The next time you lie in bed fuming over your inability to doze off, you might think of Randy Gardner. In the realm of sleeplessness, Gardner once made San Diego the center of the universe. This happened back in January 1964, when he was a student at Point Loma High School. Gardner wanted to enter the San Diego Science Fair, and when he heard about a Honolulu disc jockey who had set a world's record by staying awake 260 hours, the 17-year-old resolved to surpass that and make a science project out of the accomplishment.

From the start, the boy's plan intrigued reporters, and about a week into his attempted marathon, a Stanford University professor and physician named William Dement read about it in his local newspaper. Dement was then the most famous sleep researcher in the world (he still is), and the story electrified him. "I immediately called Randy's home, explained to him and his parents who I was, and asked if I could observe him attempt to break the record," the doctor has recorded in his 1999 book, The Promise of Sleep. Glad to have a medical observer on hand, the parents consented, and Dement flew south.

Over the next few days, Dement spent little time in his motel room. According to the doctor's written account of the adventure, Gardner had found it easy to stay awake until the third night, but from that point on, "He had to be watched every second to prevent him from inadvertently nodding off." Two of his friends had been doing this, sleeping only two hours per night themselves, and Dement joined them as a monitor and cheerleader. "If [Gardner] began to fall asleep, I would hustle him outside to the small basketball court in his backyard or drive him around the deserted San Diego streets in a convertible with the top down and the radio playing loudly."

The hours between 3:00 and 7:00 a.m. proved the most torturous, Dement records. The lad would close his eyes and protest that he was resting them, sometimes growing furious with his companions' insistence that he open them. When no amount of talking or shaking could rouse him, "playing basketball always worked," Dement says. "We almost had to drag him out to the backyard, but once he was there and got moving, he was much better."

As Gardner neared the end of his ordeal, other sources of stimulation helped to keep him awake. His family's phone began ringing constantly, and "reporters and cameramen began to gather at his home," Dement writes. The professor spent the 10th day "walking around town with Randy," and he says he was impressed by how well the teenager did. Gardner beat him time after time playing games on a mechanical baseball machine in a penny arcade, and "he easily bested me in several 3:00 a.m. games of basketball." The 11th day began with Gardner, who'd by then been awake for 263 hours, presiding over a huge press conference. He spoke without slurring or stumbling over his words, he appeared to be in excellent health, and he showed no signs of the psychotic behavior that Dement had suspected might result from prolonged sleep deprivation. After fielding questions for a while, Gardner was driven to Balboa Park, where he was connected to an EEG machine at the Naval Hospital's sleep lab. At 6:04 a.m. he sank into slumber -- having set a world record for sleeplessness that has never been broken.

Although Gardner's feat was one of the most dramatic and closely watched experiments ever conducted with sleep deprivation, it was by no means the first, points out Sean P.A. Drummond, an assistant professor of psychiatry at the UCSD School of Medicine and the VA hospital in San Diego. "The first publication about sleep deprivation was in 1896," Drummond says. "Patrick and Gilbert. They kept people awake for a couple of days and measured half a dozen different things. They basically showed that sleep deprivation leads to bad performance."

More than 100 years later, sleep deprivation remains a fruitful area for scientific inquiry. It's been the focus of Drummond's work for several years. Rather than confining his interest to the behavior of people who haven't slept, Drummond has been studying the activity in their sleep-deprived brains.

He began this research around 1996. He had come to UCSD to work on his Ph.D. in clinical psychology and had the opportunity to learn a new tool for understanding the brain -- functional magnetic resonance imaging (MRI). "Functional MRIs look at a combination of blood flow and oxygen metabolism," Drummond explains. They can be used to pinpoint what part of the brain is working on a particular task.

Drummond says another kind of brain scan -- using positron-emission tomography (PET) -- had already shattered the widespread assumption that human brains become inactive during sleep. The truth is that sleeping brains appear to be at least as active as wide-awake ones, and some parts of the brain are "probably more active," Drummond says. But although this was well-established at the time he started his work, almost no one had compared the activity in sleepy (as opposed to sleeping) brains with that in well-rested ones. Drummond set about to do so.

First he recruited a group of healthy, well-educated subjects, good sleepers all, who ranged in age from their late teens to their mid-30s. He had each volunteer take a math test after he or she had had a good night's sleep and again after going 36 hours without sleeping (the equivalent of pulling an all-nighter, then muddling through the next day without a nap). The results confirmed what researchers had already seen before: the math performance deteriorated.

