Gene Tierney is interrogated in Otto Preminger’s 1944 film noir, Laura.

20th Century-Fox/Photofest

Stranger Than Fiction

Is there any truth in truth serums?

In 1942, as World War II raged, the United States’ newly established spy agency, the Office of Strategic Services (OSS), asked its scientists to turn their attention from their other projects to the urgent mission of creating a truth drug to interrogate prisoners of war. The OSS first tried out mescaline, a hallucinogenic cactus derivative; they also experimented with scopolamine, a “twilight sleep” drug that had experienced a vogue as a truth serum in the 1920s and 1930s. Both made test subjects too sick to reveal any secrets; so the OSS researchers settled instead on marijuana, using a tasteless, colorless, and odorless liquid extract called TD that could be injected into an unwitting target’s cigarettes.

“TD appears to relax all inhibitions and to deaden the areas of the brain which govern an individual’s discretion and caution,” an OSS report said. Among other effects noted, “the sense of humor is accentuated to the point where any statement or situation can become extremely funny to the subject.”

Driven by panics over Cold War espionage and violent crime, physicians, spies, and cops repeatedly heralded would-be truth serums and mind-reading technologies.

After trying the stuff himself, in 1943 OSS narcotics agent George H. White gave the substance its first field test on August “Little Augie” Del Gracio, a drug-dealing New York gangster with connections to mobster Lucky Luciano. White arranged a meeting with Del Gracio, plied him with spiked cigarettes, and listened as the gangster became “obviously high and extremely garrulous,” according to an account by journalist Dominic Streatfeild. In the ensuing two-hour monologue Del Gracio found himself mysteriously compelled to give an in-depth account of his drug-trafficking network.

White considered the experiment a success, though a second meeting didn’t go as well. Del Gracio got too stoned, felt pins and needles in his hands and feet, and had to lie down for a nap.

A subsequent test that same year on 30 Americans suspected of being Communists was encouraging, with five admitting Communist sympathies. But when the OSS tried to use the substance for actual intelligence gathering, comic disaster befell the interrogator, who accidentally smoked the adulterated cigarettes himself and started ranting about his boss making passes at his wife; the German U-boat captain being questioned coughed up little useful information.

TD was not used again during the war, and intelligence officials eventually concluded that it was not much more effective than giving people alcohol and caffeine. Yet the quasi-scientific search for methods to extract information from unwilling subjects was far from over. Driven by panics over Cold War espionage and violent crime, physicians, spies, and cops repeatedly heralded would-be truth serums and mind-reading technologies. Occasionally these drugs and devices have seemed to work, though psychologists point out they rely on antiquated notions of how memory operates, and legal experts liken them to torture rather than genuine investigative tools. Even now entrepreneurial scientists continue to pursue new technologies that might extract truths from guilty minds.


The oldest substance for extracting secrets may be alcohol, whose tendency to loosen tongues was recognized in ancient Greece and probably earlier. Efforts to develop better tools date to the late 19th century, as insights from the developing field of psychology and growing faith in technology sparked hopes that scientific innovation could address rampant crime and public corruption.

The first modern method for coaxing out buried truths was hypnosis, which fascinated early psychologists with its promise of revealing the roots of mental disorders. As with alcohol, the results of hypnosis were never completely trusted; Sigmund Freud, for example, used the technique to uncover patients’ repressed memories of sexual abuse but eventually admitted that most of the recollections were fantasies. “Such widespread perversion against children is scarcely probable,” he concluded.

Nonetheless, interest in hypnosis grew in the succeeding decades, and it came to be used to help crime victims recall suspects’ license-plate numbers and other facts useful to police. Its apparent success in retrieving blocked memories led courts in the 1970s to admit such evidence, resulting in a number of convictions based on hypnotically induced testimony. Critics fought back by demonstrating the hypnotist’s power to influence what his subjects “remembered” and to create memories of events that had never occurred. In the early 1980s U.S. courts sharply restricted the use of testimony given under hypnosis, though it remains admissible in some states.

The first modern method for coaxing out buried truths was hypnosis, which fascinated early psychologists with its promise of revealing the roots of mental disorders.

