November|December 2004




IN 1997, A MASS MARKET POTBOILER called The Truth Machine by James Halperin anticipated what would happen if a foolproof method for detecting lies were developed. Human conflicts would be easily resolved and war would disappear, as people would be forced to negotiate with absolute honesty. Violent crime would abate, and no innocent man would ever again be executed for a murder he didn't commit. Along the way, the technology's inventor—a brilliant, exercise-driven Harvard graduate—would become the richest man on the planet.

The only memorable aspect of the novel was its prescience. The inventor in the book bears an uncanny, and unwitting, resemblance to Larry Farwell, a brilliant, exercise-driven Harvard graduate who, in 2001, unveiled the Farwell Truth Detector.

Farwell's invention, which he now calls brain fingerprinting, works like this: A headband of electrodes is placed on a subject, who watches words or pictures flash across a computer screen. Some of the images are meant to stimulate memories, which cause the brain to fire off an electrical response 300 milliseconds after the stimulus. It is known as the "P300 effect," in which the P stands for positive. The stimuli come in three categories: "target" stimuli (details of an activity that would be known to the subject), irrelevant stimuli (which would not be expected to elicit a response), and "probe" stimuli (phrases or pictures supposedly known only to a select few, like the perpetrator and investigators of a crime). If a suspect exhibits a P300 response to a probe stimulus, he is presumed guilty. If not, he is presumed innocent.

Farwell is convinced that his technique is nearly infallible. He has offered his services, sometimes pro bono, in several high-profile criminal trials, and he has assisted numerous police departments. As he tells it, he has helped exonerate an Alexandria, Va., police officer accused of drug dealing, confirmed the guilt of a serial killer in Macon County, Mo., and helped reverse a murder conviction and secure the release of a man who had served 24 years of a life sentence in Iowa. These days, Farwell is hoping that his test will exonerate Jimmie Ray Slaughter, a convicted murderer on death row in Oklahoma, and help position brain fingerprinting as the most important weapon in the forensic science arsenal. "In every case in the future when someone is arrested, they're going to say, 'Look, I'm innocent,' " Farwell predicted. "'Don't tell me anything about the crime. Give me a brain fingerprint test!' "

Farwell's enthusiasm has been contagious. Television programs like 60 Minutes have devoted segments to brain fingerprinting and newspapers around the world have written glowingly about the potential applications of Farwell's method, which range from freeing the innocent to fighting terrorism, identifying insurance fraud, and providing early diagnoses of Alzheimer's disease.

The hype, claim Farwell's critics, has outpaced the science. Many scientists accuse Farwell of making misleading and exaggerated claims. They caution that brain fingerprinting is in its infant stage and may never result in a reliable polygraph. John J. B. Allen, a professor of psychology and neuroscience at the University of Arizona, replicated Farwell's research in the past. But in a new study, "Brain Fingerprinting: Is it Ready for Prime Time?" Allen has answered no. Last year, he conducted his own experiment using a mock crime scenario, and found that Farwell's method accurately identified "guilty persons" only 50 percent of the time.

Farwell sweeps aside such objections, insisting that there is too much at stake to allow the niceties of scientific research to get in his way. He has more important matters on his mind, such as the backlog of more than 400 cases, many involving defendants on death row, clamoring for a chance to take the test that could prove their innocence. "Let's take advantage of the life-saving technology we have," Farwell said. "When I met Jimmie Ray Slaughter, he had about a 90-day life expectancy. I had to say yes or no. I couldn't say wait."

JIMMIE RAY SLAUGHTER'S TRIAL FOR MURDER in 1994 lasted five months. The transcript of the trial included more than 60,000 pages filled with gruesome details about Slaughter's ex-lover, Melody Sue Wuertz, and their 11-month-old daughter, Jessica. They were both shot in the head and found in Wuertz's house in Edmond, Okla. The mother's abdomen was carved with occult symbols, her genitalia mutilated. Prosecutors theorized that the 29-year-old Wuertz had been shot in the back and deliberately paralyzed so that she would be unable to intervene as she was forced to watch the slaying of her baby.

Slaughter, a psychiatric nurse who served as a captain during the Gulf War, had an alibi. On the day of the murders in July 1991, he was stationed 300 miles away from Edmond, in Fort Riley, Kan., where he spent the day with his wife and two children. But the prosecution countered that the family was lying to protect Slaughter. He was convicted of first-degree murder and sentenced to death.

