in waking life reported more dreams about the pandemic. In a Finnish study, algorithmic linguistic analysis showed that more than half of all dreams contained pandemic-specific content, like worries about social distancing, contamination or elderly people in trouble. Dream life and waking life were so intertwined that some researchers began to wonder if media coverage of pandemic dreams — and especially of specific motifs, like the swarms of insects that Barrett often mentioned — was creating a feedback loop, causing people to dream the dreams they had read about.

What dream theorists are still trying to understand, though, is why or how dreams remix elements that they pull from our memories and experiences into brand-new, complexly rendered stories full of unexpected details and events. Research has shown, for example, that about half the people we interact with in dreams are real people we know personally, but the other half are composites, inventions or generic roles, like police officers or librarians. (The psychologist Kelly Bulkeley has argued that people today dream about celebrities in the same way ancient Greeks dreamed about gods and goddesses: These are simply the characters whose stories live in our heads.) And though many dream settings feel familiar, only a third of them are places we fully recognize. Elements of recent life may appear, but only some of them, and not replayed just as they happened. Instead, they’re jumbled together alongside novel connections and fragments of older memories and reconstructed into brand-new narratives. Nielsen describes dreams’ dramatic reconstitution of the waking world as “experience that has been put through a chopper and reassembled with several ingredients added.” But how and why does this reassembly happen?

Wamsley, who studies the way dreams interact with memories, notes that different memory sources are so mixed up in dreams that the shock of different life stages coexisting in them can sometimes cause a sort of temporal whiplash — like when you’re dreaming about being with an ex and then remember that you’re married, or facing some problem in elementary school before recalling that, wait a minute, you’re 40. Even lucid dreamers, people who practice becoming aware that they are dreaming in order to take some measure of control over what happens in their dreams — choosing to fly or to travel to Paris, for example — are consistently confronted by how little they can control. They can choose a dream location but not the details of what it will look like, can make a dream character appear but not predict what the person will say. The mathematician Robert Wayne Thomason once had a breakthrough on a problem after his friend, Thomas Trobaugh, made a key suggestion to him in a dream. Though Trobaugh died by suicide three months before the dream occurred, Thomason argued that it was only fair to list him as a co-author on the paper.

It’s “mind-boggling,” Zadra told me, how much our conscious minds are shut out of the unconscious process of dream creation. “Anytime a character says something that surprises you, you’re really surprising yourself.”

Barrett eventually collected more than 15,000 dreams. When we first spoke, in June 2021, vaccine access was spreading. The torrent of dreams, which had shifted to include more optimistic dreams, as well as dreams about back-to-school or back-to-work anxiety, had begun to slow. Still, not a single day had passed without at least a few people wanting scientists to know what the pandemic had done to their dreams — hoping, presumably, that someone could tell them why.

Then came the second pandemic summer, and the rapid spread of the Delta variant. People began sending in dreams full of fear, reminding Barrett of the dream reports from the early days, as well as dreams that seemed to capture their disorientation amid the pandemic’s evolving risks. In one, the dreamer decided to go on vacation to a beach that had once had a problem with deadly hornets and been declared safe. But the hornets attacked after all.

Barrett viewed the dreams that people sent her about forgetting masks or accidentally going to parties as a way of “practicing” new behaviors or the navigation of new realities. (A generous reading of the carrot-ball dream casts it in this light: Most of the dream was about hosting a small party that, as if in warning, grew rapidly out of control.) This interpretation is consistent with a pair of theories put forward by the Finnish researcher Antti Revonsuo: the threat-simulation theory and its partner, the social-simulation theory. These hold that dreams developed as an evolutionary defense mechanism and that they offer a chance to practice our responses to threatening situations as well as social ones — because sociability is key to primate survival — in a sort of virtual reality. In Finland, Katja Valli, a researcher at the University of Turku, is analyzing pandemic dreams to see whether they support the hypothesis of dreams as threat simulation. To better contextualize the dreams, she also asked study participants to record their “mind-wandering” while awake and various details about their mental well-being. Early results, not yet published, show dreamers facing many more threats than they did before the pandemic.

