After nearly 12 hours of testimony yesterday by scientists, a hearing on whether lie detection technology based
on fMRI scans of brain activity should be admitted in court continues today. Many scientists are skeptical of this technology, but several recent cases
suggest that lawyers are increasingly interested in using it to sway juries. Based on yesterday's testimony in a federal
court in Tennessee, it's hard to say whether one side or the other is gaining the upper hand, says University of Pennsylvania cognitive neuroscientist
Martha Farah, who flew to Memphis earlier this week simply to be a spectator.
"To me, it's very exciting," Farah says. "They're trying to decide if this [technology] should have a role in a criminal trial." She says she was
surprised by the first day's events, which she recapped with ScienceInsider last night. Going in to the hearing, she says, she assumed that the
fMRI lie-detection method "didn't have much of a chance" to ruled admissible by the judge, Tu Pham. But now she's not so sure. The scientists behind
the method have "some impressive sounding evidence, and understanding the problems with it requires fairly involved scientific and statistical
reasoning," she wrote.
The first to testify in yesterday's hearing was Steven Laken, a cell biologist and founder and CEO of Cephos,
a Massachusetts company that offers fMRI lie detection services.
Late last year, Cephos was retained by the defendant in the Tennessee case, Lorne Semrau, a psychologist who is fighting charges that he defrauded
Medicare and other health insurers with wrongful claims. Semrau's attorney hopes to introduce fMRI scans performed by Cephos as evidence that he is
telling the truth when he says he had no intent to commit fraud.
According to Farah, Laken "spoke in very strong and confident terms" about research finding that certain regions of the brain become more active when a
person lies. Laken also described the procedure undergone by Semrau, which, to make a very long story short, involved having his brain scanned as he
answered questions about details of his case, and comparing those scans with others taken as Semrau answered a different set of questions, unrelated to
the case, after being instructed to intentionally lie on certain questions and respond truthfully to others. (The methods are similar, but not
identical, to those used in two studies published by Laken and other researchers affiliated with Cephos: a 2005 paper in Biological Psychiatry, and a 2009 paper in the Journal of Forensic Sciences.)
On cross-examination, it came out that Cephos had originally planned to scan Semrau twice, with each scan addressing different aspects of his alleged
misdoings. The first scan indicated he was telling the truth, but the second indicated he was lying. Laken noted that the method has a high rate of
false positives (33% according to the 2009 paper) and argued that the second scan may have been compromised because Semrau appeared to be fatigued.
When Cephos repeated that scan 2 weeks later, it indicated that he was telling the truth. Prosecuting attorneys questioned why fatigue would influence
the results and argued that this do-over was arbitrary and indicated a lack of standardization in the procedure, one of several guidelines used to
establish the admissibility of scientific evidence.
After lunch, the court heard from Marcus Raichle, a neuroimaging expert at Washington University in St. Louis. Farah says Raichle raised questions about
the strength of evidence that increased activity in the brain regions examined in the Cephos scans are specifically related to deception. The same
regions become active during a variety of mental tasks, Raichle said. He also noted that Semrau was in his 60s when the scans were taken, considerably
older than the 18- to 50-year-old subjects who participated in the published studies.
Farah sounds like she would have liked to chime in at this point about some things that weren't getting enough attention. "No one asked me, but the thing we have
not a drop of data on is [the situation] where people have their liberty at stake and have been living with a lie for a long time," she says. She notes
that the only published studies on fMRI lie detection involve people telling trivial lies with no threat of consequences. No peer-reviewed studies
exist on real world situations like the case before the Tennessee court. Moreover, subjects in the published studies typically had their brains scanned
within a few days of lying about a fake crime, whereas Semrau's alleged crimes began nearly 10 years before he was scanned.
Next on the stand was Peter Imrey, a biostatistician at the Cleveland Clinic Foundation in Ohio. Farah says he began with a lengthy exposition on
statistics and several examples from epidemiology that didn't seem directly relevant to the Semrau case. "I worry that some of the problems raised by
Imrey may seem like academic hair splitting," she said as the court adjourned for a break late in the day. But later in the evening she texted to say
that Imrey was landing some more direct hits on Cephos, raising serious questions about their methods.
At the end of the day, it wasn't clear who was winning, Farah said. But she says that Judge Pham seems determined to hear everyone out. "I think that
we are getting a fairly complete picture of what's known and not known about the validity of this method."
For more on that, and updates on the hearing, check back with ScienceInsider.
* This article has been amended to more closely reflect Farah's description of the testimony regarding differences between the published studies on fMRI lie detection and the procedure used in the Semrau case.