Drummond's math-testing sessions also included an important new element: each time a subject was given the math exam, both sleepy and awake, a brain scan was obtained. Today when the scientist calls up on his computer the pair of slides that summarize his findings, you don't have to be a brain specialist to see the difference between the two. Whereas the well-rested math-engaged brain bears several splotches of bright color -- signs of intense activity in specific locations -- the same parts of the sleepy math-engaged brain appear almost lifeless. Drummond says, "You can see that after sleep deprivation, each of these areas that responded to the task demands while rested showed significantly less responsiveness." At least when it comes to doing math, the sleepy brain apparently "shuts down."

"So the next thing we looked at," the researcher continues, "was a different kind of test. We gave people a bunch of words, one at a time, and said, 'Memorize these.' We wanted to see what areas of the brain are involved in the memorization process."

The two slides that summarize the results of this testing paint a much more complicated picture. Instead of appearing lifeless, the brain scans taken while the sleepy people were trying to memorize the words revealed more activity than was evident when the memorizers were rested. The activity in one area of the sleepy subjects' brains appeared to expand and -- even more remarkably -- another area (in the right prefrontal cortex) that normally has nothing to do with language appeared to "come online during sleep deprivation," in Drummond's words. This increased brain activity appeared to pay off in terms of the subjects' memorizing performance: after 36 hours without sleep, they were able to memorize the list of words almost as well as when they were rested.

"Essentially, what we took from this is that -- under certain circumstances -- the brain appears to be able to compensate for sleep deprivation," the scientist says. It doesn't always happen, he stresses. If you have to do mathematical calculations while you're sleep-deprived, don't expect perfect accuracy. But for language-memorization activities, Drummond says the scans suggested that the brain has the ability to "recruit additional cognitive resources" that it doesn't normally use and "thereby allow somebody to perform reasonably well."

Since doing the initial set of scans, Drummond says he's been able to replicate the results and ask additional questions. These have included trying to find out if the compensatory responses can be predicted and what are the limits of the brain's ability to compensate. "We're extending sleep deprivation to 62 hours -- two and a half days," Drummond says. "We're scanning in the morning versus the evening, to see if that makes a difference. We're looking at people who naturally get very little sleep, say less than six hours, versus those who naturally get a whole lot -- more than nine hours -- to see if their brains are different. We're doing a lot of different things to see what's going on."

Drummond says the major funding for this work has come from the Department of Defense. "I think there's a lot of recognition in the military that they drive people very hard, and yet there's really no choice in time of war, especially now with very long range operations. You might have pilots who take off from Saint Louis and go bomb Kosovo, then come back. They really need to know what happens when their soldiers and sailors that fly aren't sleeping. And they're also very interested in the recovery process: once you push someone and don't let them sleep for a long time, how long does it take to get back to normal?"

In the case of Randy Gardner, that recovery appeared to be swift. According to Dement's account, the teenager at first slept for 14 hours and 40 minutes; then he awoke spontaneously at a few minutes before 9:00 p.m. By midnight, after giving more interviews, "He was very wide awake and decided to stay up and go to school in the morning." He slept for 10H hours the following night and for 9 hours the third. In three subsequent follow-up visits at the Naval Hospital, his normal sleep requirement appeared to be a bit less than 7 hours nightly.

Although Drummond says much remains unknown about the process of recovering from lost sleep, he points out that the body doesn't seem to need to make up every hour it loses. Rather, some parts of sleep appear more expendable than others. "Sleep comes in different stages," he says, with the most important phases thought to be the deep sleep of Stages 3 and 4 (in which the brain waves, heartbeat, and respiration become the slowest and most regular) and REM sleep (so called for the rapid eye movement that characterizes it). Drummond says when sleep-deprived individuals finally get to bed, they move quickly into the REM and slow-wave phases, spending much less time than normal in the light, early phases of slumber. They also may sleep more efficiently -- falling asleep more quickly once they're in bed and waking up fewer times in the night. "So without even getting extra hours of sleep the next night, you can recover some of what you lost."

Drummond says when the results of his initial MRI work reached publication, "Several people said to me, 'Oh, so what you're showing is it's no big deal not to sleep.' But that's not the case," he argues. "What we're showing is it's not necessarily lethal to not sleep, but the brain is definitely working less efficiently. Even if it doesn't perform completely poorly, it's still going to be slower. It's still going to be less accurate. My guess is that you also are probably less flexible. You'll tend to get stuck in a rut in terms of the way you make decisions or face problems." The consequences of these sorts of impairments can range from trivial to catastrophic: sleep deprivation has been implicated in the grounding of the Exxon Valdez oil tanker, the Chernobyl and Three Mile Island nuclear accidents, and the Challenger space shuttle explosion, among countless less spectacular tragedies.

