Wednesday, April 1, 2009 | A juvenile sex-abuse case in San Diego has brought the issue of whether MRI machines should be admissible as lie detectors in U.S. courts farther than ever before — again pitting entrepreneurs who say the technology is ready for the courtroom against scientists and lawyers who say allowing it would set dangerous legal precedent.
Last month, a report based on the results of a brain scan done by San Diego-based No Lie MRI was offered as evidence by the defense in a child protection hearing in San Diego County Juvenile Court. The report from the scan, called functional magnetic resonance imaging, was submitted to show that a man was telling the truth when he denied sexually abusing a child.
It was the first time that experts on the subject are aware of that results of such a scan were offered as evidence of a person’s guilt or innocence in a U.S. court. And though the report was withdrawn last week by defense lawyers, the case will go down as an important juncture on the road to determining whether machines can reliably tell if someone is lying.
“The mere fact that someone felt the technology was far enough progressed to attempt to introduce it into evidence is significant to both the legal field and neuroscience,” said Senior Deputy San Diego County Counsel Gary C. Seiser, the county’s lawyer in the juvenile hearing. “Just a few years ago nobody believed it was ready.”
A functional MRI, or fMRI, is essentially a scan from a standard MRI machine that has software tuned to track oxygenated blood flow throughout the brain. The theory behind using fMRI scans as lie detectors is that lying requires the brain to do more work than telling the truth. So, when a person tells a lie oxygenated blood flows to specific parts of the brain, and this blood flow can be shown via the fMRI data.
Few will dispute the potential of fMRI to someday have a great impact on jurisprudence. But for years the vast majority of scientists and lawyers have agreed that researchers still have a long way to go in proving fMRI’s reliability as a lie detector before it should be anywhere near a courtroom. Even one of the inventors of the technology is dubious of its use in a real-world setting.
Yet the prevailing opinion hasn’t stopped Joel Huizenga, CEO of No Lie MRI, from trying to convince a judge anywhere that the technology is reliable enough to be used in a legal case. “There have been 20 or 30 scientific journal articles about this — all have concluded that it works,” Huizenga said. “Cases will come to court; they just have to come to the right venue.”
That is what worries Hank Greely and Emily Murphy at the Center for Law & the Biosciences located at Stanford University. “I don’t think it is anywhere near sufficiently proven to be used today,” said Greely, a law professor who heads up the center. “And premature use not only risks harming people’s lives, but potentially harming neuroscience by overpromising and under-delivering,” on the technology.
Greely and Murphy acknowledge that fMRI has shown positive results as lie a detector in laboratory settings, but they emphasize that the data have come from highly controlled experiments, a far cry from the reality of a courtroom. And they sharply criticize Huizenga for not allowing any research he as done to be scrutinized by the scientific community.
“For him to say ‘it works’ is evading the real issue,” said Murphy, who is a neuroscientist and fellow at the center.
Huizenga said Greely’s and Murphy’s opposition to his work is politically motivated. He said they were “part of a political action committee funded by the MacArthur Foundation.”
Greely receives grant money from the MacArthur Foundation. But the foundation does not have a political action committee. As a 501(c)(3) charitable organization, it is forbidden by law from doing so. Greely called Huizenga’s charge “ridiculous.”
Very few U.S. courts allow test results from lie-detecting machines to be admitted as evidence. Only New Mexico courts allow results from a polygraph, the most widely used method of lie detecting, to be admitted.
Functional MRI technology was invented in the early 1990s for brain mapping and as a diagnostic tool for Alzheimer’s disease and other neurological disorders. Earlier this decade psychiatrists Daniel Langleben, at the University of Pennsylvania, and F. Andrew Kozel, now at the University of Texas Southwestern Medical Center, began using the technology for lie detection.
Both have patents pending. In the meantime, Langleben has licensed his technology to No Lie MRI, and Kozel licensed his to Massachusetts-based Cephos Corp., which also markets fMRI-based lie detection. Neither scientist, however, is currently a paid representative of the company he licensed his work to.
The scans have correctly detected lies in many experiments done by Kozel, Langleben and others. In a typical experiment, a subject will be given a playing card and told to lie about the value of the card. In others, subjects are told to steal items and then lie about the theft. The machines’ accuracy rates in the experiments are regularly over 90 percent, Huizenga said.
“There is no one in conflict with the data,” Huizenga said.
Greely and Murphy, among others, said their issue is not with the results of the experiments, it is with the experiments themselves. For one thing, subjects in a laboratory are not under anywhere near the stress that someone involved in a court case would likely be under. And it is not at all clear how such stress could affect blood flow in the brain.
“What happens when you are charged with sexual molestation or murder?” Seiser said. “What happens when it is high stakes?”
Also, none of the experiments adequately address the actions someone might take to beat the test, Murphy said. It has long been known, for example, that someone can beat a polygraph test by doing things like stepping on a tack or taking sedatives. There is no data on how such actions, or others, would affect the results from an fMRI.
“People can deceive themselves or others in very different ways,” Murphy said. “For example, there is no evidence that shows whether these things work if it is a lie that someone has convinced themselves is true.”
Even Kozel, one of the inventors of the technology, agrees that it is not yet ready for the courtroom. “Although we have established that it is a reliable technique at individual level in the laboratory setting,” he said. “How valid it is under different circumstances is unknown.”
Under law, the case file of the juvenile protection hearing in San Diego is sealed, and all parties involved in the case are forbidden from speaking about it. However, Murphy recently posted an item on the Center for Law & the Biosciences blog that offered some details. At issue, according to the blog post, is whether a minor suffered sexual abuse at the hands of a custodial parent.
The blog post reports that the parent contracted with No Lie MRI to do a lie detection test. The report produced by the company asserted that the test showed that the parent told the truth when answering “no” to direct questions about whether he had sexual intercourse with the child.
The report was offered into evidence by the attorneys McGlinn & McGlinn, but then withdrawn. Ryan McGlinn said only the following: “As far as the juvenile court preceding we were trying to admit it into evidence. We believed it was reliable. Now, at this point, we have decided to not try to get it admitted.”
Huizenga said the decision had nothing to do with the validity of the report No Lie MRI produced. “There was nothing wrong with our evidence,” he said. “This is something the lawyer decided to do for his own reasons.”
Seiser, like McGlinn, would say very little concerning the case. But he did say that an fMRI scan would not come close meeting the evidentiary requirements of California courts.
A new scientific technique must meet the following requirements before it can be admitted into evidence: It has to be generally accepted in the relevant scientific community; that general acceptance has to be presented by a qualified expert; and it must be shown that proper procedures were following in the particular case.
Seiser said it would be easy to prove the general distrust of fMRI in the scientific community. To that end, he had arranged for Marcus Raichle, the world-renowned neurologist from Washington University in St. Louis whose work formed the physiological basis of fMRI, to be an expert witness for the county.
And, Seiser said, fMRI would also have problems clearing the second hurdle because the only qualified experts of the technology are also its inventors, who have an inherent conflict of interest. Others said the third requirement would be a problem because Huizenga has refused to share his methods with the scientific community.
“[fMRI] is promising, there is no question about that,” Seiser said. “But there is still a lot more testing that needs to be done. And it is showing us that law and science really have to be communicating.”
Clarification: The original version of this story incorrectly characterized how the MacArthur Foundation funds the work of Hank Greely, the Stanford law professor who heads up the Center for Law & the Biosciences. Greely receives grant money from the MacArthur Foundation for his research into brain imaging being used in court cases. However, the center does not directly receive money from the foundation.