There are a host of other theories that might be applied to pandemic dreams. Blagrove, the British psychologist, has a theory that accepts that dreams themselves are vestigial echoes of the mind-wandering mode in which our waking brains spend much of their time, but suggests that they nevertheless have an evolutionary function: promoting bonding and empathy when people share them with one another. (He came up with this theory after he and the artist Julia Lockheart began hosting online events in which they explored health care workers’ pandemic dreams.) Ernest Hartmann, a psychiatrist who was the founding editor of the journal Dreaming, has proposed that dreams, like therapy, offer a safe place to process memories, especially traumatic ones, allowing us to store memories without being overwhelmed by the strong emotions they carry. The psychologist Rosalind Cartwright, in a series of pioneering studies, found that in a group of depressed patients going through divorces, those who dreamed about the failing relationship were the ones who later recovered from their depression. “Dreams are a natural healer,” she wrote. “They work during sleep in the same way a good psychotherapist does, by relating the new to older patterns of problem-solving that have gotten us through bad spots in the past.”

The difficulty of isolating the effects of dreams has always made theories of dream function hard to test empirically. Pandemic dreams are subject to these old problems: isolating the effects of dreaming from those of sleep, accessing an experience that’s purely subjective and quickly forgotten. Zadra told me that he believes the value of a dream is not in remembering it later, as you might a movie that you watched. Instead, what matters is creating it and experiencing it — feeling emotions, having reactions to people and events, watching how dream characters and situations react to you — and then waking up a slightly different person, whether the conscious memory of the dream stays with you or not.

Researchers are always working to design more targeted studies in which they are better able to isolate various confounding factors. About 20 years ago Stickgold and some colleagues at Harvard and the Massachusetts Mental Health Center designed a now-famous study. It showed that people who played Tetris, dreamed and then played again improved more than those who didn’t get a chance to dream — and that they often dreamed images of the game. Even more notable was Stickgold’s inclusion of a small group of people who had lost their short-term memory to brain lesions. Members of this group soon forgot that they had played Tetris, or even what Tetris was, but they still reported dreams about arranging falling shapes.

Since then, many more studies have tested the relationship between dreams and learning, including a recent effort by Nielsen to use associative sounds to trigger dreams about a specific task. We know that people often dream about the new things that they are learning, and that those who do so can often perform a task better after it appears in their dreams. (Some researchers point out that those who were the worst at the task to begin with tend to be the ones who dream about it and then improve: Does dreaming itself help them, or does it simply reflect that their brains continue to work on a problem that vexed them?)

Some researchers now theorize that dreams are central to sleep’s role in helping us move memories from short- to long-term storage, a process that involves choosing which memories to preserve, which to forget and how to store them. In one study, people who were asked to memorize a list of words, and then to recite them after spending time in REM sleep, were worse at remembering the original words but had replaced them with “gist” words that grasped the patterns that connected them. Infants, who spend more than half of their already substantial time asleep in the REM phase, seem to need the sleep to help process new memories and knowledge. In one study of 15-month-olds, only those who’d had a chance to nap shortly after hearing a simple, invented language could make sense of its rules and patterns, noticing when its “grammar” was right or wrong in new uses. During the pandemic, Erik Hoel, a professor of biology at Tufts University, drew on the field of machine learning to posit that human brains, like computers, can get stuck in analytic ruts when there’s not enough new stimulus to test their theories against, and that the most bizarre and nonsensical elements of dreams keep them from being “overfitted,” or unable to make sense of new information.

When REM sleep was first discovered, psychologists called it “paradoxical sleep,” as the brain’s electrical activity during REM sleep looks more like the waking brain than it does like the brain during other sleep stages. But researchers now know that the dreaming brain is tuned quite differently. Certain key neurotransmitters (like noradrenaline, which surges at times of stress) are suppressed during REM sleep, while others (like dopamine and acetylcholine, associated with emotion and memory) are increased; regions associated with visual stimulation, movement, autobiographical memory and emotion all become very active, while regions that engage in logic, decision-making and impulse control are suppressed. The brain also seems to show a preference for abstraction, novelty and hyperassociativeness that it doesn’t show while it’s awake: Research subjects woken from REM sleep and asked to solve word puzzles do so faster than when fully awake, apparently with less careful deliberation and more intuition, and are more likely to favor weaker, more distant and less obvious associations between concepts. “No longer are we constrained to see the most typical and plainly obvious connections between memory units,” Walker explains. “On the contrary, the brain becomes actively biased toward seeking out the most distant, nonobvious links between sets of information.”