Drummond concurs "absolutely" with the chorus of voices denouncing how little sleep most Americans get. "We've become a 24-hour society," he says. Work demands, television, the Internet, and other modern blandishments all compete for time spent on the pillow. "People seem to think sleep can be sacrificed for, quote, productivity, even though we're much less productive when we haven't slept," he notes. Various consequences have been shown to flow from partial sleep deprivation. Drummond says researchers have "shown that something like seven nights at six hours of sleep is like 24 hours with no sleep, and seven to nine nights at four hours of sleep is like 48 hours of sleeplessness." After a week of five-hour nights, "You actually get endocrine changes that look like you're developing diabetes. You don't process glucose well. You don't use insulin efficiently. You get increases in ghrelin, which is a hormone that makes you hungry.

"The other interesting factor I like to use to illustrate the effect of sleep deprivation is that some very good studies have shown that if you've been awake for 18H hours -- which for most of us is extending our day by only a couple of hours -- your driving performance is the same as if your blood alcohol content was .05," Drummond says. "And if you've been awake 21 hours, which is not even a full all-nighter, your driving performance is the same as it would be if you had a blood alcohol of .08."

The scientist adds that the subjects of those studies were instructed to sleep eight hours before the day of their testing. Although some individuals require more sleep than that and some less, the majority of people are thought to need between seven and nine hours per night. The fact that substantial differences in sleep needs do exist, combined with the cultural pressures to sleep less, can make it hard for many people to know just how much sleep they should be getting. Renata Shafor, a local specialist in sleep medicine, says people often ask her to tell them how much sleep they need. She suggests, "Look back at your family members, especially the one you resemble the most." Sleep needs have a strong genetic component, she believes. "Or look at how much sleep your body requires when you're on vacation," she counsels.

Drummond points out additional clues to sleep deprivation. He says most people take about 20 minutes to wake up, but after that, "You should get up feeling refreshed. And then you should not feel sleepy during the day. That's the bottom line. If you are falling asleep or wanting to fall asleep in the middle of the day, it means one of two things: either you're not getting enough sleep or your blood sugar is way too low and you're not eating enough. It's one or the other. Typically, it's going to be the lack of sleep. If at three or four in the afternoon, you can sit for an hour in a dark room during a boring talk and stay awake, you're probably sleeping enough. But if your head starts to nod and you want to fall asleep, you're probably not getting enough sleep. You should not feel the need to sleep during the day if you're getting enough sleep at night." (He adds that in cultures with siestas, people typically sleep less at night.)

For someone who's sleep-deprived because he's staying up for Letterman or blogging into the wee hours, the alternative to feeling exhausted may be as simple as scheduling more time in bed and making sleep a priority. But for a substantial number of Americans, opportunity alone isn't sufficient; once they're in bed, they still can't sleep as well as they would like. According to the National Sleep Foundation's 2002 "Sleep in America" poll, 74 percent of those surveyed reported having at least one of the following problems a few nights a week or more: difficulty falling asleep; waking a lot during the night; waking up too early and not being able to get back to sleep; waking up feeling unrefreshed; snoring; having unpleasant tingling feelings in the legs; and having pauses in nocturnal breathing.

Those are the sorts of torments that come to light in Renata Shafor's office, a modest two-story stucco building on Third Avenue, just north of Elm, downtown. For almost 20 years, Shafor has been a specialist in sleep medicine. It's a field that didn't exist when she attended medical school in the early 1960s in Chisinau, Moldova (then part of the Soviet Union). "I wanted to be a neurologist, but there really was no possibility for me at that time," she says. Instead she worked as a family practitioner in Odessa, where "I made a lot of house calls." In 1979, she and her family emigrated, settling in Saint Louis, and Shafor eventually passed the examinations necessary to earn her American medical license. She was delighted to find a residency in neurology at the University of Missouri, Columbia, "and when questions came up about a fellowship, the dean said to me, 'There is a fellowship in an absolutely new field. I'm sure you're going to like it. And it's in a wonderful place: San Diego.' " Although Shafor says she was dubious at first, and her husband kept asking, "Are you sure this isn't a gimmick?" the fellowship proved to be "a priceless experience."

