In the field of sleep and dreams, these are promising times. But there's been no year more momentous than 1953. Until then, scientists had equated sleep with flicking off a desk lamp. For more than two decades they'd been able to record brain activity in sleep, but the feeling was, why bother? Why waste reams of costly graph paper making electroencephalogram recordings of what was thought to be a neurological desert? With no strong expectation of finding otherwise, University of Chicago researchers Eugene Aserinski and Nathaniel Kleitman decided it was worth doing, monitoring 10 subjects in a laboratory. Their findings turned our understanding of the sleeping brain upside down. What they discovered was a sleep state in which the brain is, in many ways, every bit as active as when it's awake; a state in which, compared with other stages of sleep, the heart beats faster, breathing quickens, blood pressure and blood flow to the brain (and sexual organs) rise, while the eyes move rapidly beneath their lids. Brain waves are low-voltage and high-frequencythe opposite to the brain waves of deep sleep, more like what goes on when a person is awake, thinking and talking. Awoken from this paradoxical state that Aserinski and Kleitman called Rapid Eye Movement (REM) sleep, subjects could usually recall vivid dreams. In a single swoop, the pair had not only uncovered what many regard as a third state of consciousness, but raised expectations that the mysteries of how and why we dream might soon be solved. The quest for answers has been hindered by doubt: is dreaming a mystery worth solving? Science has long had an uneasy relationship with our nocturnal imaginings. While some brilliant practitioners have workedand do workin the field, its links with mysticism and Freudian theory have repelled others like a bad odor. Everybody dreams and most people talk about theirs now and again. But once, as children, we learn to distinguish these delusions from reality, dreams usually become no more than a sideshow, sometimes disturbing, occasionally poignant, but mostly something to be forgotten, quickly and completely, if they were remembered in the first place. It's easy, in other words, to lose sight of what a remarkable phenomenon dreaming is. Every night, devoid of external sensory stimulation, our brains screen internally-projected films concocted from pieces of our own thoughts. Nearly always in the lead role, we flee from danger, triumph and flop in our areas of endeavor and enjoy passionate encounters with people we yearn for or hardly know. We do these things and countless others not in a state of detachment but rather, despite the bizarre distortions typical of dreams, convinced the events are real and with our emotions and senses engaged. That the movies we watch in theaters were so engrossing. The discovery of REM sleep was supposed to be an opening through which science could throw light on dreaming. And it was, up to a point. But it has also been an obstacle to progress. What happened was that many scientists became convinced that REM sleep and dreaming are more or less the same thingthat they're underpinned by the same physiological processes. REM became known as "dreaming sleep," but that is almost certainly an oversimplification. "Sadly, there's an authoritative school of thought that will not let go of this idea," says Mark Solms, professor of neuropsychology at the University of Cape Town, South Africa. "As a discipline we've been lost because we've been conflating the two." There are tensions in the field, but "it's starting to break out of its malaise," says Robert Stickgold, associate professor of psychiatry at Harvard Medical School's Center for Sleep and Cognition. Recently, scientists have abandoned stagnant theories of dreaming and postulated new and intriguing ones, with experiments underway in various parts of the world aimed at establishing the function of our nightly hallucinations. If recent work suggests anything, it's that there is such a function, or more than one, and that dreams aren't just neural waste. They may improve the quality of our sleep. They may prepare us for danger. They may embed memories. There are more theories than answers. But after years in the doldrums, dream research is moving forward again. for all the recent advances in brain science, it shouldn't surprise that the riddle of dreaming hasn't been cracked. "We still don't know why we sleep, let alone why we dream," says Dorothy Bruck, professor of psychology at Melbourne's Victoria University. It seems commonsensical that sleep is a restorative phase for brain and body, and there's some evidence that the effects of sleep deprivation are the result of minor brain damage that would normally be repaired when we're asleep. But despite their best efforts, scientists have been unable to pinpoint what's going on in sleep that makes it essential. Still, sleep research is a breeze compared to studying dreaming. In the former, you can at least be certain you're observing a sleeping person. "But we don't have a device that looks into a person's head and sees dreams happening in real time," says Russell Conduit, a lecturer in the school of psychology, psychiatry and psychological medicine at Melbourne's Monash University. Instead, dream