“Brain Fingerprinting” Project at University of Arizona

Eric Swedlund reports for the Arizona Daily Star in an article titled, “UA on Security’s Cutting Edge.” Excerpt:

John Allen, a psychology associate professor, will try to answer this question: “Is Brain ‘Fingerprinting’ Ready for Prime Time?”

Conventional polygraphs measure factors such as heart rate and sweaty palms to determine nervousness or anxiety, but “brain fingerprinting” examines brain waves for particular responses associated with recognition.

The technique is currently being used to assess memory, but it has potential applications in criminal investigations.

Allen will test subjects in a mock crime scenario and note how recognition of a specific fact will elicit different brain activity. Applications could involve testing spies to determine if they recognize particular acronyms, pictures or phone numbers.

Allen said the procedure must accurately identify guilty people without incriminating the innocent. Tests thus far indicate about 90 percent accuracy on both accounts.

“In any attempt to increase homeland security, you have to protect the citizenry against false accusations,” Allen said.

“$4M Project at UA Targets Deception”

Eric Swedlund reports for the Arizona Daily Star on a taxpayer-funded research program at the University of Arizona. Excerpt:

To boost national security, the Defense Department is paying for a $4 million UA research project on detecting deceit in communication.

In the electronic communication age, the military faces more challenges because analysts cannot always rely on conventional models of lie detection.

“We know deception is commonplace everywhere,” from daily conversations to military endeavors, said Judee Burgoon, the principal investigator.

Burgoon said the project’s significance “has grown astronomically” since Sept. 11. “Obviously it’s extremely timely.”

“How might we have been able to possibly . . . provide earlier intelligence” about attacks? she asked.

“Information power is more important than firepower.”

Military intelligence officers and criminal investigators would like to have fully automated tools, Burgoon said. There was great hope for the polygraph and other devices, but none are fully reliable.

“The best hope is a highly trained human augmented by tools,” Burgoon said. “Humans are extraordinary information processors, especially when they’re well-trained.”

Burgoon, a communication professor and director of human communication research at the University of Arizona’s Center for the Management of Information, reports to the Air Force Office of Scientific Research on the five-year project.

Collaborating with Burgoon and her UA team are researchers at Florida State University, Michigan State University and Wright-Patterson Air Force Base in Ohio.

One goal is to devise computer software or hardware tools that could detect potentially deceptive patterns or characteristics in electronic transmissions such as e-mail or cellular phone conversations.

Computer software could offer alerts at different levels of danger by searching for words or phrases that warrant further investigation. Such a program could offer only a probability of truthful discourse and couldn’t determine if a particular message or person is deceptive.

“Virtual Lies Face Foolproof Software”

Fiona Harvey of the Financial Times reports on new software that purports to detect deception in electronic text messages. Excerpt:

Software that can detect when people are lying in their e-mails sounds a bit far-fetched, but its manufacturers declare it is true.

SAS Institute, which makes fraud-detection systems for banks and phone companies, will on Monday announce a product that can sift through e-mails and other electronic text to catch elusive nuances such as tone.

“The patterns in people’s language change when they are uncertain or lying,” says Peter Dorrington, business solutions manager at SAS. “We can compare basic patterns in words and grammatical structures versus benchmarks to detect likely lies.”

Perhaps SAS business solutions manager Peter Dorrington’s claims regarding this new software should be scanned for likely lies.

“Medical Detection of False Witness”

Brandon Spun reports in the 4 February 2002 issue of Insight magazine on experimental detection of deception/concealed information techniques. Spun specifically addresses Dr. Larry Farwell’s “brain fingerprinting” technique, Dr. Daniel Langleben’s fMRI research, and Dr. James Levine and collaborators’ thermal imaging technology.

“Thermal Camera May Detect Lying”

United Press International (UPI) reports on the development of a new system that purports to detect lies. Excerpt:

An experimental new lie detector that measures sudden flashes of heat from around the eyes may soon provide another line of defense against terrorism . . .