In psychology, hypnosis hung on for a while longer despite the disillusioning experiences of Freud and others. The 1980s and 1990s saw many people again claiming to have recovered memories of sexual abuse, leading to criminal and civil suits, countersuits against the accusers’ therapists, and the introduction of “false memory syndrome” into the popular lexicon. Having lost its patina of scientific respectability, hypnosis was largely relegated to smoking-cessation programs and bachelorette parties.

Hypnosis wasn’t entirely useless; it sometimes helped investigators uncover genuine facts. But its reliability could never approach the evidential level the judicial system requires to send someone to jail. Even spies, who more easily set aside legal and ethical niceties, viewed hypnosis skeptically. It couldn’t be used on a resistant subject, and studies indicated that the information obtained “need not be accurate and may in fact contain untruths, despite hypnotic suggestions to the contrary,” a CIA researcher noted in 1960.

Still, the CIA had taken the trouble to investigate hypnosis as a way to induce compliance during interrogations, just as it investigated marijuana, LSD, and a number of other psychoactive substances. The persistent hunt for a working truth technology, despite the faulty theories on which the concept rests, reflects its intense, irrational allure. Truth serums promise to give governments access to dissidents’ thoughts and enemies’ strategies, employers a tool to manage workers, and police and prosecutors a powerful crime-fighting weapon, without the messiness and consequences of physical coercion. They tantalize with a vision of a direct link to the seat of an individual’s thoughts and personality, and with it an unprecedented, Godlike view into their souls.

As Martin Lee and Bruce Shlain write in Acid Dreams, their history of LSD, the CIA’s quest to develop truth drugs “resembled a skewed version of a familiar mythological theme from which such images as the Philosopher’s Stone and the Fountain of Youth derive—that through touching or ingesting something one can acquire wisdom, immortality, or eternal peace.”


The great power of hypnosis, and its fatal flaw, is the suggestibility it induces. The hypnotist may command confession (“Tell me about your childhood!”) but also may create new memories (“Your father beat you, right?”). The control it grants the therapist is reminiscent of the power imparted by the polygraph exam, another process that requires the administrator to wade through the muck of unspoken psychic flow in search of solid evidence. Both technologies involve complicated negotiations with the person being interrogated and the interpretation of subtle, ambiguous signals.

By contrast, a truth serum would theoretically skip such mediation and prompt the delivery of facts directly through the subject’s voice: just give an injection and record the results. That was the goal of the OSS’s cannabis trials and of experiments with scopolamine, the one truth drug that gained wide acceptance for a time.

Scopolamine, derived from plants in the nightshade family, had for years been used to put women giving birth into “twilight sleep,” a state of sedation in which patients are numb to pain but still conscious. In 1916, Texas obstetrician Robert House noticed another of the drug’s effects. He asked the husband of a sedated patient where to find scales in the house to weigh the newborn, and the sedated woman suddenly answered the question. According to historian Alison Winter, the drug apparently eased both emotional and physical pain, including the psychological discomfort of revealing upsetting or otherwise repressed facts. It was initially used by interrogators to confirm assertions of innocence by criminal suspects; if they kept to their story after being drugged, their claims were judged valid and they could be released.

In her book Memory, Winter maps the drug’s successes and failures. In 1922 it was used to “confirm” the innocence of an accused murderer in Dallas; the witnesses against him refused to take the scopolamine test, and he was released. In 1924 it drew confessions to a string of two dozen ax murders in Alabama. Yet a chauffeur who confessed to a kidnapping and murder in Hawaii subsequently maintained his innocence during a second scopolamine interview, and a different suspect was convicted of the crime.

Among the various ways memory has been imagined, the file cabinet packed with folders is perhaps the most deeply embedded in modern society.

To explain how the drug worked, House, who was not a psychologist or neurologist, drew from common, unproven conceptions of the mind. The hearing nerve, he explained during a 1926 deposition, mechanically sends a question to “that part of the brain where the answer is stored for future use, and . . . the brain sends the answer to the nerve of the tongue.” When a person is asked his name, “His will power can prevent the tongue from articulating the name, and the power of reason can also . . . make the tongue tell a lie, but when the will power and the power to reason are removed, the replies are automatic.”