researchers rely on what he calls the "faulty methodology" of waking subjects and asking them what was going on in their heads immediately before they were woken. But because certain parts of the brain are switched off during sleep, it shouldn't be assumed that subjects' answers will be accurate. They may have been having a dream but simply weren't paying attention, can't remember it, or both. "If you took a lot of the dream research to a physicist," says Conduit, "they'd laugh at you." Before delving into the latest theories, then, it's worth reviewing what we know about dreaming and the sleep state in which we seem to do most of it. REM follows four stages of sleep known collectively as non-REM sleep, in which brain activity becomes progressively more subdued. In REMwhich occurs four or five times a night and lasts about 30 minutes at a timeour muscles become paralyzed, which could be a mechanism for preventing the acting out of our dreams. About half of a baby's sleep is REM compared to a quarter of an adult's. The body craves it: when cut short in experiments REM returns more insistently in subsequent periods of sleep. Neuroimaging has told us a lot about the dreamingor at least the REMbrain. Those areas that are active include the brain stem (responsible for basic functions like heartbeat regulation), the limbic system (which mediates emotions, learning and memory) and parts of the forebrain involved in processing sensory information. Shut down, meanwhile, are the bits responsible for the most sophisticated mental processes, such as logical and ordered thought. It's a profile that fits neatly with the subjective experience of dreaming: vivid images, strong emotion and snippets of memory, but a shortage of coherence. Rather like a David Lynch film. All mammals experience REM, but some have more than others. It appears the less mature a species is at birth, the more REM it has. Assuming for a moment that REM equals dreaming, the opossum and armadillo are among the most prolific dreamers, while dolphins do very little of it; humans are in the middle. Research suggests our dreams become more complex as our mental abilities develop. The dreams of very small children don't just seem prosaic because tikes lack the eloquence to bring them to life in the retellingthey really are prosaic. Two-thirds of dreams are almost exclusively visual, a quarter feature sound and a smaller fraction smell and taste. Nine out of 10 contain emotion, most commonly mild anxiety or frustration. Our dreams tend not to be reproductions of past events but rather, according to research led by Tore Nielsen, director of the Dream & Nightmare Laboratory at Sacré-Coeur Hospital in Montreal, reinterpretations of events that happened at two distinct time periods: yesterday and about a week ago. We also dream about upcoming events and conversations we'd like to have. Other times the content is so weird who knows where it comes from? Recent developments in dream research won't make sense without first touching on the academic thunderbolt of 1977, when a paper by two Harvard neurophysiologists, Allan Hobson and Robert McCarley, ran in the American Journal of Psychiatry. At the time, Sigmund Freud's theory of dreams (which holds, in part, that dreams preserve sleep by distracting the brain with reflections of the unconscious) was a pillar of psychiatry. In The Brain as a Dream State Generator: An Activation-Synthesis Hypothesis of the Dream Process, the Harvard pair challenged Freudian theory on virtually every point. They argued that dreams are nonsense created when the forebrain makes "the best of a bad job in producing even partially coherent dream imagery from the relatively noisy signals" sent up to it from the brain stem at the onset of REM. Their paper served to yank dreaming from the realms of the psychological and plonk it in a dreary, physiological bucket. Later, the English molecular biologist Francis Crick, a co-discoverer in the 1950s of the structure of the DNA molecule, drained a little more romance from dreaming. His and theoretical biologist Graeme Mitchison's "reverse learning" theory held that dreams rid the brain of superfluous notions, and that without this regular flushing brain overload would manifest as hallucinations and obsessions. There are echoes of this idea in the perspective of Drew Dawson, director of the University of South Australia's Centre for Sleep Research: "I tend to think of dreaming as a bit like backwashing the swimming pool filter." While hugely influential, Hobson and McCarley's Activation-Synthesis model attracted hordes of critics, who protested that many dreams aren't merely cognitive fragments nor a succession of chaotic images, but so story-like, sequential and dramatic that the thinking brain must surely have played a more substantial role in their production than the last-minute editing of a pile of neural bloopers. And there's the matter of lucid dreaming, in which people become aware in the course of a dream that they are, in fact, dreaming, and are able to control the course of eventsa phenomenon that strengthens the case for higher-brain involvement in dream construction. The lucid dreamer can apparently apply certain techniques to prolong the dream and take it in delightful