Last year, after one of Slaughter's lawyers saw a 60 Minutes episode about brain fingerprinting, she contacted Farwell, who traveled to Oklahoma State Penitentiary in February with six steamer trunks filled with equipment. Free of charge, Farwell performed his test on Slaughter, and Slaughter showed no memory of the crime. A big part of Slaughter's final appeal, now pending before the state's court of criminal appeals, is riding on that test.

In Macon County, Mo., Sheriff Robert Dawson learned about the method from his secretary, who had also seen it featured on television. In 1999, Dawson ordered a test on J. B. Grinder, accused of raping and murdering a 25-year-old woman. Grinder had admitted and denied the allegations so many times that, according to Dawson, "We didn't know what to believe anymore." Confronted with the test results, which seemed to confirm one of Grinder's many confessions, Grinder pled guilty to the charges and also admitted to killing three other girls in Arkansas. When another murder investigation ran into problems earlier this year, Dawson turned again to brain fingerprinting. He refrained from discussing the details of the case with the suspect and with the media so that the P300 probes would be valid. While the suspect denied knowing anything about the case, Farwell's test suggested otherwise.

His method's biggest breakthrough, according to Farwell, was its role in freeing convicted murderer Terry Harrington, who had been serving a life sentence in Iowa State Penitentiary for killing a night watchman in 1977. In 2001, Harrington requested a new trial on several grounds, including conflicting testimony in the original trial.

Farwell was faced with an immediate and obvious problem: 24 years had passed since the trial. Evidence had been presented and transcripts published long ago; the details of the crime had long since come to light. What memories of the crime were left to probe? But Farwell combed the transcripts and came up with obscure details about which to test Harrington. High weeds that Harrington would have had to run through after fleeing from the scene of the crime. A ditch. Some parked cars. On the strength of those prompts, Farwell concluded that Harrington had no memories of them and was, by inference, innocent of the crime.

Farwell has hailed the case as a landmark victory, noting that it is the first and only time a court has ruled brain fingerprinting admissible as evidence. That claim has been repeated in newspapers and magazines and on TV, but it exaggerates what happened.

The presiding district judge, Timothy O'Grady, did agree to consider brain fingerprinting as scientific evidence. But he questioned Farwell's methods and ruled that the test results did not warrant a new trial. By Farwell's own account, the "probes" intended to jog a suspect's memory should be "salient details," known only to the perpetrator and investigators of a crime. The probes that Farwell chose for Harrington did not meet those standards, the judge found. They were not significant enough to be remembered 24 years later. In addition, all the details had come up at trial and had been depicted in photographs that had been passed among the jury, the prosecution, and the defense. Harrington likely saw the photographs, heard the evidence, and discussed many of the details with his lawyers. If he now had no memory of these details, as his brain fingerprint showed, what did that mean? "Brain fingerprinting is a bunch of hogwash," said Matt Wilber, the district attorney in the county where Harrington's case originated.

Harrington was granted a new trial when it was discovered that some of the original police reports in the case had been missing at his initial trial. By 2001, however, most of the witnesses against Harrington had either died or had been discredited. Finally, when a key witness heard that Harrington had "passed" his brain fingerprinting test, he recanted his testimony and the prosecution threw up its hands. Harrington was set free.

Among the expert witnesses who testified in that hearing, Emanuel Donchin was one of the most vociferous in opposing brain fingerprinting. Donchin was Farwell's dissertation adviser at the University of Illinois from 1984 to 1992. Together, in 1991, they co-authored a seminal study about the P300 effect, in which they demonstrated that a "guilty" person's brain produced an electrical charge when presented with an incriminating detail in a mock trial scenario.

But while Donchin, now at the University of South Florida, stands by the science of their original paper, he decries what Farwell has done with it. "Larry has made all kinds of extravagant claims and gross oversimplifications," he said, criticizing Farwell's claim that his method was accepted by the Iowa Supreme Court. "The man is a businessman and therefore what you get are the ethics of business and standards of marketing."

BEGINNING IN ELEMENTARY SCHOOL, Larry Farwell wanted to be a physicist like his father, George, who worked on the Manhattan Project. "Physicists have to work hard," he scrawled in big block letters on a piece of paper that his mother saved, "but I don't care."

He graduated from Harvard University in 1973 with a degree in physics and an interest in transcendental meditation. In his 1999 book How Consciousness Commands Matter, he claimed that he could move matter by the power of his mind alone. "My task was to command matter through consciousness . . . using nothing but the influence of consciousness alone," he wrote. Those mental efforts, he continued, caused three words to appear in large blue letters on his computer screen: Anything is Possible.

