One researcher has built a headband outfitted with lights and detectors able to "see" blood-flow changes in the brain. Another uses magnetic resonance imaging to snap several split-second pictures.
Britton Chance, a biophysicist at the University of Pennsylvania, leads the headband project, which uses near-infrared light to peek at the brain's prefrontal cortex, the place where people make decisions — and where lies are born.
Research subjects wearing the headband are told to answer some questions truthfully and others deceptively.
At the moment a subject makes the decision to lie, before even uttering it, there's a milliseconds-long burst of blood flow. Those bursts are read by the sensors and show up as spikes on a laptop computer.
One day, Chance said, the headband might not be needed at all. Perhaps one would need only point a sensing device at people — making it possible to test someone's truthfulness without their knowledge.
"We're interested in covert detection of prefrontal activity, where the subject may not be told the experience is occurring. That's in the future but it is possible," he said. "Obviously, there are ethical problems."
"There's only one thing worse than a lie detector that doesn't work, and that's a lie detector that does work," said physicist Robert Park, a longtime polygraph critic. "It's the last invasion of privacy that you can imagine, and it frightens me that we seem to be almost able to do it."
Traditional lie detectors, known as polygraphs, measure heart and respiratory rates as a person answers questions.
Critics claim polygraphs are easy to beat — they say something as simple as stepping on a tack placed in a shoe can skew results in the test-takers' favor — and largely unreliable, as evidenced by people like former CIA agent Aldrich Ames, who passed polygraphs, concealing his work as a Russian spy.
Though federal agencies use polygraph tests to screen workers and job applicants, courts do not allow the tests to be admitted as evidence.
Researchers believe the technologies they're working on could change that — though it could take several decades to get it right.
"I doubt that anything in life will ever be 100 percent reliable, including lie detection. But will we have a technique that's good enough to be taken as one source of evidence? Probably," said Stephen Kosslyn, a Harvard University psychology professor who is studying the brain scans of liars.
As Chance develops his headband, another Penn researcher, psychologist Daniel Langleben, is putting volunteers inside a type of magnetic resonance imaging machine and telling them to lie as it photographs their brains.
Langleben's MRI detects which part of the brain is active in response to specific stimuli. Volunteers were told not to divulge a playing card they were given. They were then placed within an MRI scanner and "interrogated" by a computer. When volunteers lied, Langleben said, part of their brains lit up.
Chance and Langleben contend that people can't change what happens in their brains during a lie, so a machine accurately measuring those changes would be next to impossible to beat. Polygraphs, on the other hand, essentially measure the fear of getting caught lying, symptoms that can be beaten.
"It strikes me as odd that people seem rarely to see the positive side of a reliable lie detector," Kosslyn said. "If you're innocent, wouldn't it be nice to have a way to support your claims?"
Researchers say more accurate lie detectors could help courts and police.
Doctors could also determine whether patients are being less than truthful in describing their symptoms. Corporations could check whether their employees — or perhaps even their chief executives and accountants — are truthful.
Other scientists are looking at "thermal imaging" (training a heat-sensitive camera on people's faces that would register increased blood flow around the eyes) and "automated face analysis" (a computer that analyzes the tiniest expressions in the face) as potential lie detectors.
Lawrence Farwell, an Iowa-based neuroscientist who runs Brain Wave Science Inc., has developed what he calls "brain fingerprinting." It focuses on a specific electrical brain wave, called a P300, which activates when a person sees a familiar object.
A convicted murderer petitioning for a new trial has already tried to use brain fingerprinting as evidence in an Iowa court. The test showed that the defendant, Terry Harrington, had no memory of the crime scene, but the judge refused to accept it as evidence.
Barry Steinhardt, director of the American Civil Liberties Union's technology and liberty program, warns that none of the new technology has been proven to work like the scientists claim.
But if it does, Steinhardt said, "then it would become another weapon in the arsenal of those who want to put us into a surveillance society where every action, every deed and one's very thoughts can be monitored, categorized and correlated."