IN THE SUMMER OF 2010, Ben Blackwell, a five-year-old living near Dublin, Ireland, began complaining of headaches and a squealing noise in his head. And even though he’d stopped taking naps three years earlier, he now randomly fell asleep—while watching television, reading a book, sitting in the car. When Ben’s parents, James and Natalie, succeeded in rousing him, he often snarled at them and seemed terrified. But a few hours later, he’d be fine.

Natalie read a newspaper article about dramatic changes in the behavior of some children who had been vaccinated against the flu. Because of the disproven link between vaccines and autism, James associated any talk of vaccine-related problems with conspiracy theories and pseudoscience. But Natalie dug up more news items, and it turned out that a number of children across Northern Europe—from Ireland to Finland—did indeed appear to have developed a sleeping disorder after inoculation against the previous winter’s swine flu. Ben had received the vaccine at school. Now a test of his spinal fluid showed a complete lack of the hormone hypocretin, which regulates wakefulness. An absence of hypocretin can indicate the sleep disorder narcolepsy, and that was Ben’s diagnosis. “In a sense, it was a huge relief,” says James Blackwell. “Finally we could start working on making it better.”

But the larger question of what caused the narcolepsy of Ben and other children remains unresolved. Four years and several country-wide studies later, there now is an established link between Pandemrix, a vaccine against the H1N1 strain of the flu that was widely used in Europe, and some cases of narcolepsy. (The GlaxoSmithKline vaccine wasn’t used in the United States, and the manufacturer’s statement about the European cases stresses that “a detailed understanding of the mechanism by which Pandemrix may have contributed to an increased risk of narcolepsy remains elusive.”) Absolute numbers of vaccine-associated narcolepsy cases have been small—upwards of 900 so far among the 30 million people vaccinated with Pandemrix. And it may be that the vaccine only accelerated progression of narcolepsy in children who would have developed it anyway.

Still, the quest to explain the Pandemrix link to narcolepsy may lead to a breakthrough in understanding the disorder. Scientists have long speculated that the immune system might be involved in narcolepsy. They theorized that in people who were genetically susceptible, an accidental assault by cells normally tasked with fighting infection  destroyed neurons crucial for regulating sleep. The problems with Pandemrix could finally clarify the kind of trigger that prompts this self-attack. “It’s the first time we can really prove that a vaccine or an immune response can trigger an autoimmune disease,” says Emmanuel Mignot, director of Stanford University’s Center for Sleep Sciences and Medicine.

On that front, the narcolepsy mystery intersects with a much larger story. A growing number of apparent neurological and psychiatric disorders appear to be driven by autoimmune processes. And a subset of those may be treatable not with antidepressants but rather with therapy directed at the immune system.

ONE OF THE EARLIEST BREAKTHROUGHS involving autoimmunity and the brain came in 2007. Josep Dalmau, a neurologist then at the University of Pennsylvania in Philadelphia, was studying several women who suddenly began experiencing a syndrome that included psychiatric symptoms, memory problems, seizures and abnormal movements. They all had a type of ovarian tumor called a teratoma. Dalmau discovered that teratomas can feature synaptic receptors and proteins that also occur in the brain. The women’s immune systems­—in mounting a response against the synaptic receptors present in the ovarian growths—inadvertently attacked a receptor in the brain called N-methyl-D-aspartate receptor, or NMDAR. What had looked like a serious neuropsychiatric disorder was really a kind of autoimmune disease. After having their tumors removed and receiving immunotherapy, most of the women improved or recovered.

Dalmau dubbed the condition “anti-NMDA receptor encephalitis.” He then reported the cases of others he had discovered, including men and young children, who had no tumors but who carried those NMDAR-directed antibodies. Many of their neuropsychiatric symptoms also improved with immunotherapy, usually a combination of steroidal immuno-suppressant drugs and intravenous infusions of immunoglobulin-G, though many patients required more aggressive immunotherapy. This was the same condition that afflicted journalist Susannah Cahalan, who in 2012 described her experiences in a best-selling memoir, Brain on Fire: My Month of Madness.

In one sign of how far such investigations have come, a center of research on autoimmune neurology at the Mayo Clinic in Rochester, Minn., ran some 140,000 tests last year for antibodies that could indicate autoimmune diseases of the central nervous system.

Tracey Cho, a neurologist at Massachusetts General Hospital, calls these developments “a game changer.” He points to a recent retrospective analysis by the California Encephalitis Project that analyzed the blood of 761 patients that had been gathered and stored during the previous decade. The results suggested that the condition wasn’t as rare as assumed. For patients under 30, autoimmune encephalitis was actually more common than encephalitis caused by viral infection.