“On the other hand, when thinking about the possibility of someone with explosives in his shoes boarding a plane, given the technology’s security potential, I think most of us would want this application to be accelerated as quickly as possible,” Levine added. “We’re making advances in science, and I think the ethical issues need to be dealt with when the advances are being made. Otherwise ethics gets left behind,” he said.

The device consists of a high-definition thermal imaging camera the size of a shoebox. The scientists also have developed a miniaturized version of the camera, roughly the diameter of a postage stamp. Both are hooked up to a filing cabinet’s worth of computer hardware.

“As people lie, there is a massive increase in blood flow around the eyes, and associated with that there is sudden warming around the eyes, where the color changes to white in the thermal imaging system,” Levine explained.

The research team had 20 volunteers commit a mock crime and then assert innocence under experimental conditions at the Department of Defense Polygraph Institute in Fort Jackson, S.C.

Eight of the volunteers stabbed a mannequin and stole $20 from it, while the rest had no knowledge of the crime. The device accurately detected lying roughly 80 percent of the time, a precision level comparable to standard lie detecting polygraph tests performed by experts.

Even if the device discussed in the article were 80% accurate, the extremely low base rate of terrorists would make it completely useless as a tool to screen airline passengers.

“The Lie Detector That Scans Your Brain”

Clive Thompson reports on Dr. Lawrence A. Farwell’s brain fingerprinting technique in this New York Times magazine article. Excerpt:

The police have tried for years to get into the heads of criminals. But the accuracy of polygraphs, which measure pulse rates and blood pressure, has frequently been questioned — since steely-nerved liars can quell these physiological cues. Now a new technique called “brain mapping” promises to add a new (if creepy) weapon to crime fighting: a device that can scan the brain of suspects and hunt for incriminating thoughts.

The idea of monitoring brain waves isn’t new. Scientists have long known that certain recognizable waves occur when people are surprised, pleased or frightened. But recently the technique has become much more precise. At Brain Fingerprinting Laboratories, a company in Fairfield, Iowa, the chief scientist, Lawrence Farwell, interrogates suspects by checking their EEG’s for “P300 waves.” These waves are produced when the brain encounters words or images that it recognizes; thus the police, Farwell claims, can present a suspect with information that only a criminal would know and see if the brain recognizes it.

“Climbing Inside the Criminal Mind”

Sarah Sturman Dale reports on Dr. Larry Farwell’s brain fingerprinting technique in this short article. Excerpt:

He went to Harvard, works in Iowa and loves swing dancing. That’s not the typical profile of an anticrime crusader, but Lawrence Farwell is an unusual guy. While developing technology that would allow the vocally paralyzed to speak, he stumbled across a trove of seemingly extraneous signals stored in the brain. He began looking for a way to put that information to use. Result: a new forensic technology he calls brain fingerprinting.

“Decoding Minds, Foiling Adversaries”

Sharon Berry reports in the October issue of SIGNAL magazine on a new technology which which Dr. John D. Norseen of Lockeed Martin Aeronautics Co. says may be able to determine if a person is lying. Excerpt:

Whether a threat comes from pilot error or enemy aggression, scientists are finding that multisensor mapping and analysis of the brain lead to systems with human-machine interfaces that can correct human error, aid counterintelligence work and guard against attacks.

A technology, known as bio-fusion, combines sensors to examine biological systems to understand how information and neural structures produce thought and to display the thought in mathematical terms. By creating an advanced database containing these terms, researchers now can look at brain activity and determine if a person is lying, receiving instructions incorrectly or concentrating on certain thought types that may indicate aggression.

[Dr. Norseen]…has been asked by military and law enforcement agencies to show how brainprints can be used to determine probable cause, which could be used for an anti-terrorism situation. “If someone is walking through the airport and he goes through the security checkpoint and we get a feeling that this person is preoccupied with certain numbers or certain thought types that may indicate hostility or aggression we could ask him questions and verify the answers. Then it gives you probable cause to say, ‘Sir/Ma’am, may we step aside with you and ask you additional questions?’ It allows you to find a problem set within a large group.” Norseen is confident that if such a system were fully developed it would be accepted if it meant everyone would be safer at the airport gate. The data he collects may not only show probable cause but also truth verification, he adds. The brain, which uses energy, does not want to expend it needlessly, he says. If someone is telling the truth, it is kept on the outside portion of the brain in low-energy domain areas of the brain. “If someone starts to light up in more areas of the brain and at a higher energy level, it means that the person is now starting to confabulate or obfuscate.” Research so far indicates a 90 to 95 percent accuracy rate.