directions. "The experiences are so convincing," says Victoria University's Bruck, "it seems as if another level of reality exists." But the wrecking job on the notion that dreams are a random by-product of REM sleep was carried out by the South African neuroscientist and psychoanalyst Solms, who was working at the Royal London Hospital in the 1990s when he made his career-defining discoveries. Solms wasn't alone at the time in realizing that dreaming occurred outside periods of REM, that it was also common at sleep onset and shortly before waking in the morning. But he found an even weaker spot in the Hobson-McCarley hypothesis. If their theory was right, then people with damage to a part of the brain stem called the ponsthe on-off switch for REM sleepshouldn't be having dreams. Solms, however, had five patients with lesions in precisely that region, and while they weren't having REM, they were nonetheless reporting dreams. Even more interesting to Solms were 53 Royal London Hospital patients with healthy brain stems who said they'd stopped dreaming. Most of them had damage to the part of the brain that generates spatial imagery. That made sense: if you can't create pictures in your mind, how are you going to dream? It was the circumstances of the remaining nine patients that fascinated Solms. They had damage to the white matter of the ventromesial quadrant of the frontal lobes, an area linked to the transmission of the chemical dopamine and crucially involved in motivation, urges and cravings. These patients still experienced REM sleep but reported having lost both the capacity to dream and all sense of spontaneity, drive and love of life. They did what they were told and that was about it. It seemed to Solms that dreams must themselves be associated with driving urgesa very Freudian takebut he needed more evidence in the form of more people with lesions in this particular spot. Nowadays, damage to that part of the brain is rare, normally a result of strokes or tumors. But it was a lot more common in the '50s and '60s when some mental illnesses were treated by removing it in an operation called a prefrontal leukotomy. Solms waded through the literature and found hundreds of case studies in which the effects of this procedure were described. To his amazement, reported loss of dreaming was one of them. "So I thought I'd discovered something new," says Solms, "but it turned out to be something we'd documented long ago but had forgotten." In the field of dreams, however, his findings were no less significant for that: Solms had shown REM and dreaming to be dissociable states and produced a compelling case that the higher brain has the central role in dream creation. Solms, who believes science is getting closer to answering the key questions about dreaming, is leading two studies at the University of Cape Town with that goal in mind. One involves using functional magnetic resonance imaging to try to disentangle the REM brain from the dreaming brain. He wants to obtain images of the dreaming/non-dreaming brain at sleep onset and note the differences between the two. "If we can image what's going on at that point, then we'll get a clear handle on what mechanisms are important for dreaming as opposed to REM sleep," says Solms, who predicts a key role for the motivational part of the brain. His other study involves comparing the sleeping ability of subjects who dream with that of subjects who can't because of brain lesions. Solms argues that because our motivational drive is fully active while we're asleep, our brain's way of keeping us asleep and undertaking the necessary repairs is by tricking us, through dreams, into thinking we're up and about and pursuing our desires. It's the neurological equivalent of putting on a DVD for the kids so the main players in the house can get some shut-eye. "Dreams replace the real actions that are instigated by our motivational impulses while we're awake," Solms says. His hunch is that the non-dreamers in his study will wake up more often during the night than the dreamers, especially during REM sleep: "The dream," he says, "is what keeps you asleep." Others approach dreams from a different angle. An argument that resonates with many is that whatever the explanation for dreaming, humans must do it for the same reason that all mammals have done it for more than 100 million years; any theory must make as much sense when applied to a rabbit as to a person. In the same Darwinian vein, sleeping and dreaming must serve important functions because they're vulnerable states and natural selection would have eliminated them if they didn't provide compensating benefits. In the ancestral environment, human life was short and perilous; ever-lurking predators threatened survival and reproductive success. The biological function of dreaming, argues Antti Revonsuo, professor of psychology at the University of Turku, Finland, is to simulate threatening events so to prepare the dreamer for recognizing and avoiding danger. The threat-simulation theory, first presented in 2000, "is built on the actual empirical evidence we have concerning the content of dreams," Revonsuo says. "It's surprising how many theories of dreaming there are that are not based on any systematic review of the evidence." He