Why would an industry beat a hasty retreat from a market that continues to  boom? (Recent surveys indicate that mental illness is the leading  cause of impairment and disability worldwide.) The answer lies in the history of  psychopharmacology, which is more deeply indebted to serendipity than most  branches of medicine—in particular, to a remarkable series of accidental  discoveries made in the fifteen or so years following the end of the Second  World War.

In 1949, John Cade published an article in the Medical Journal of Australia describing his discovery that lithium sedated people who experienced mania. Cade  had been testing his theory that manic people were suffering from an excess of  uric acid by injecting patients’ urine into guinea pigs, who subsequently died.  When Cade diluted the uric acid by adding lithium, the guinea pigs fared better;  when he injected them with lithium alone, they became sedated. He noticed the  same effect when he tested lithium on himself, and then on his patients. Nearly  twenty years after he first recommended lithium to treat manic depression, it  became the standard treatment for the disorder.

In the nineteen-forties and fifties, schizophrenic patients in some asylums  were treated with cold-induced “hibernation”—a state from which they often  emerged lucid and calm. In one French hospital, the protocol also called for  chlorpromazine, a new drug thought to increase the hibernation effect. One day,  some nurses ran out of ice and administered the drug on its own. When it calmed  the patients, chlorpromazine, later named Thorazine, was recognized in 1952 as the first drug treatment for  schizophrenia—a development that encouraged doctors to believe that they could  use drugs to manage patients outside the asylum, and thus shutter their  institutions.

In 1956, the Swiss firm Geigy wanted in on the antipsychotics market, and it  asked a researcher and asylum doctor, Roland Kuhn, to test out a drug that, like  Thorazine, was an antihistamine—and thus was expected to have a sedating effect.  The results were not what Kuhn desired: when the schizophrenic patients took the  drug, imipramine, they became more agitated, and one of them, according to a  member of the research team, “rode, in his nightshirt, to a nearby village,  singing lustily.” He added, “This was not really a very good PR exercise for the  hospital.” But it was the inspiration for Kuhn and his team to reason that “if  the flat mood of schizophrenia could be lifted by the drug, then could not a  depressed mood be elevated also?” Under the brand name Tofranil, imipramine went  on to become the first antidepressant—and one of the first blockbuster  psychiatric drugs.

American researchers were also interested in antihistamines. In 1957, Leo  Sternbach, a chemist for Hoffmann-La Roche who had spent his career researching  them, was about to throw away the last of a series of compounds he had been  testing that had proven to be pharmacologically inert. But in the interest of  completeness, he was convinced to test the last sample. “We thought the expected  negative pharmacological results would cap our work on this series of  compounds,” one of his colleagues later recounted. But the drug turned out to  have muscle-relaxing and sedative properties. Instead of becoming the last in a  list of failures, it became the first in a series of spectacular successes—the  benzodiazepenes, of which Sternbach’s Librium and Valium were the flagships.

By 1960, the major classes of psychiatric drugs—among them, mood stabilizers,  antipsychotics, antidepressants, and anti-anxiety drugs, known as  anxiolytics—had been discovered and were on their way to becoming a  seventy-billion-dollar market. Having been discovered by accident, however, they  lacked one important element: a theory that accounted for why they worked (or,  in many cases, did not).

That didn’t stop drug makers and doctors from claiming that they knew.  Drawing on another mostly serendipitous discovery of the fifties—that the brain  did not conduct its business by sending sparks from neuron to neuron, as  scientists previously thought, but rather by sending chemical messengers across  synapses—they fashioned an explanation: mental illness was the result of  imbalances among these neurotransmitters, which the drugs treated in the same  way that insulin treats diabetes.

The appeal of this account is obvious: it combines ancient notions of illness  (specifically, the idea that sickness resulted from imbalanced humors) with the  modern understanding of the molecular culprits that make us suffer—germs. It  held out the hope that mental illness could be treated in the same way as  pneumonia or hypertension: with a single pill. Drug companies wasted no time in  promulgating it. Merck, the manufacturer of Elavil, commissioned the  psychiatrist Frank Ayd to write a book called Recognizing the Depressed  Patient, in which he extolled the “chemical revolution in psychiatry” and  urged doctors to reassure patients they weren’t losing their minds, but rather  suffering a “common illness” with a “physical basis” and a pharmacological cure.  Merck sent Ayd’s book to fifty thousand doctors around the country. In 1965,  Joseph Schildkraut, a psychiatrist at the National Institute of Mental Health,  reverse-engineered antidepressants and offered an actual theory: at least when  it came to depression, the imbalances were to be found in the neurotransmitters  he thought were affected by the drugs, dopamine and norepinephrine. Seven years  after antidepressants were invented, and five years after Ayd asserted that  depression was a chemical problem, psychiatrists finally had a precise,  scientific explanation for why they worked. The paper quickly became one of the  most cited articles in the medical literature.