“Truth and Justice, by the Blip of a Brainwave”

Barnaby J. Feder reports for the New York Times about Dr. Lawrence A. Farwell’s brain fingerprinting technique. Excerpt:

Since the World Trade Center bombing in 1993, Dr. Lawrence A. Farwell has been arguing that terrorist operations can be investigated through careful monitoring of the brain waves emitted by suspects during interrogation. The claim did not get very far with the Federal Bureau of Investigation or any other major law enforcement agency then.

Now, since the Sept. 11 terrorist attacks, Dr. Farwell and a number of supporters are pressing for a much more thorough consideration.

Their effort is another instance of the typical innovator’s natural impulse to dress up old visions in front- page news. But Dr. Farwell’s investigative technique, which he likes to call brain fingerprinting, may also be seen as a typical story of conflict over how to develop real-world applications from promising bodies of research.

Dr. Farwell’s concept is an offspring of a vast body of research on the electrical activity of the brain. Most of the research has focused on easily observed phenomena like alpha and beta waves, which have been respectively linked to activities including sleep and heightened alertness. But one subset beginning in the mid-1960’s homed in on extremely brief electrical wave patterns associated with recognition of familiar sounds, smells and sights.

The most widely studied of such event-related changes is a split-second bump in electrical activity that starts anywhere from 300 milliseconds to 800 milliseconds after a recognized stimulus. Many researchers have studied how the bump, called p300, appears to be affected by various diseases of the brain. Some have pondered how it may be used to help severely disabled people control computers. Starting in the 1980’s, Dr. Farwell and a few other neuroscientists began exploring whether the phenomenon could be used to detect concealed knowledge.

One reason for their interest is that the most widely used lie detectors, known as polygraphs, have long been considered an embarrassment by many scientists. Polygraphy measures a suite of physical reactions to interrogation. The underlying premise is that people being questioned about crimes in which they were involved will involuntarily exhibit telltale increases in their pulse, blood pressure, breathing rate and sweat levels.

But polygraphy has been under fire ever since it was invented in the 1920’s. Supporters say that experience in framing questions and the constant improvement in the monitoring equipment has made polygraphy highly reliable. Critics say such testing is flawed because it measures emotion rather than knowledge. They say the guilty can train themselves to respond in ways that deceive their questioners while many easily flustered people have been wrongly branded as guilty.

“Thought Police Peek Into Brains”

Julia Scheeres reports for Wired News. Excerpt:

U.S. investigators are facing the daunting task of sorting through more than 700 suspects in connection with the Sept. 11 terrorist attacks. A neuroscientist from Iowa says he’s got the perfect tool to help them do it.

Lawrence Farwell says he has devised a test that will ascertain whether the suspects have criminal knowledge of the terrorist attack by measuring their brainwaves. He calls it “brain fingerprinting.”

It a nutshell, it works like this: A subject’s head is strapped with electrodes that pick up electrical activity. He sits in front of a computer monitor as words and images flash on the screen. When he recognizes the visual stimuli, a waveform called the P300 reacts and the signal is fed into a computer, where it is analyzed using a proprietary algorithm.

In police cases, a suspect is shown data that only a person with intimate knowledge of the crime would know, such as details about the crime scene or the weapon used. If the suspect’s P300 waves reacted to the data, it would be an indication of guilt, Farwell said.

For the terrorist investigation, suspects could be shown pictures and terminology known only to members of a terrorist group, such as the word “al-Qaida” written in Arabic, or the instrument panel of a 757, he said.

“There is no question from a scientific perspective that this is an extremely useful tool in the war against terrorism,” said Farwell, who says he has tested more than 170 people and has a 100 percent accuracy rate in determining a recognition response. “It’s extremely important to the national interest to implement this as soon as possible.”