A petite woman with a gentle but confident demeanor, Shafor still speaks with a heavy Russian accent. She says when she arrived in San Diego in 1984, she was the first fellow in sleep disorders in the United States, and she wound up getting exposed to a little bit of everything. Although she spent the majority of her time in Scripps Clinic's then-new sleep-disorders clinic, she also was introduced to the Naval Hospital and VA research programs, and she went up to Stanford for a month and immersed herself in the work there. Today a photo from her Northern California sojourn hangs on the wall in her office. A girlish, glowing Shafor stands between two men, one in his mid-60s and the other his senior. The younger man was Dement, Shafor says, while the older one was Dement's mentor, Nathaniel Kleitman, "the guy who opened up the sleep field."

Another native of Chisinau (born back when Moldova was part of Romania), Kleitman today is recognized as the first scholar in the world to concentrate on sleep. He joined the faculty of the University of Chicago in 1925 and by the late 1930s had devised an experiment that fired the nation's imagination. Kleitman and a colleague spent a month living deep within Kentucky's Mammoth Cave, far removed from the regulating influence of sunlight and human society. Their aim was to study whether they could adjust to a schedule of 28-hour days. Although the results of their stay were inconclusive (one man adjusted, while the other did not), subsequent investigation revealed that human beings do appear to have an innate biological clock that runs on a cycle just under 25 hours long.

Kleitman also established at the University of Chicago the world's first sleep laboratory, and in 1952 his facility was the setting for a pivotal breakthrough. He and a graduate student had discovered that if they pasted electrodes near a research subject's eyes, they could pick up electrical signals when the eyes were moving, and these signals could be used to move pens on a chart recorder. Although the researchers knew that eye movements near the onset of sleep tend to be slow and rolling, they were shocked to discover that at other times, their sleeping subjects' eyes began to dart around, as if watching a scene. In the fall of 1952, when Dement (then a medical student) began working as a volunteer in Kleitman's lab, he was given the job of waking up the sleepers when they were in the rapid-eye-movement states and asking them if they were indeed dreaming. "The results were as dramatic as a bombshell," Dement has written of that effort. "When awakened during bursts of rapid eye movement, subjects had vivid, prodigious dream recall." When no eye movements were present, they seldom remembered anything.

Dement went on to conduct other fascinating experiments related to REM. Suspecting that schizophrenics might be crazy because they lacked the ability to dream, he ventured into a state mental hospital and recorded the sleep of inmates, in short order proving his theory wrong when it became clear that the schizophrenics had both REM sleep periods and dreams. In the University of Chicago sleep lab, he pushed to collect a continuous record of sleepers' brain waves and eye movements throughout the night, rather than the sporadic recordings his colleagues (not wanting to waste paper) had been taking. From the night-long evidence, Dement began to decipher how normal sleep is organized, and he published the first scientific paper describing that complex architecture. Today we know that sleepers' brain-wave patterns usually move through four stages before the first REM sleep appears, about 90 minutes into the evening, with REM occurring more and more frequently in the second half of the night. This pattern changes when someone is REM-deprived. (Dement discovered this when he systematically interrupted sleepers' dreamtime.) At such times, the brain "tries to compensate by having longer REM periods and having them sooner."

In 1963, Dement moved to Stanford, where he continued his research and a year later opened a clinic to treat people with narcolepsy, a sleep disorder characterized by overwhelming daytime sleepiness and brief spells of weakness and paralysis. The clinic lost money, however, and soon closed. It wasn't until 1970 that Dement tried again, opening the world's first general sleep-disorders clinic. In addition to treating narcolepsy sufferers, it welcomed "any and all patients with insomnia."

Since then the field of sleep medicine has undergone tremendous expansion, according to Shafor (who opened her own facility in San Diego in 1985). "It grew from just insomnia and excessive daytime sleepiness into so many different diagnoses and conditions," she says -- more than 100. To figure out which problem someone has, Shafor likes to begin by meeting with both the patient and his or her "bed partner." That's because people often aren't accurate in describing their own sleep-related behaviors. Serious snorers often "don't hear themselves," she says, or they'll say they don't feel sleepy during the day; it can take the partner to point out how the other person dozes off at lunchtime. "There's a lot of denial," the physician says. She also examines each patient, searching for physical conditions that might be sabotaging healthy sleep.