In House’s model the information the interviewer seeks is a discrete object stored permanently and unchangingly in the brain, like a letter in a filing cabinet. The nerves act like pulleys, or perhaps telephone lines, dumbly conveying the wanted data to and fro. Only the intervention of two other forces—the will and reason-driven imagination—can prevent this otherwise automatic action, and suppressing them with drugs frees the nerve to operate perfectly. Scopolamine not only eased confessions but also confirmed that professions of innocence were genuine and not forced by the will.

Among the various ways memory has been imagined, the file cabinet packed with folders is perhaps the most deeply embedded in modern society. Other metaphors include the metal etching, photograph, filmstrip or videotape, and computer circuit. But while memory is still not completely understood, decades of neuroscience research have shown that these popular notions are far too simplistic.


Remembering is actually a complex electrochemical process involving living cells in multiple parts of the brain. Different aspects of memories—such as physical activity, geographical location, and emotion—are produced in different places as they rise to consciousness as a single memory. The quality of recollection depends on the presence of cues related to the original processing of the information, such as smells and feelings.

Memories are also frail and changeable. They are pared down over time, losing the specificity so crucial in court testimony, and eventually become harder to retrieve. They may be misattributed, with the face of an innocent person replacing a criminal suspect in a witness’s mind. They are fabricated fairly easily: when researchers ask people about events that never happened, a significant minority say they remember the events, especially after repeated questioning. Memories also change to better match a person’s beliefs or mood.

Remembering, then, might better be called reconstructing. Each time a person recalls an event or undergoes a related stimulus, the memory may change, as if it were being experienced and remembered anew. Research by neuroscientist Daniela Schiller and others has shown that memories can be altered by having the person recall and “reconsolidate” them with a different emotional tinting. “Our memory is not like a video camera,” says Donna Jo Bridge, a neuroscientist at Northwestern University who has studied memory change in the hippocampus region of the brain. “Your memory reframes and edits events to create a story to fit your current world. It’s built to be current.”

Memory was even less well understood during House’s time than it is now, and the lack of a credible explanation for how scopolamine worked bred skepticism in the press and judiciary. House’s 1926 deposition concerned a scopolamine interview he had conducted with an alleged rapist. He concluded the man was innocent, but the interview was rejected as unscientific and the suspect was convicted, according to Winter. An appeal ended with a scathing dismissal of the case by Judge Robert Walker Franklin, who called the truth serum’s origins “nebulous” and its effect “uncertain,” and castigated its advocates for inducing credulous folk to “believe in the magic powers of philters, potions, and cures by faith.”

The lack of a credible explanation for how scopolamine worked bred skepticism in the press and judiciary.

House had envisioned scopolamine as an advance that would end false convictions and prevent law-enforcement abuses by providing a humane, objective way to produce the truth. But when he died in 1930, his successors instead took up the drug as a brute weapon in their fight against crime.

In this way it recalls the polygraph, which does a poor job of detecting lies but derives great power from the fear and confessional mode it induces. Both technologies came to be seen as just more varieties of the “third degree” that police used on criminal suspects, along with chloroform and beatings with rubber hoses. In a 1931 U.S. government report a policeman described how officers would tell a resistant suspect that, if necessary, they could simply knock him unconscious by drugging his coffee and then inject him with scopolamine. “In many instances, the terrified suspect talked to avoid being tested with the truth serum,” the officer said.

Scopolamine’s popularity eventually faded owing in part to its distracting side effects, such as hallucinations and painfully dry mouth. The CIA concluded that it did not work as a truth serum, and its clinical use as a “twilight sleep” drug ended because of the danger to patients. Scopolamine survives as a treatment for motion sickness and certain other ailments, and as a knockout drug: in Colombia, scammers have used it on tourists and other victims to turn them into temporary “zombies” who withdraw cash from automatic teller machines on command.


No other truth drug had as successful a run as scopolamine, but many others were given a try. Sketchy intelligence and press reports in the 1940s and 1950s suggested the Chinese and Russians were experimenting not just with scopolamine but also with amphetamines, such as Benzedrine, and barbiturates, such as sodium thiopental (also known as Sodium Pentothal). In one such account, Lieutenant John Ori, an American, said he found a sweet white powder in his food while detained in a prison camp during the Korean War. After eating, “I found myself talking and talking,” he told a journalist. “I was hardly able to control what I was saying. I talked a blue streak.”