Indeed, Cho has seen patients previously diagnosed with various psychiatric symptoms who turned out to have anti-NMDAR encephalitis and who recovered with immunotherapy. “In my field, this is definitely the biggest advance in recent times,” he says. Neurologists now “cure” patients who, just a few years back, would have been considered intractable cases.

Sean Pittock, a neurologist at the Mayo Clinic, has his own miraculous recovery stories—patients diagnosed with autoimmune dementia and epilepsy who got better just days after beginning immunotherapy. He thinks that thousands of people around the country, some of them in nursing homes, have potentially treatable autoimmune diseases of the central nervous system. They’ve just been misdiagnosed.

W_15_betrayal_within_spot_2__630x420

AGAINST THE BACKDROP OF THOSE rapid advances, the scientists studying narcolepsy hope for their own breakthrough. They are, in a sense, still searching for their version of Dalmau’s anti-NMDA receptor encephalitis—definitive proof that narcolepsy is an autoimmune disease of the brain.

An estimated one in 2,000 Americans suffers from narcolepsy, which is sometimes accompanied by seizures, hallucinations, muscle paralysis and anxiety. In some narcoleptics, strong emotions can trigger a sudden loss of muscle tone, called cataplexy, that leaves them lying on the floor, body limp and paralyzed but with their minds perfectly awake.

In the early 1980s, scientists at the University of Tokyo found that all of their narcolepsy patients had the same form of the human leukocyte antigen, or HLA. This is a U-shaped molecule on the outer membrane of white blood cells that functions like a grasping claw. It grabs proteins from invading pathogens, displays these signature proteins to immune cells, and “trains” the cells to recognize and pursue hostile bacteria or viruses that have those proteins. Developing such a memory of the pathogens’ molecular “fingerprint” is supposed to help the body form long-lasting immunity to particular bugs.

The apparent link between narcolepsy and a specific form of HLA—HLA-DR2—suggests that the immune system might be involved in the disease. In this hypothesis, the variant plays a key role. When people who carry it encounter certain invading pathogens that produce humanlike proteins, their immune cells might be unusually prone to attacking their own body’s cells as they pursue the interloper. Subsequent studies found other connections to the immune system. Virtually all narcolepsy patients with cataplexy carried the gene variant HLA-QB*06:02. Other gene variants associated with the function of T-cells, an immune system workhorse, would later come to light.

But for years, no one knew what protein the immune system might pursue that would produce uncontrollable sleepiness. One clue came in the 1990s, from observations of Doberman pinschers that would collapse on the floor, paralyzed but awake, after excited play, in a kind of canine cataplexy. Mignot and his colleagues identified a spontaneous genetic mutation in those animals that incapacitated a receptor in brain cells for hypocretin. Other researchers, meanwhile, found that hypocretin was crucial for staying awake.

People normally carry relatively few of the neurons that produce that hormone—about 70,000 of them, located in the hypothalamus region deep in the brain. But when researchers analyzed the brain tissue of narcoleptics taken during autopsies, they found practically none of those neurons. Spinal fluid from living narcoleptics also has revealed an almost complete absence of hypocretin.

Studies of genetically identical twins have shown that when one twin develops narcolepsy, the other tends not to. That strongly suggests that there must also be an infection or another kind of environmental trigger that starts the disease process.

In their search for such a trigger in people who had become narcoleptic, Mignot’s group found elevated levels of antibodies to Streptococcus pyogenes, a bacterium that causes strep throat and has been implicated in other autoimmune diseases, such as rheumatic fever. And a University of Washington study found that among people who have the HLA form thought to predispose them to developing narcolepsy, having had strep throat early in life increased the risk of actually getting the disease more than fivefold.

Scientists hypothesized that the problem with S. pyogenes might be “molecular mimicry,” in which a molecular pattern on the bacterium resembles proteins on those hypocretin neurons. The immune system, going after the strep bug, could inadvertently pursue and destroy the neurons. Yet it still wasn’t known what protein might be the culprit. And without it, the autoimmune hypothesis for narcolepsy languished—until the swine flu pandemic of 2009.

IN EARLY 2010, LARS PALM, A PEDIATRIC neurologist at Skåne University Hospital in Sweden, noticed an alarming trend. He usually follows one to three cases of narcolepsy continuously. But that spring, he saw six new cases.

