The traditional lie detector, the polygraph, has existed for many years. It relies on physiological reactions--increased heart rate, respiration, blood pressure and sweating--to indicate that a person being questioned is fearful of getting caught and is therefore lying. Although this machine has been used in criminal investigations, critics insist it can easily be defeated. Some people are very good at controlling their physiological responses. Others secretly invoke alternative sensations at the same time, which can confuse the polygraph. And simple fear of a false reading can cause reactions in an honest person's body chemistry that register as a false response. New brain-imaging techniques have reawakened interest in lie detection.

Near-infrared brain scan is a test of blood flow devised by Britton Chance, a biophysicist at the University of Pennsylvania. A headband containing near-infrared light emitters and detectors is placed on the subject's head, which reputedly senses changes in the prefrontal cortex, the site of decision making that is also stimulated by deception. According to Chance, the sensors can detect changes that occur when a person makes a decision to lie--before the lie is actually articulated. Subjects are given a series of questions, some to be answered truthfully, others not, to chart the changes that occur. Although the device is still in development, Chance claims it will soon be capable of detecting covert activity in the prefrontal lobe.

Thermal imaging uses a heat-sensitive camera to detect increased blood flow around the eyes. Some scientists claim that when people lie, their eyes give off more heat than when they are telling the truth. This technique is in the early phase of development, and how accurate it may be is still an open question.

Functional magnetic resonance imaging (fMRI) is also being investigated. Daniel D. Langleben of the University of Pennsylvania recently studied 18 volunteers who were given certain playing cards. The subjects were then placed in an fMRI scanner. A computer presented them with images of specific cards and asked them if they had those cards in their possession. When the subjects lied, their anterior cingulate cortex and superior frontal gyrus lit up more than when they told the truth. The anterior cingulate cortex, which has connections to the limbic system and the prefrontal cortex, is involved with emotional processing, decision making and conflict resolution. It seems to be frequently activated when a lie is being told.

Yet the anterior cingulate cortex is also involved with decision making in general, which is a confounding consideration. A subject may activate this region from mere anxiety about the lie detection event. Yet it appears that telling the truth does not create a distinctive brain print, just diffuse activity. More research must be conducted to assure that the evaluations of fMRI patterns are highly specific to lying.

Refinements of fMRI will probably be made so that information flowing between various brain regions could be traced, giving more insight into what the test subject is feeling. Functional MRI could also be linked with trans-cranial magnetic stimulation to produce a powerful lie detection system. The magnetic apparatus could block out or enhance activity from select parts of the brain, in effect eliminating interference in the fMRI signal and improving the accuracy of detection in critical brain regions. Using both technologies together could also create greater sensitivity to lies that are camouflaged with confounding thoughts invoked by the subject.

Brain-fingerprinting results have already been admitted into evidence in one case--a reexamination of an Iowa murder conviction--even though the technique has not gained wide acceptance in the scientific community. In this approach, developed by Brain Fingerprinting Laboratories, a subject wears a helmet of electrodes, creating an electroencephalogram (EEG) that records changes in electrical potentials in the brain. The subject is presented with words, phrases or pictures while the EEG records her brain-wave activity. As with polygraphs, an investigator presents information that ostensibly only the offender would know. If the suspect knows the information but lies, a specific brain wave known as P300 is elicited. The P300 pattern is activated when the brain recognizes information (or a familiar object) as significant or surprising. The goal is to determine if the subject has the information stored in her brain even though she denies knowing it.

Lawrence A. Farwell, the inventor of brain fingerprinting, claims an accuracy of nearly 100 percent, but there are several problems. First, the presence of drugs and alcohol can adversely affect the reception and storage of information. Second, the investigator has to have detailed information that only the participant would know, requiring much investigation; FBI and police reports are not so detailed. Brain fingerprinting therefore will most likely prove significantly useful in situations in which unique factual information is available to investigators. Furthermore, with advances in behavioral genetics, it might be possible in the long term to correlate gene profiles with brain-fingerprinting waves. That could enhance the statistical validity of the test results by factoring out confounding conditions such as anxiety or fear and factoring in biological conditions such as psychopathy that are known to be highly associated with antisocial behavior.