The United States responded with its own research programs, continuations of the effort that had George White testing cannabis extract on August Del Gracio. The general approach was to develop a stronger and more carefully controlled version of the alcohol-and-caffeine effect: a shot of Pentothal to knock down the subject’s self-censorship faculties and a touch of Benzedrine to perk him up and get him blabbing. Pentothal’s tendency to overcome memory blocks was already well known; it was used widely during World War II to help shell-shocked soldiers remember and recover from traumatic experiences, much as Freud had done more painstakingly with hypnosis and free-association conversation.

As a truth serum Pentothal proved no more useful than scopolamine, though it is safer to administer and still has its advocates. After 9/11, as intelligence agencies turned to secret prisons and waterboarding of suspected terrorists, former CIA and FBI director William Webster said the United States should consider using Pentothal on al-Qaeda and Taliban detainees imprisoned at Guantanamo Bay. More recently, the judge in the trial of mass shooter James Holmes alarmed legal experts when he ruled Holmes could undergo so-called narcoanalysis, probably using a barbiturate, to determine whether Holmes was sane when he killed 12 people in a Colorado movie theater in 2012. It’s unclear whether the examination ever took place.

In India, one of the few nations where some courts allow narcoanalysis, barbiturates are occasionally used in high-profile cases. In Uttar Pradesh a wealthy businessman and his servant were injected with a truth serum in 2007 and supposedly confessed to the gruesome murders of 17 children and women. The one surviving gunman from the 2008 Mumbai attacks was also reportedly interrogated while on Pentothal; he was later convicted based on other evidence and hanged.

A few would-be truth serums departed from the depressant-and-stimulant formula, including marijuana and LSD, which in the 1950s the CIA briefly considered fabulously successful at eliciting secrets. “For years they had searched, and now they were on the verge of finding the Holy Grail of the cloak-and-dagger trade,” Lee and Shlain write in Acid Dreams. “As one CIA officer recalled, ‘We had thought at first that this was the secret that was going to unlock the universe.’ ” Of course LSD failed, too; it induced “marked anxiety and loss of reality contact,” along with hallucinations and other side effects, making interrogation difficult. The drug went on to a much grander career as a pathway to spiritual revelation and a touchstone of hippie culture through the 1960s and 1970s.

‘We had thought at first that [LSD] was the secret that was going to unlock the universe.’

The CIA never did find a real truth serum, and the U.S. Supreme Court ultimately barred confessions under scopolamine and similar substances in 1963, reasoning that the resulting testimony was unconstitutionally coerced. Such drugs are similarly banned in most countries.

In another context truth serums have been massively successful. They have become a staple of popular entertainment, an irresistible dramatic tool in novels, movies, and TV. They can serve as a traditional poison or as a device that withholds or releases new information to advance the story. In the 1947 film noir High Wall, Pentothal fails to help a man remember the circumstances of his wife’s murder but in the end makes the true killer confess. In Big Jim McLain, a heavy-handed 1952 political thriller starring John Wayne, a villainous Communist psychiatrist tries to use the drug on a House Un-American Activities Committee investigator but accidentally kills him.

At times, truth drugs have been depicted in ways that illustrate their unreliability. In Thomas Pynchon’s 1973 novel Gravity’s Rainbow, the barbiturate sodium amytal induces fantastical LSD-type visions loaded with ambiguous cultural significance. In the popular 1961 film The Guns of Navarone, a British officer is deliberately given false information so he can blab it to the Germans when they administer scopolamine.

Truth serums are also often used to comic effect, as in Meet the Fockers, where Robert De Niro plays a retired CIA officer and a caricature of a suspicious father-in-law. When he injects Pentothal into his son-in-law, played by Ben Stiller, the younger man confesses his lust for his mother-in-law to a roomful of relatives. In another, truer-to-life variation used on the TV show Get Smart and in many other productions, the drug works too well: the character reveals everything on his mind, including random trivia, boring the interrogators to tears.


At the heart of truth-detection techniques is an enduring power fantasy that repeatedly adapts to the latest quasi-scientific belief and technological innovation. From medieval racks to 17th-century dunking of witches—whose buoyancy indicated their accursed nature—through to the blood and breath measurements of polygraphy and the theory of facial microexpressions, policing organizations have latched onto the latest fad to extend their intelligence-gathering capabilities deeper into their subjects’ bodies.