But Schildkraut was wrong. Within a few years, as technology expanded our  ability to peer into the brain, it became clear that antidepressants act mostly  by increasing the availability of the neurotransmitter serotonin—rather than  dopamine and norepinephrine, as previously thought. A new generation of  antidepressants—the selective serotonin reuptake inhibitors (S.S.R.I.s),  including Prozac, Zoloft, and Paxil—was developed to target it. The ability to  claim that the drugs targeted a specific chemical imbalance was a marketing boon  as well, assuring consumers that the drugs had a scientific basis. By the  mid-nineties, antidepressants were the best-selling class of prescription  medications in the country. Psychiatry appeared to have found magic bullets of  its own.

The serotonin-imbalance theory, however, has turned out to be just as  inaccurate as Schildkraut’s. While S.S.R.I.s surely alter serotonin metabolism,  those changes do not explain why the drugs work, nor do they explain why they  have proven to be no more effective than placebos in clinical trials. Within  a decade of Prozac’s approval by the F.D.A. in 1987, scientists had concluded  that serotonin was only a finger pointing at one’s mood—that the causes of  depression and the effects of the drugs were far more complex than the  chemical-imbalance theory implied. The ensuing research has mostly yielded more  evidence that the brain, which has more neurons than the Milky Way has stars and  is perhaps one of the most complex objects in the universe, is an elusive target  for drugs.

Despite their continued failure to understand how psychiatric drugs work,  doctors continue to tell patients that their troubles are the result of chemical  imbalances in their brains. As Frank Ayd pointed out, this explanation helps  reassure patients even as it encourages them to take their medicine, and it fits  in perfectly with our expectation that doctors will seek out and destroy the  chemical villains responsible for all of our suffering, both physical and  mental. The theory may not work as science, but it is a devastatingly effective  myth.

Whether or not truthiness, as one might call it, is good medicine remains to  be seen. No one knows how important placebo effects are to successful treatment,  or how exactly to implement them, a topic Michael Specter wrote about in the  magazine in 2011. But the dry pipeline of new drugs bemoaned by Friedman is an  indication that the drug industry has begun to lose faith in the myth it did so  much to create. As Steven Hyman, the former head of the National Institute of  Mental Health, wrote  last year, the notion that “disease mechanisms could … be inferred from drug  action” has succeeded mostly in “capturing the imagination of researchers” and  has become “something of a scientific curse.” Bedazzled by the prospect of  unraveling the mysteries of psychic suffering, researchers have spent recent  decades on a fool’s errand—chasing down chemical imbalances that don’t exist.  And the result, as Friedman put it, is that “it is hard to think of a single  truly novel psychotropic drug that has emerged in the last thirty years.”

Despite the BRAIN initiative recently announced by the  Obama Administration, and the N.I.M.H.’s renewed efforts to stimulate research  on the neurocircuitry of mental disorder, there is nothing on the horizon with  which to replace the old story. Without a new explanatory framework,  drug-company scientists don’t even know where to begin, so it makes no sense for  the industry to stay in the psychiatric-drug business. And if loyalists like  Hyman and Friedman continue to say out loud what they have been saying to each  other for many years—that, as Friedman told Times readers, “just because  an S.S.R.I. antidepressant increases serotonin in the brain and improves mood,  that does not mean that serotonin deficiency is the cause of the disease”—then  consumers might also lose faith in the myth of the chemical imbalance.

Gary  Greenberg is a practicing psychotherapist and the author of “The Book of Woe: The DSM and the Unmaking of Psychiatry.” 

Correction: Due to an editing error, the antidepressant Tofranil was  originally identified as Elavil.

Photograph by Paul Skelcher/Science  Faction/Corbis.