If you were to walk up to any stranger and ask if they could detect a lie, there is a high probability that you would receive an affirmative answer. Research however indicates that our ability to detect a lie is no better than chance. Still, we persist and in a sense the criminal justice system reflects just how tenaciously we hold to this mythological belief.
The most well known means for detecting deception via the measurement of physiological responses involves William Moulton Marston’s early 20th century polygraph machine. The original polygraph measured systolic blood pressure whereas later versions included galvanic skin response (skin conductance) measurements. By the early 1950s, the Department of Defense Polygraph Institute (DoDPI), funded several projects in search of the ultimate truth serum while continuing to explore various versions of psychophysiological detection of deception (PDD) tools and procedures. Among these were voice stress analysis[1] and periorbital thermography. It was also during this period that sodium amytal and sodium pentothal interviews were studied extensively. The results were inconclusive.
By the 1960s, voice stress analysis began making headway into various law enforcement agencies around the nation, and in 1978, Paul Ekman, PhD developed a system that he coined, facial action coding system ( FACS). As we near the one century mark of the polygraph, the scientific community remains divided with regard to PDD. Proponents claim deception detection instruments meet scientific rigor whereas skeptics claim they remain in the realm of art and are therefore not valid and reliable. Functional magnetic resonance imaging (fMRI) and functional near-infrared imaging (fNIR) are the latest PDDs to be added to the mix. At this time PDDs are inadmissible in a court of law. This is largely due to the 1923 precedent setting case of Frye v United States.
In 1923, James Alphonzo Frye, who had been convicted of murder in the second degree, appealed his conviction on the basis of a “single assignment of error.” Frye’s attorney had attempted and was denied the testimony of an expert witness regarding the results of a “systolic blood pressure deception test” that indicated Frye was telling the truth when he stated he did not commit the murder. In a December 3, 1923 precedent setting opinion, Chief Justice, Van Orsdel, upheld the ruling, stating the systolic blood pressure deception test had not yet gained wide acceptance by the scientific community.[2]
The result came to be known as the Frye standard (also known as the Frye test), which would be used to determine expert testimony admissibility for the next 50 years. By 1975 expert testimony admissibility criteria, which focused upon scientific method, was codified in Rule 702 of the Federal Rules of Evidence (FRE).[3] In 1993, during the appeal of another precedent setting case, Daubert v Merrell Dow Pharmaceuticals, Justice Blackmun noted that the court should act as a “gatekeeper” to ensure expert testimony met FRE 702 scientific standards, specifically in the areas of testability, peer review, error rates, and general acceptance.[4]
Expert testimony admissibility was further clarified in the 1997 General Electric Co. v. Joiner opinion, wherein Chief Justice Rehnquist ruled that expert testimony should not be excluded solely upon the basis that an expert might only be able testify to a limited degree of certainty–which some argue does not rise above guesswork and/or biased speculation.[5] In the 1999 case of Kumho Tire Company v Carmichael, Justice Breyer recognized Dennis Carlson Jr’s ipse dixit expertise, noting that FRE 702 does not distinguish between scientific knowledge and special knowledge, thus further clarifying expert testimony admissibility criteria.[6] Although expert testimony admissibility criteria have changed during the past century, several states continue to rely upon the Frye test. Therefore, for the fMRI to be accepted as a viable deception detection instrument in a court of law, it must, at the very least, meet the Frye standard of general acceptance within the scientific community. In this context, the fMRI looks promising. What is it that makes the fMRI look promising? Source measurements.
For example, the polygraph measures changes in the sympathetic nervous system–a system, some claim, that can be easily manipulated via relaxing and/or tensing, such that end results are inconclusive or generate a false negative or positive. The fMRI however measures changes in the central nervous system, which researchers argue is very hard, if not impossible, to control. Or is it? The brain activity measurements come in the form of changing blood flow, which in turn helps testers to determine which areas of the brain are activated at any given point in time. Researchers hypothesize lying involves emotion, conflict, and cognitive control, thus resulting in increased neurological excitation. To test this hypothesis, several researchers designed and implemented experiments involving the well-known deception paradigm coined the Guilty Knowledge Test (GKT).[7]
In 2004, Kozel et al attempted to replicate a previous blood oxygen level-dependent (BOLD) fMRI lie detection study to determine if correlations exist between lying and activation of specific brain centers. Ten (10) subjects were tested utilizing the GKT model. This model relies upon forced-choice deception tasks and hypothesizes statistically significant neurological differences will be present when the subject is lying. The results of Kozel et als research indicated increased activity in the right frontal areas, right anterior cingulate gyrus, right precentral gyrus, left temporal area and left cerebellum.[8] In 2005, Davatzikos et al used the BOLD fMRI to measure the brain activity of twenty-two (22) subjects participating in GKT activities. Their results indicated increased brain activity in the right prefrontal cortex and temporoparietal cortex.[9] In 2006, Mohamed et al tested eleven (11) subjects in a GKT mock shooting scenario.[10]
In all three research experiments, subjects targeted to lie were told they would receive a financial reward (usually $25) if they followed the rules and successfully escaped detection. To control for confounds, all subjects were tested for handedness, and Kozel et al, whose research included both male and female subject, also tested for and eliminated subjects for DSM-IV-TR Axis I disorders, drug use, and, in the case of the females, pregnancy. Each of the aforementioned researchers hypothesized that deception activates specific areas of the brain whereas truth telling activates alternative areas. To test their hypothesis, Kozel and Davatzikos asked subjects to lie about seeing a playing card, whereas Mohamed et al instructed their subjects to fire a starter pistol with blanks. Of Mohamed et als eleven participants, six were instructed to lie when asked if they had fired the pistol. All experimenters also used the polygraph to measure deception.
By 2004, fMRIs were being touted as the lie detector of the future. Since that time, several small companies have begun marketing fMRI lie detection services.[11] With increased interest however came increased concerns. Among these involve the age-old questions of validity and reliability,[12] and perhaps the equally important though less well defined question of bioethics. The prior, validity and reliability, is perplexing when considering the experiments conducted thus far. For example, in attempting to control for confounds, researchers have limited their test sample to college-aged, healthy, drug-free (i.e., caffeine, tobacco, alcohol, prescribed, and over the counter medication), mostly male, participants. Additionally, no time is allowed between the target incident (that which the subject is to lie about) and the lie. Thus results could be confounded by lack of lie preparation as would normally occur in a forensic setting.
The latter is especially troublesome when it comes to fNIR and the ability to “surreptitiously scan the brain.” Furthermore, the issue of bioethics was raised and addressed in the mid-1980s when employers began utilizing polygraphs for pre-screening potential employees. The use of PPDs by employers was quickly halted with the enactment of the 1988 federal Employee Polygraph Protection Act. And finally, Hughes et al (2005) proposes that results indicating deception as an executive function support the hypothesis that deception is an evolutionary necessity. Their hypothesis is further supported with regard to a number of acute psychological disorders in that executive control of deception appears to be “broken.“[13] In summary, although emerging PDD technologies look promising for forensic settings, it is important to proceed with extreme caution and aforethought due to potentially negative unintended societal consequences.