Shafor says in many cases this initial evaluation yields enough information to point the way to a treatment. Some people have poor sleep habits and need instruction in the basic rules of sleep hygiene. Shafor says over the years she's expected these to change, as understanding of sleep has evolved. "But they really haven't," she says. "I'm amazed." One of the foremost guidelines is to keep regular hours. "You go to bed at the same time and get up at the same time," the doctor instructs. "Don't have alcohol less than three hours prior to going to bed. It makes people sleepy and it's easy to fall asleep." But a few hours later, alcohol awakens many people. Heavy exercise and cold showers often have a stimulating effect, so they should be avoided in the late evening. "Don't watch TV in your bedroom," Shafor continues. "That should be for sleep and sex."

"If you do have problems sleeping, then it becomes important to avoid going to bed until you're sleepy," she advises. Some individuals can feel sleepy, but "as soon as they hit the pillow, they're wide awake again. So if 10 to 15 minutes go by and you're not falling asleep, get up again. Do something -- as long as it's not exciting." Shafor says she often recommends "sleep restriction." For someone who goes to bed at 10:00 and gets up at 6:00 but only sleeps five hours, she'll say, "Don't go to bed at 10:00. Go to bed at 1:00 a.m. And get up at 6:00 a.m. But make sure the alarm is working." She tells such individuals to "do that until you learn to use those five hours and feel that you can sleep more." Then they can start to move their bedtime to an earlier hour.

She says some of the patients who come to her are reaping the consequences of a period in which they have ignored their body's normal sleep requirements. "I had an attorney who spent several nights and days not sleeping at all," Shafor says. The woman explained that she'd been launching an exciting new project, but when that work wound down, she was dismayed to discover she couldn't fall asleep. "She would say, 'I go to my bedroom, but just by thinking of bed, I'm wide awake.' She had conditioned herself to look at the bed and face the struggle for sleep." Each day she would try going to bed a little earlier, only to wind up logging more hours of frustration. Shafor says she found no evidence of any other physical ailments. "She had just sort of damaged her body clock -- damaged her homeostasis." So the doctor says she worked with the woman "to reinforce her body clock and create a new schedule."

She says other patients come in with what's called "delayed-phase sleep syndrome." They sleep just fine. The only problem is, they do it at the wrong time. They may be unable to doze off until 4:00 a.m., for example, but then they'll enjoy a blissful repose until noon. While that schedule may work well on a college campus, it can cause agony for someone with a more conventional work schedule. So the sleep doctor works with such patients to readjust their biological clock. "The problem is, you cannot move the clock backwards," she says. "You have to move around the clock."

Shafor estimates that perhaps 40 percent of the time her initial patient evaluation fails to reveal the cause of the patient's problem. "If I cannot tell from talking [to the patient], then I do a sleep study." At Shafor's clinic, these sessions take place on Sunday and Monday nights. Last spring, Alex Rochells was the technician who was conducting most of them. A Navy veteran, Rochells started out as Shafor's receptionist but became so intrigued by the sleep studies that he began learning how to prepare patients for their night of medical scrutiny. On a Monday night last March, he figured he had administered more than 1200 of the assessments.

That evening, Rochells had just learned that one of the two patients scheduled for that evening would not be coming in. "The main problem with doing sleep studies is the number of people who don't show up," he commented. A bearded man with receding brown hair, Rochells was dressed in khaki scrubs and wearing wire-rimmed glasses. The studies require a big commitment of time, he acknowledged. Most patients arrive at 8:30 or 9:00 p.m., and they not only spend the night but also must remain at the clinic most of the following day. In the daytime portion of the study, they're required to lie down in one of the darkened sleeping quarters four or five times for 20 minutes apiece. Noting whether they fall asleep and how long it takes for them to do so provides a way to quantify their daytime sleepiness. (Dement and a fellow sleep researcher developed this so-called Multiple Sleep Latency Test back in 1975, and today it's a fixture of sleep medicine.)

A few minutes before 9:00 p.m., the front door of the clinic opened, and Raquel Cervantes, a short, plump 70-year-old with a head of thick, steel-gray hair, walked in, clutching a pillow. Rochells directed her to don her sleeping clothes (a well-worn pair of blue flannel pajamas), then escorted her to a crowded room at the back of the office suite. Here he began the tedious process of attaching all the wires that would expose what would be happening to Cervantes's brain and body during the coming hours.