Farwell made similar claims after graduating from Harvard. Before enrolling in graduate school at the University of Illinois, he hung around the University of Washington, where his father was a professor and administrator. Once, during a graduate psychology seminar, he told a group of professors that he knew how to levitate. When asked to prove it, he demurred, saying he could do so only in the presence of believers.

As a graduate student, Farwell began working with Donchin, a nationally respected psychologist. They designed an early brain-computer program to allow severely paralyzed patients to communicate by linking their thought mechanism to a computer. The prototype was based on the same P300 effect that is involved in brain fingerprinting, which Farwell often worked on late at night. With electrodes glued to his scalp, Farwell would dash back and forth between the subject's chair and his computer, measuring his own brain waves.

Farwell recognized the potential right away—and so, initially, did the CIA. In the early 1990s, it gave Farwell $1 million to research and develop his invention. He designed one study that sought to find unidentified U.S. Navy medics within a group of people, and then he repeated that study with FBI agents. (The CIA didn't want to use their own agents for security reasons.) In both situations, he tested for information that only the operatives would recognize; both studies, he said, achieved 100 percent accuracy.

The CIA ultimately concluded that brain fingerprinting was not worth further investment. The agency was looking for a general tool to screen current and prospective employees, one that didn't require such a detailed level of prior information. In October 2001, a General Accounting Office report on brain fingerprinting found that both the CIA and FBI believed the technology had "limited applicability and usefulness." Each cited Farwell's reluctance to provide algorithmic information about how his test works, on the grounds that his technique was proprietary. "One of the problems you have with new scientific techniques is that the developer often is not the best investigator of the validity of that technique," said David Faigman, author of Legal Alchemy: The Use and Misuse of Science in the Law. "They tend to be true believers and tend to interpret data as supporting their hypotheses when there is so much at stake."

Memory is at the heart of Farwell's work, and it defies easy analysis. Despite a widespread perception of the brain as the ultimate video tape player, memory is far more complicated. According to Daniel Schacter, chair of the psychology department at Harvard, memories are constantly recreated and reconstructed rather than retrieved from a mental data bank.

Despite abundant research about memory, as well as ongoing testing for drugs to enhance and erase it, huge gaps in understanding remain. For example, how does a heightened state of arousal affect the memory process? If a man kills his girlfriend in a rage, will he remember that she was wearing a chartreuse blouse? What if he was high on drugs or alcohol at the time of the crime? "We don't know enough of how memories are formed during crimes," said William Iacono, a professor of psychology at University of Minnesota who testified on Farwell's behalf in the Harrington case.

Iacono believes that a positive result on the test is almost always accurate—how else to explain a result showing that someone has a stored memory of an event that is not a matter of general knowledge? Much harder to explain is the absence of the P300 when other evidence indicates it should appear, which is why, he said, it would be easy to get a false negative. A guilty person might pass a brain fingerprinting test without difficulty, either because he no longer remembers the details or because he didn't notice them in the first place.

"There are many reasons why a person may not have memory of a crime, apart from just not remembering," he explained. "If you're committing a crime, you're thinking about how to cover your tracks, how you're going to get out of there without being seen or leaving evidence behind. It's not clear what you're going to remember."

THOUGH THE BRAIN FINGERPRINTING MAY NEED TINKERING, most scientists believe we will one day be able to read minds—or at least expose lies. At the University of Pennsylvania, Dr. Daniel Langleben is using functional magnetic-resonance imaging, a kind of brain scanning, to identify distinct patterns of brain activity that occur when people lie. He tests his subjects by asking them to pick a card from a deck, and then lie to him about what it is. If they are able to convince the computer that they're telling the truth, he pays them $20. Langleben has found that specific parts of the brain show increased activity when a person lies. Britton Chance, who also works at the University of Pennsylvania, has developed a cognosensor, a device that measures brain patterns using infrared light. His research shows that the brain reflects more light when a person lies. As Jonathan Moreno, director of the Center for Biomedical Ethics at the University of Virginia, put it, "Science is catching up with fiction and challenging our ability to deal with it."

The novelist who anticipated Farwell's method dismisses such concerns. By 2040, James Halperin predicts, "truth machines" will become so routine that people will wear them on their lapels. Earlier this year, he invested in Farwell's company.