Lately the hunt has focused on the brain’s electrochemical properties. Using functional magnetic resonance imaging, or fMRI, researchers can track blood flow in the brain and watch in real time as neurons react to external experiences.

Such experiments occasionally give striking results, as in 2011, when researchers at the University of California, Berkeley, attempted to convert brain activity into videos of what people were seeing. It was quite a bit more laborious than straightforward mind reading. First, while subjects watched movie trailers, researchers conducted brain scans on the volunteers to determine which neuron patterns corresponded to which viewed images. Then the researchers had the subjects watch new videos and recorded their brain activity again. Finally, drawing from a selection of YouTube clips, the researchers extracted images that seemed to correspond to the new brain recordings and combined those images into an “average” video.

The re-created videos, which may be viewed online, are little better than blurry, oozing inkblots that roughly correspond to faces and shapes in the trailers. But for an optimistic police investigator or prosecutor, they may hint at a future in which witnesses’ memories of events and conversations can be downloaded and played for a jury, like surveillance video of a gas-station robbery. The dream of unmediated access to the mind would finally be realized. But the risks that doomed older technologies—the malleability of memory and the tendency to use such devices coercively—still apply.

At any rate, that future is a long way off. So far fMRI’s only investigative use has been as a dodgy would-be lie detector that a couple of companies are trying to market. To use one, a test subject lays his head in a bulky, donut-shaped fMRI machine; an examiner asks test questions, noting the different brain-activity patterns corresponding to true and false statements, and then asks about the crime or issue being investigated.

In 2006 an accused murderer in Maryland tried to get his fMRI interview results admitted to demonstrate his innocence, but the judge denied the request, ruling there was no scientific consensus that the technology is valid. A comprehensive review in 2013 identified several fundamental questions fMRI lie detection would have to answer before it could be considered for legal use, including whether it is actually detecting lying and not some other mental phenomenon, whether the machine can be tricked, and whether it works outside the lab.

“Contrary to public expectations, lie detectors like fMRI are not mind readers, do not actually detect deception, and will never provide details of what has actually happened in complicated cases,” the study authors wrote. “Rather, they merely detect and measure manifestations of thoughts through changes in oxygenated blood which proponents consider denotes lying.”

Another method dubbed “brain fingerprinting” has gained slightly more acceptance. It uses the electroencephalogram, or EEG, to detect electrical activity in the brain via electrodes on the scalp. In the mid-1960s scientists discovered a pattern of signals that occurs about 300 milliseconds after a person experiences a familiar stimulus. In brain fingerprinting a suspect is read a series of true and false statements, and if his brain produces the signal after he hears a confirmed truth (“She was strangled with a yellow rope”) that means he knows the details of the crime and must be guilty, according to the technology’s proponents.

Contrary to public expectations, lie detectors like fMRI are not mind readers, do not actually detect deception, and will never provide details of what has actually happened in complicated cases.

Brain fingerprinting has met stiff criticism for having, at best, weak support from scientific studies. In addition, it requires keeping some criminal evidence hidden until it is sprung on the suspect. Its results haven’t been admitted in U.S. courtrooms, except in one Iowa case where it played a minor role in exonerating a man wrongly convicted of murder.

But the technology has secured a small foothold: a company called Brainwave Science says it has sold EEG gizmos to police in Singapore and Florida, and a similar test has gained a striking level of acceptance in India, where in 2008 a woman accused of poisoning her former fiancé with arsenic was convicted based on EEG results. “The relevant nooks of her brain where memories are thought to be stored buzzed when the crime was recounted,” a state investigator told the New York Times.

For now none of these newfangled devices or any other truth technology appears likely to win acceptance by Western legal systems. At most they are used on rare occasions to assess a suspect’s mental condition.

The picture is naturally less clear with intelligence agencies. According to journalist Mark Bowden, after 9/11 the intense pressure to prevent further terrorist attacks led agents back to the drugs the OSS experimented with 70 years ago, though without any pretense that they would reliably provoke useful admissions. At best they could make a terrorist trip over his lies or become addicted and dependent on his interrogators.

“According to my intelligence sources, drugs are today sometimes used to assist in critical interrogations, and the preferred ones are methamphetamines tempered with barbiturates and cannabis,” Bowden wrote in 2003. “These tools can help, but they are only as effective as the interrogator.”