First the technician rubbed an exfoliating cleanser and some extra-fine sandpaper on a spot on each of the woman's lower legs. He connected electrodes at each place; these would reveal any abnormal movements of her legs during the night, a common cause of insomnia. People with "restless legs syndrome" feel crawling or uncomfortable sensations in their limbs. While movement relieves the sensations, it also interferes with sleep. Those with the similar and often related "periodic limb movements disorder" experience involuntary jerking motions in their toes and feet that may be strong enough to awaken them. Medication can alleviate both conditions.

Next Rochells parted Cervantes's hair at pre-established locations, rubbing the scalp with Q-Tips, then pressing on electrodes filled with conductive gel. He positioned a total of nine electrodes in this manner, covering each with a glue-soaked piece of gauze to anchor it. "See, it is a lot like hairdressing," he joked. He attached still more wires near the woman's eyes (to reveal periods of rapid motion) and on her chin (to pick up signs of teeth-grinding). Last to be connected were an airflow sensor that strapped on below her nostrils and two belts positioned around her chest and abdomen that were designed to measure her respiratory effort. Along with a sensor that Cervantes would wear on one fingertip, these devices would reveal whether she suffered from the disorder known as sleep apnea.

That's what her doctors suspected, Cervantes disclosed. She herself had long known she snored; she was so loud family members refused to sleep in the same room with her. But she'd never thought of her snoring as anything requiring medical attention. Then, after hitting her head during a fall, she had talked to a neurologist, and he had sent her to Shafor.

Not everyone who snores stops breathing for intervals long enough to lower blood oxygen to dangerous levels. But snoring is a symptom of impaired breathing, and it may lead to or include periods of apnea (a Greek word meaning "without breath"). First recognized in the mid-1950s, apnea occurs (as does snoring) because the soft tissues at the back of the throat tend to collapse during times of relaxation. Obese individuals in particular are susceptible to this malady. The respiratory pauses may occur dozens or even hundreds of times per night, and in the worst cases, they may stretch out for a minute or longer, sending blood-oxygen levels plummeting. The sleeper totters on the brink of suffocation before heaving a convulsive gasp for air, which in some instances never comes. As a result, "Between 2000 and 3000 people a year in America die in their sleep," according to Shafor.

Remarkably, apnea victims often are unaware of their mortal struggles during the night, even though the malady has been estimated to rob sufferers of up to a third of their sleep. Chronic exhaustion results, along with its attendant dangers. (Some studies, for example, have estimated that apnea sufferers have ten times the number of car accidents as the general population.) Moreover, the nightly cycle of oxygen deprivation and pulse-pounding recovery takes a heavy toll on the cardiovascular system, contributing to high blood pressure and coronary artery disease. The National Commission on Sleep Disorders Research in 1992 estimated that apnea caused 38,000 fatal heart attacks and strokes every year in the United States.

"I have high blood pressure," Cervantes acknowledged as Rochells bundled together all the wires attached to her and stuffed them into a plastic sleeve. It was a few minutes after 10:00 p.m. -- earlier than her normal bedtime, the woman said. "Usually I sleep until about 9:00 in the morning, and I go to bed at 12:30 or 1:00. But tonight I'm tired." She expected to doze off quickly but did express one concern: she was accustomed to going to the bathroom several times a night. How could she do that if all these wires were plugged into monitoring devices? Rochells reassured her that all she had to do was to call out. A microphone in her sleeping chamber would alert him to come and disconnect her.

The three rooms where Shafor's patients sleep are spartan, but they're dark and quiet, in contrast to the light and activity in the room where Rochells spends most of his working hours, facing six monitors of various sizes and brands. Now one of them displayed a grainy image of Cervantes tossing and turning in the unfamiliar bed, while on two other monitors, an array of bristly tracings moved across the screens, graphic representations of the information being captured by the electrodes. "These are alpha waves," the technician said, pointing to one section of somewhat more symmetrical patterns. "When you close your eyes, you will always have alpha waves. Alpha waves are when you're in thought, basically."

When Rochells was learning to work as a sleep technician, Shafor still used traditional equipment that recorded patients' tracings with ink on paper. Rochells preferred these, he said, because they made an attention-catching noise whenever a connection malfunctioned. But the old-style machines also gobbled up close to 1000 sheets of paper in the course of one night of recording an average patient, so around the beginning of 2003, the clinic switched to using computers to record and display the readings. Although Shafor says programs have become available to perform some of the analysis, she doesn't trust them; the physician "scores" every patient's sleep record herself.