These sometimes wild leaps, Zadra and Stickgold argue, account for the bizarreness of our dream worlds. They would not happen “in the glare of day, when our brains are dealing primarily with new incoming sensations and the balance of neurotransmitters in our brain is optimized for processing the here and now.” While that kind of optimization serves an obvious purpose, they believe that the opposite mode is necessary, too. During the day, our brains are busy sorting and responding to what’s real and immediate, but at night their job is to wander as widely as they can.

“For every two hours we spend awake, taking in new information,” they write, “it appears that the brain needs to shut down all external inputs for an hour to make time to figure out what it all means.”

When young children are asked where their dreams are taking place, they often give answers such as “in my bedroom” or “in front of my eyes.” It typically takes years of cognitive development for them to understand that what they are seeing and experiencing is really happening inside their heads, separate from reality, and that it is invisible to other people. We are born, in other words, believing in dreams, and we unlearn that belief in order to be able to live in the world.

But of course, we don’t fully unlearn it. We insist on flailing through morning recitations of nonsensical and half-remembered nocturnal adventures. Sometimes we continue to feel residual emotions from dreams — nursing a grudge against someone who did nothing wrong in real life, for example — long after we’ve forgotten the dream itself. We support a whole industry of dream dictionaries.

Even John Allan Hobson, who helped originate the theory that dreams are a result of accidental neural firings, was fascinated by the particulars of his own dreams. He kept decades of detailed dream diaries — his favorite, he once said, was about running so lightly through the Swiss Alps that he was nearly flying — and published a book about 13 of his own dreams, in which he analyzed the particular confluence of life events and brain stimulation that might have brought them into being. Eventually, he began to argue that dreams could be both random and deeply meaningful: that they enabled the development of waking consciousness itself.

When Barrett took phone calls from reporters about her sample of pandemic dreams, she noticed that many seemed to have a personal motive as well as an educational one: One after another, they described dreams about masks or parties or disease or isolation. Some simply seemed curious about how the dreams fit into broader trends, while others seemed to see their dreams the same way that uncountable generations before them had: as mysterious messages, sent either by an outside force or their own subconscious, waiting to be decoded by the right translator.

Barrett tried to let them down easy. She had always seen dreams as both more complicated, and simpler, than the theories that swirled around them. “We’d never ask, ‘What is waking thought for?’” she told me. “It’s for everything.” She sees dreams as essentially another way of thinking, one governed by a brain state with its own set of rules and possibilities. “Dreaming is, above all, a time when the unheard parts of ourselves are allowed to speak,” she once wrote. “We would do well to listen.”

Shortly after her dream about the library, Barrett had another dream that she recognized as being about the pandemic. In it, she took her cat, Morpheus, outside through a cloud of toxic air, both of them clad in protective hoods that looked a bit like gas masks. After that, for a very long time, she didn’t dream about the pandemic in any recognizable or memorable way, even as she spent her days immersed in the intimate and often fearful dream lives of others.

And then, after months, she had another dream. She was holding a device like a phone, but it had an app on it that generated glowing, holographic projections of other people’s dreams. She flipped through them, one by one, and each projection showed a still image that somehow conveyed all the detail and depth of emotion that the dream carried with it. Here was a 1950s living room; here was a doglike creature with big, sharp teeth. As each image surged out of the phone, Barrett found herself overcome with concern for its dreamer, but unable to do anything to help.

She flipped to a sickly yellow monster with a head like a coronavirus, a crown of spike proteins on a stylized and repulsive humanoid body. On an impulse, she reached out to hug it, taking its body in her arms. The hug, to her surprise, felt as though it went right past the monster. She had finally managed, in a dream, to do what her waking self had been longing to do: reach through the dream and find the person who was dreaming it.

Brooke Jarvis is a contributing writer for the magazine. She last wrote a feature about Covid-19 and the science of smell. Amandine Urruty is an illustrator in Paris known for her whimsically macabre drawings.