While proponents of lie detectors swear by their accuracy up to 90 percent, critics across the country have successfully argued against allowing results from such tests into evidence during court trials.
Defense attorneys most recently failed to pass the Frye-Reed test — Maryland’s standard for the admissibility of evidence — during a motions hearing Tuesday in Montgomery County Circuit Court for a first-degree murder trial. The defendant, 29-year-old former U.S. Army Ranger Gary Smith, passed multiple lie detector tests, including a Functional Magnetic Resonance Imaging scan, when asked about his role in the 2006 death of his roommate and fellow army combat veteran Michael McQueen.
Under the Frye-Reed standard, scientific evidence like DNA or lie detector tests must be accepted by a general consensus of the relevant scientific community before it is allowed as evidence during a trial, a standard Judge Eric M. Johnson cited after a careful review of academic journals and studies presented by prosecutors and Smith’s defense attorneys Tuesday.
“There’s no quantitative analysis of this procedure available yet,” Johnson said before making his ruling.
The fMRI, typically used to measure brain activity for medical research by detecting changes in blood flow to different regions of the brain, could also allow testers to identify which areas of the brain are active when a person lies, said Joel Huizenga, the founder and CEO of No Lie MRI, the California-based company that administered Smith’s test.
“There is always room to do more research in anything, the brain’s a complex place. There have been 25 original peer reviewed scientific journal articles, all of them say that the technology works, none of them say that the technology doesn’t work,” Huizenga said. “… that’s 100 percent agreement.”
But just because scientists generally agree that the use of magnetic resonance scans to detect falsehoods might someday work, that does not mean the technology is anywhere close to meeting the high standards required to be admitted in U.S. courts, said Dr. Hank Greely, a Stanford University law professor specializing in the legal implications of new technology. Greely also provided written testimony for Montgomery County prosecutors in Smith’s case.
“I don’t think there’s a serious neuroscientist who thinks that they are ready to be used in court and there’s a split in neuroscientists who think they’re any good at all,” he said. “... There’s always the possibility that it will someday eventually become reliable enough to be used in court, but there’s an enormous amount of scientific research that still needs to be done, first.”
Proponents of fMRIs often brush over significant holes in their research, including the complete lack of reliable, study-based evidence that certain portions of the human brain are concretely linked to lying, Greely said. Also, because the scans measure brain activity, test subjects could, for instance, develop undetectable countermeasures — like counting back from 100 or trying to remember past World Series winners — that might throw off test results, Greely said.
Steven Laken, president of the Boston-based Cephos company, provides his clients with a variety of forensic services, including fMRI lie detection and DNA analysis. He believes fMRI results are held to an unfair standard. Laken described, for instance, the hurdles DNA evidence once faced before it became widely accepted by judges.
“But the standard is set even higher for these lie detectors because of this idea that the judge or the jury are basically the final determiners of whether someone is lying on the stand,” he said. “The courts are unfairly putting a higher bar on that than they are on other scientific evidence like DNA.”
Like DNA evidence, which is generally accepted by courts today, members of a jury should be allowed to hear fMRI test results to make their own determination on their reliability at trial, Laken said. Judges are only doing a disservice to the justice system by denying all the relevant evidence to the jury, he said.
Stephen E. Fienberg, a professor of statistics at Carnegie Mellon University, also took part in a congressionally-mandated study on the reliability of polygraph lie detectors for the U.S. Department of Energy in 2002. Polygraphs, which seek to detect lies through indicators such as a person’s blood pressure, pulse and respiration, are also largely disallowed as evidence in a criminal court trial, Fienberg said. He has also studied fMRI detector reliability rates.
“There is no reason to believe that the fMRI would be more accurate than the polygraph, and lots of reasons to believe that it would be less accurate because of the subtleties in the studies,” he said.
For the time being, at least, judges around the country have flatly rejected test results from fMRI lie detectors as evidence in court based on the lack of a general consensus among scientific experts.
“DNA is awfully good and we can’t insist that all other evidence reach that level,” Greely said. “… but the crucial thing about DNA is not its accuracy, it’s the standards that were worked out over several years as to how it should be tested; we haven’t even begun to do that for fMRIs.”