"Some people don't sleep all night long," Rochells said. Others start snoring the instant their heads hit the pillow. The technician confided that on the single occasion when he spent a night in the clinic wired to the monitors, "I discovered that my own conception of time stinks! I had thought it took me around 45 minutes to get to sleep, but later I discovered that it was like 7H minutes. My impression was way off!"

It took Cervantes around 15 minutes to start to relax, Rochells observed, pointing out the rolling landscape forms on the monitor that revealed the wandering motion of her eyes. This precedes the descent into slumber. Her brain-wave patterns also were becoming lower, and a while later, the technician spotted the bursts of brain waves called "spindles," as well as "K complexes" (isolated brain waves that resemble asymmetrical little mountain peaks). "Now she's in Stage 2 sleep," Rochells pronounced.

Later, when Cervantes moved into the deeper sleep levels, Rochells began seeing evidence of sleep apnea: pauses in the sleeping woman's breath; plunging blood-oxygen levels. So about halfway through the night, he woke her up and had her strap on a bulky, insectile-looking mask that covered her nose. This device delivers a gentle stream of air, creating what's known as "continuous positive airway pressure" (or CPAP). Such devices keep the breathing passages open; they've been shown to prevent obstructive apnea in almost 100 percent of those who can adapt to using them.

A few days after the sleep test, Cervantes confessed to panicking when Rochells asked her to put on the mask. She's had a longstanding horror of anything touching her nose; she compares the sensation to claustrophobia. "I feel that I cannot inhale. So when I saw the mask, I thought, 'I'm not gonna wear that.' " Despite her reluctance, she finally complied, "And after a while, when the oxygen started going through, you know -- then it felt good!" She fell asleep again. Now she was willing to try making the mask a part of her daily life, even if it made her look "like somebody from another world." She could understand the need to deliver enough oxygen to her body, she said.

According to Shafor, it isn't ideal to have to wake apnea sufferers halfway through the night to have them try on the CPAP devices. "By missing the second half of the night, you may miss a lot of information. The right thing to do medically and ethically is to have one night's sleep study diagnostically to see what's going on, and then to bring [the patient] back the second night and start titrating the treatment: CPAP, oxygen, whatever, and see how it goes." But most insurance providers, especially Medi-Cal, won't pay for two nights of the expensive observation. (The night portion of the sleep studies ranges from $800 to $1200, and the day portion can be an additional $600 to $700.)

Still Shafor sounds grateful that insurance companies have come as far as they have. Just five or ten years ago, "They wouldn't reimburse sometimes at all for sleep and diagnostic studies." She says today virtually all insurance companies cover the majority of sleep studies and referrals for sleep disorders. She says the studies often are necessary to confirm a suspected problem, before she can prescribe medication or some other form of treatment. That's the case for what's known as REM behavior disorder, "one of the most fascinating problems," in Shafor's estimation. "[Couples] come to me and say that the husband is almost killing his wife, beating her at night and screaming." Sleep studies of such individuals have documented that their violent outbursts occur during REM sleep. "REM sleep is notorious for the mind being actually in an awake state," Shafor explains. "We see things, we make decisions, etc." Although a powerful physiological response paralyzes the arms, legs, and trunk of most normal sleepers before each period of REM, people with REM behavior disorder regain their muscle tone, "and they begin acting out their dreams. I have several guys like that," Shafor says. "They see in the dream, for example, that a snake is crawling up their wife. So they try to kill the snake. And in the meantime, they hit the wife!" Medications can correct the problem, she says. "Couples are able to move back together in the same bedroom."

Shafor says the sleep studies also can demonstrate that some patients have no sleep problem. "There's such a thing as Sleep/Wake Misperception. These people say, 'I'm in bed. I keep my eyes closed. But I don't sleep at all.' Then they come to the lab and say the same thing has happened here. But I can look at the record and see that they were in deep sleep. Not only light sleep, but deep sleep. When these people have their sleep interrupted, all they know is it was just a second ago they were awake, and now they're awake again. They misjudge the period of time they were asleep."

The physician says such individuals often feel relieved to learn about their misperception. Many haven't been feeling bad during the day, but rather they've worried about the ill effects of not sleeping. "They wonder, 'Am I going to get sick soon? Or go crazy?' " Shafor says she's always careful about how she breaks the news to them. "If you say, 'What are you talking about? You were asleep! Look at this!' you can create antagonism. Instead you have to tell them, 'You have to understand that sometimes it feels like we don't sleep, but the brain does catch some period of sleep.' "

Despite the increased willingness of insurance companies to reimburse the costs of sleep medicine, Shafor believes that many insomnia sufferers still lack access to effective treatment. Some shrink from talking to their doctors about their problems, and doctors who do hear such complaints tend to "just feed [their patients] sleeping pills. A lot of pharmaceutical companies leave a lot of samples, and the doctors just give them out. 'Take it. Come back in three months.' And the patient takes it and comes back and still doesn't sleep well."