Fears of an H1N1 influenza pandemic had spurred a mass vaccination campaign in Sweden and other European countries, and Sweden had achieved a public health triumph, getting 60% of its population vaccinated within a few months. But now, Palm heard the same story from the parents of several narcoleptic children. Their symptoms—including disturbed sleep cycles and, in some cases, cataplexy—had come on after vaccination for H1N1.

Palm alerted Swedish health authorities, and subsequent population-wide studies in Sweden, Finland, Ireland and the United Kingdom suggested that vaccination for H1N1—but only with the Pandemrix vaccine—had increased the risk of developing narcolepsy between four- and 13-fold in children and adolescents, and between two- and four-fold in adults. Nearly all of the affected children had the HLA-QB*06:02 gene variant already linked with narcolepsy. (Ben Blackwell, in Ireland, also has it.)

The most obvious difference between Pandemrix and the vaccine against H1N1 that was used in the United States—which hasn’t been associated with narcolepsy—was that Pandemrix included a strong adjuvant, a substance intended to provoke a heightened immune response to proteins in the vaccine. Pandemrix contained a particularly powerful adjuvant called AS03. (Adjuvants are not used in flu vaccine preparations in the United States, and Pandemrix was not licensed for use here.)

That discrepancy between vaccines used in Europe and those used in the United States seemed to impugn the Pandemrix adjuvant. But in China, where comparatively few people were vaccinated, authorities also registered a nearly sevenfold spike in new narcolepsy cases during the swine flu year. And Mignot, in his Stanford clinic, also noted an uptick. He and others believe that these cases were associated with infection by the live H1N1 influenza virus itself, rather than with the Pandemrix vaccine.

As anyone who has suffered through the aches, fever and delirium of the flu can attest, the influenza virus can prompt a major inflammatory response, with the live virus essentially acting as its own adjuvant. Other infections can also cause major inflammation, though few of those are associated with narcolepsy. These associations suggest that in people with a genetic predisposition to get narcolepsy, a molecular pattern on this virus can trigger the disease, but only when there is significant inflammation.

MIGNOT THOUGHT HE HAD SOLVED THIS puzzle once before. In late 2013, he published what seemed to be a breakthrough study of patients with narcolepsy, which identified self-reactive T-cells that pursued the hormone hypocretin. In line with the “molecular mimicry” idea, he also thought he had identified viral proteins that resembled hypocretin. But in July, after failing to replicate his earlier results, he retracted the study.

Other research teams have proposed different proteins that may prompt an autoimmune attack. Three groups independently studying people with narcolepsy identified antibodies against the same protein, tribble 2, that’s found on hypocretin neurons. In 2014, another group looking at narcolepsy pinpointed autoantibodies to various other proteins in the brain.

But some researchers doubt that self-directed antibodies cause the disease at all. To begin with, not all narcoleptics have the antibodies that have been implicated. For instance, in the case of tribble 2, the antibodies are present in 14% to 26% of cases. And some people who are perfectly healthy turn out to have the same antibodies, suggesting that they are neither necessary nor sufficient to explain the disease.

“I’m skeptical,” says Angela Vincent, a scientist at the University of Oxford and a pioneer in the field of autoimmune neurology. Her own analysis of blood from narcoleptic patients vaccinated with Pandemrix revealed no clear evidence of self-directed antibodies. She believes T-cells, which don’t produce antibodies, destroy the hypocretin-producing neurons directly. Yet in her view, there is now wide consensus that the immune system does play a role.

And finding an immunological marker for narcolepsy is important for reasons beyond just proving that it’s an autoimmune disease. Such a discovery could, in theory, allow for timely intervention. In anti-NMDAR encephalitis and other autoimmune disorders of the brain, the earlier immunotherapy begins, the more complete the recovery.

Yet even if narcolepsy is proven to be autoimmune, its treatment faces unique hurdles. Symptoms appear only after more than 90% of the hypocretin-producing neurons have been destroyed. That may be why immunotherapy for narcolepsy has so far yielded lackluster results. The intervention comes too late.

For now, most narcoleptics have to manage the disease with old-fashioned methods. Ben Blackwell, now 10, naps very frequently and takes the stimulant Ritalin to remain alert. Because he gets an energy boost when he eats, his parents strategically space his meals throughout the day. In the absence of a cure—which would presumably involve regenerating hypocretin-producing neurons—the Blackwells hope better drugs become available soon.

A better understanding of autoimmunity’s role in narcolepsy might open the door to clinical solutions, large and small. Ben Blackwell and others who find their bodies’ own defenses turned against them anxiously await a breakthrough. But for now, the mystery of narcolepsy still has “to be continued” stamped on the file.