Brain scans detect lies more accurately than the polygraph
It’s a common scene from an episode of aTV crime drama: A suspect is strapped to apolygraph machine. He sweats. He squirms.His eyes dart from one direction to the other.The needle on the polygraph machineswings wildly, indicating thatthe suspect is lying.
In fact, increases in one’s blood pressureor pulse, or the presence of sweaty palms, areno longer seen as telltale signs of lying: Sincethe 1980s, polygraphs have been deemedinadmissible as legal evidence in a majorityof jurisdictions in the U.S. and for most pre-employmentscreenings.
However, polygraphs remain the only biologicalmethod of lie detection in use in theU.S. in certain instances, such as governmentbackground checks and security clearances.
In recent years, researchers have successfullyused functional magnetic resonanceimagining, or fMRI, to map brain activationpatterns associated with intentional deception.Neuroimaging studies have also shownthat fMRI is fairly accurate in detecting deceptionin individuals. However, expertsagree that use of fMRI in real-life police orforensic work remains impractical withoutextensive testing and validation.
Daniel Langleben, a professor of psychiatry;his colleagues at the Perelman Schoolof Medicine; and Ronald Barndollar, a FBIsupervisory agent and polygraph examiner,set out to fill in one of those gaps by comparingthe polygraph and fMRI in lie detection.The study, published in the Journal of ClinicalPsychiatry, found that experts reading fMRIdata were 24 percent more likely to detect alie from a study participant than professionalexaminers reviewing polygraph recordings.
The study was the first to compare polygraphtesting and fMRI under nearly identicalconditions and in the same participants—which was no small task, says Langleben,since the tests are administered and evaluateddifferently.
He explains that with a polygraph, anexaminer asks questions of the subject untilthere is enough data to make a determination.The recordings are usually read bythe same person who administers the test.Although automated software packages forscoring polygraph data are included withmost digital systems, examiners rarely relyon them to make determinations.
Comparatively, fMRI tasks are programmedin advance and do not change betweenparticipants; the scanner operator interactswith the testee via intercom and onlyin an emergency. Experts do not compareresults. It is, as Langleben says, “an almostideal experimental environment.”
The challenge, he says, was in setting upan experiment where two dissimilar technologiesare used under conditions similarenough to allow a head-to-head comparison.
“We decided that since it’s new technologythat needs to prove itself against theold—or to be polite, ‘state of the art’ technology—the old technology should be the startingpoint. In our case, this meant that MRIexaminers were asked to act as if they werepolygraph examiners,” Langleben says. “Thepoint of the study was not to achieve the bestperformance either technology is capable of[but] to compare their performance underidentical techniques.”
The researchers gave 28 participantsthe “Concealed Information Test,” whichis designed to see if people have specificknowledge that in a real-life situation couldbe incriminating, through a series of questions,some of which have known answers. Aresearcher asked participants towrite down a number betweenthree and eight, keep it hidden,and not admit to having writtena number. Next, participants werehooked up to a polygraph or werescanned by a fMRI machine andasked whether they had writtena certain number. People wereinstructed to answer “no” to allquestions, making one of theiranswers a lie. At the end of thesession, participants handed theirhidden notes to the designatedteam member who placed themin a sealed envelope in a lockedbox. The results were analyzed by three polygraphand three neuroimaging experts independentlyof one another and then comparedto see which technology was more accurate indetecting the lie of each participant.
Overall, fMRI experts were 24 percentmore likely to detect the lie. Langleben andhis colleagues discovered another importantpoint, too: When everyone agreed on thenumber that the participant was lying about,fMRI and polygraph experts turned out tobe 100 percent correct. This happened in 17cases and suggests the two tests may complementeach other.
Langleben says the study was not designedto prove this unexpected observation,but it points to a potentially useful approachin countries that adhere to the “guilty untilproven otherwise” principle. If the suspectcan pass sequential MRI and polygraph tests,“the chances he/she is lying are very low,” hesays. “This could be brought up as evidenceof ‘reasonable doubt’ leading to acquittal orreduced punishment.”
Langleben says people should be open tothe possibility and power of a brain scan.
“Imaging is the most powerful technologythat we have so far for looking at howthe living brain works,” he says. “This workis important not only as a lie detection project,but especially because it opens doors andwindows to all kinds of related fundamentalphenomena, such as delusions, self-deception,and insight.”