Shafor thinks sleeping pills have their place. For the person losing sleep because of a stressful situation -- a family crisis or an upcoming surgery -- "then it can be a good idea to take a sleeping pill, as long as there's no contraindication," the doctor says. She sometimes uses sleeping pills to help patients whose biological clock is out of sync with the demands of society. But for chronic sleep troubles, "Sleeping pills most of the time become ineffective at best," she declares.

"At worst, they require more and more pills to the point where patients sometimes take bunches of them," Shafor says. This was true of the barbiturates used as sleeping medication from the early years of the 20th Century well into the 1970s. These drugs (which had trade names such as Seconal, Amytal, and Nembutal) posed a high risk of accidental overdose and, when combined with alcohol, worked as an effective suicide tool for everyone from starlets to suburban housewives. A new class of hypnotic -- benzodiazepines with trade names such as Librium and Valium -- began reaching troubled sleepers in the '70s, and by the end of that decade, one version, called Dalmane, had become the most popular sleeping pill in America. But Dalmane takes a very long time to break down in the body, according to Shafor. "So people would take, say, a half milligram, and the next night another half milligram, but they would still have some in their system," so the overall amount accumulated.

That problem doesn't seem to occur with Ambien and Sonata, two drugs that belong to a third, still newer, class of sleeping pills. Patients also "don't require more and more of them," Shafor says. "There's not much grogginess the next day, although some degree's still there." Unlike the benzodiazepines, the newer drugs don't appear to change the brain waves and patterns seen during sleep. And they don't seem to be addictive "in the sense where it becomes part of your metabolic system, and you feel that you cannot survive without it."

Even with the newer drugs, patients develop a relationship that Shafor says might be thought of as a psychological dependency. "Patients create the relationship," she says. "They choose something they like and become firmly attached to it." It might be Motrin. "They'll say, 'Naprosyn is no good. I like Motrin. I have to have my Motrin in the house.' " Even things that are not physically addictive can create this sense of need. "Maybe it just provides a sense of security: if I have this cane here, I can go around the house. Or in the past, people might have had a glass of warm milk. But they don't want to give it up, even if they know they maybe don't need it."

Studies of the effects of taking Ambien and Sonata night after night for months or years have not been done, Shafor points out. So when she's asked what's wrong with long-term usage, she replies, "We don't know. We're just afraid that maybe something's wrong, because we don't know."

She says yet another type of sleeping pill will reach the marketplace soon, one based on the hormone melatonin. Released by the pineal gland, melatonin works on the tiny cluster of brain cells that function as the biological clock (the suprachiasmatic nuclei). Health-food stores have been selling forms of melatonin for years, but Shafor says the sleep-medicine community has been skeptical. "First of all, it's not crystallized. It's not dosed. You really haven't known what you were buying.... At one point somebody checked something labeled melatonin that was brought in from Mexico, and half of it was benzodiazepine!"

At the same time, solid research into melatonin's sleep-inducing properties has been progressing, and a product appears to be close to receiving FDA approval, Shafor says. She looks thoughtful about the potential for this drug or one of its successors. "Let's say hypothetically, in the future -- remote or not -- we finally know how to manipulate the biological clock. So we can make soldiers go three or four days without bothering to sleep. Is this good or bad? Are we changing their personality? What else are we changing there? Who are we after we don't do whatever human beings have always done?" She offers no answers; rather, she's just posing the questions.

The sleep doctor says sooner or later she expects an effective long-term sleeping pill to be developed. "But I think we're getting smarter and smarter about identifying the group of people who would need that." Some individuals will probably always need something more than sleeping pills, "starting with psychotherapy and changing the whole regimen of medical treatment. Take Synthroid [the synthetic hormone taken by people who don't produce enough thyroxine]. It definitely affects sleep. It keeps people on edge." So do asthma medications, "While other medications have just the opposite effect." No one sleeping pill is likely to ever work for everyone, Shafor thinks.

"We still think insomnia is the enemy," she reflects. "But it's a dragon with so many heads."