cites studies showing that, typically, dreams are too seldom sweet, and that negative feelings, dangerous scenarios and aggression are over-represented. Based on ongoing work with PhD student Katja Valli, Revonsuo estimates that the average "non-traumatized" young adult has, conservatively, 300 threat-simulation dreams a year. In the dreams of both men and women, male strangers and wild animals are most often the enemy, and the dreamer's typical responses are running and hiding, often in a state of terror. If dreams are biased toward simulating ancestral threats, the traces of these biases would be strongest early in life, before the brain has adjusted to the realities of the contemporary environment. Sure enough, Revonsuo says, research shows that animals make up about 30% of all characters in the dreams of children aged 2-6 compared to 5% of adults'. True, the animals of children's dreams are often fluffy and harmless, but almost half the time they're frightening creaturessnakes, bears, lions, gorillasthat children would seldom, if ever, have encountered in waking life. The reason we don't dream about reading and writing, Revonsuo speculates in his original paper, is not because these activities don't engage our emotions but because they're "cultural latecomers that have [yet] to be effortlessly hammered into our evolved cognitive architecture." Revonsuo knows that his theory pleases neither Freudians nor neuroscientists. "If, for the supporters of the psychological theories, it grants too little meaning to dreams," he says, "for the supporters of neurophysiological random-noise theories, it grants dreaming far too much." Harvard's Stickgold believes dreams have a different function entirely. "I think it's pretty clear now that sleep and dreaming serve to process memories from the last day and all the way back," he says. "Sleep can strengthen memories... and help extract the meaning of events by building associative networks with other memories. Dreaming is probably a high-level version of this processing." Clearly, he adds, you don't have to remember your dreams for these processes to work. "The brain is tuning your memory circuits as you sleep, and remembering the imagery created during this process may be fun, may be instructive, but is almost undoubtedly a freebie." Stickgold's evidence includes an experiment he led in 2000 when Harvard researchers were able to elicit the same dream in a bunch of people as they drifted off to sleep. They did this by exposing 27 subjects to an intensive three-day course in the computer game Tetris, which involves assembling geometric shapes. By the second night of training, 17 subjects had reported having the same dream imagefalling Tetris piecesindicating to Stickgold that the need to learn prods the brain to dream. More of these kinds of studies are needed, he says, "because as we learn to manipulate dream content, we can start to figure out what the rules are that the brain uses in selecting material for our dreams." Though not sold on the memory-consolidation theory, the Dream & Nightmare Laboratory's Nielsen sees merit in it. Of course, if dreaming does embed memories it's doing it in ways we don't understand, he says. "Perhaps memory needs to be sliced and diced and then reassembled in odd ways in order for consolidation to be maximized." Psychotherapists tend to regard a lot of the research into dreaming as missing the point. Scientists, they say, can theorize all they like about dreaming's function and physiological underpinnings, but why dreams matter is their effect on the dreamer. The man contemplating an extramarital affair dreams of the dire consequences of having one. He awakens to feel not only exquisite relief that he was dreaming but determined to walk the line. If, as Solms believes, dreams spring from the motivational part of our brain at a time when other parts that inhibit us are off-line, "it follows that there's value in interpreting dreams," he says. They provide a "privileged, unfiltered access" to what's on a person's mind. Mark Blagrove, a lecturer in psychology at the University of Wales, where he runs a sleep laboratory, thinks it's possible the search for a biological function of dreaming could be futile. "It could just be," he says, "that our elaborate dreams are a side effect of the fact that we have a highly evolved imagination." All the competing theories on why we dream may be wrong. One or more of them could be right. "I have no doubt that dreams can be enjoyable, informative, even revelatory to the dreamer," says Harvard's Stickgold. "But dream analysis is a more tricky question. The more dogmatic and doctrinaire the beliefs of the analyst, the less useful and potentially more destructive the analysis process becomes." People should understand, he adds, that dreams aren't constructed with the goal of delivering a message; they don't have an inherent meaning. "But when you look at your dreams after you wake up... you can often feel the associative networks that were activated during dream construction, and trace them back a ways, and maybe discover a new way of looking at events in your life, of looking at yourself, at others, or at the world at large." Maybe that's worth a third of our lives asleep, perchance dreaming.