Lie Detection: The Science and Development of the Polygraph

You are seated in a room, alone with an examiner. The examiner hooks wires and straps to your body. You are nervous and your heart begins to pound. He starts asking you questions. You try to stay calm, since you are innocent and have nothing to conceal. However, agonizingly, you feel your hands sweat, as you worry that the metal box and computer next to you will devastate your future.

This is one possible reaction that a test subject might have to the type of polygraph tests that are conducted throughout the nation every year. Although polygraph tests are inadmissible as evidence in most states, polygraphs are administered to many witnesses and suspects to streamline police investigations and project new directions of examinations. Over 20 federal agencies use polygraphs to screen job applicants and current employees [1]. Occasionally, polygraphs make their way to the news headlines and provoke heated debate. For example, when California Congressman Gary Condit was involved in the scandal of missing intern Chandra Levy, there was much debate about whether he should submit to a polygraph test. Under media pressure, he took the polygraph test and passed. A few months later, police linked Levy’s murder to a criminal in the Washington, D.C., area. In this case, the polygraph provided accurate results, but serious doubts remain about validity of the test in all cases. The ability to critically evaluate the accuracy of the polygraph is predicated on an understanding of the fundamental theories behind the device’s functionality.

For centuries, humans have sought a reliable means to detect lies. In ancient Hindu and Chinese civilizations, authorities “detected” lies by asking the suspect to chew a grain of rice and spit it out. In China, a dry grain of rice would be indicative of the dry mouth of a liar [2]. In India, rice stuck to the mouth would be the sign of guilt [1]. Although these methods were primitive and non-scientific, they nonetheless highlight the fundamental assumption humans make in lie detection: the psychological state of lying can be detected by physiological signs.

Of the three physiological signs recorded in modern polygraphs, cardiovascular activity was the first sign to be associated with lying. This association was first documented by an Italian criminologist named Lombroso in the late nineteenth century [3]. Lombroso noted an increase in a subject’s blood pressure and heart rate when the subject was giving untruthful answers [3].

In the 1910s, some studies suggested that the ratio of inhaling and exhaling time during breathing changes when a person lies [3]. Using this theory along with Lombroso’s findings, John Larson, a medical student working for the Berkeley Police Department, invented the first polygraph in 1921 [2]. This first polygraph simultaneously traced a subject’s blood pressure and respiration [3]. Under Larson’s assumptions, irregularities in blood pressure and breathing patterns would indicate lies.

The third physiological channel used in modern polygraphs “skin resistance” was added later in the 1920s by Leonarde Keeler [2]. Keeler assumed that a lying person would sweat more than a truthful person, which would decrease skin resistance due to a higher concentration of negatively charged chloride ions on the surface of the skin. By adding an extra metric for detecting lies, scientists had hoped to decrease the probability of errors in the polygraph test. Since this addition, there have been no new underlying assumptions introduced to improve the polygraph.

Technological breakthroughs in past decades have improved the retrieval and analysis of data obtained from a subject during a polygraph. However, the fundamental assumptions behind the polygraph and the parameters measured have not changed in over eighty years. For example, in the 1970s and 80s, the analog polygraph was used. This machine converted physiological responses into electrical signals which induced needles to move and record the physiological data on a scrolling roll of paper [4]. Although the analog polygraph incorporated advances in electrical engineering and helped examiners to conveniently record data, it still worked on the same principles as the first polygraphs.

With advances in computer programming in the 1990s, most modern polygraphs are administered through computers. The development of sophisticated software has allowed signals to be displayed on computer monitors, which allows examiners to identify irregularities more efficiently. However, lie detection using the polygraph still relies on the same three physiological vectors: cardiovascular, respiratory, and perspiratory activities. Modern polygraphs also use the same underlying assumption derived hundreds of years ago: lying is associated with certain changes in physiological conditions (see flash demo).

Despite​ its name, the lie detector, also known as the polygraph, is not capable of detecting lies. The device can only display certain physiological signs that are assumed to reflect the psychological state of a liar. For example, the modern day polygraph displays measurements of heart rate, blood pressure, respiration and perspiration in the fingers (see Fig. 1). When a person lies or is asked a sensitive question, his or her heart may begin to race, raising the body’s blood pressure. The test subject may also hold his or her breath, take in a deep breath, or begin to sweat. These physiological irregularities are detected by the polygraph and interpreted by the polygraph examiner. It is the province of the examiner to decide whether or not sudden changes in the data signify dishonesty. The assumptions that lying will produce physiological changes and that polygraph examiners will be able to interpret these signs accurately have raised questions about the accuracy of the polygraph. What if a test subject does not exhibit abnormal physiological signs while lying? Can human errors occur in the polygraph interpretation process? Indeed, polygraph testing is not an exact science, and there is certainly margin for error in the interpretation of the three metrics.

Blood pressure and heart rate data are collected by an arm-encircling cuff placed on the upper arm. The cuff is filled with air and connected to the polygraph machine through air-filled tubes. Changes in blood pressure modulates the air pressure in the cuff. These changes are recorded by the polygraph machine and displayed on a computer screen [4]. This data is simultaneously displayed with respiratory and perspiratory data on a computer monitor, and all are interpreted by the polygraph examiner.

The respiration pattern of the subject is detected by two pneumograph, devices which record thoracic movements, or volume change during respiration. One of the pneumograph tubes is strapped around the chest and the other is placed around the abdomen. Each pneumograph has an air-filled rubber tube connected to the machine. When the subject breathes in and out, the air pressure inside the tubing changes and is recorded by the polygraph machine.

The measurement of sweat, which is scientifically known as the measurement of galvanic skin resistance, is conducted by a two-piece galvanometer attached to two of the subject’s fingertips [4]. The galvanometer works by sending a small electric current into the skin from one of the fingerplates and records how much current was allowed to pass through on the other fingerplate. Dry skin is not a very good conductor of electricity. If a subject perspires, however, the water and salt from the sweat reduces the resistance of the skin. This decrease in resistance allows a larger amount of electric current to travel along the surface of the skin. Therefore, the amount of electric current recorded by the galvanometer reflects the amount of sweat that was produced in the subject’s fingertips.

There is normally a preliminary interview between the polygraph examiner and the examinee before the important questioning begins. During this pretest period, the examiner explains the polygraph testing procedures. The examiner also gets the opportunity to understand the subject’s behaviors and responses to information [4]. Understanding how the subject reacts to questions gives the examiner an idea of the subject’s normal physiological responses, so he or she can detect irregularities if the person lies during the actual test.

Crafting questions to ask in a polygraph test is often very complicated and requires thorough analysis of the case. First, it is of the utmost importance that questions are constructed with simple, unambiguous language in order to reach conclusive results. For example, an examiner wants to ask the subject if he has been in court before as a defendant, and might ask, “Have you ever been in court?” The subject can easily misinterpreted the question and say “Yes” even if he or she was just accompanying a friend to court [3]. Second, the examiner has to structure questions in such a way that he or she is able to decisively distinguish between truthfullness and falsehood. Furthermore, the examiner can construct questions that are sensitive to the real suspect but meaningless to rest of the examinees [5]. For example, the examiner might ask, “Is this the place where the body was hidden?” The real suspect might have a much greater emotional response than an outsider who has no knowledge of the crime. Therefore, the questions asked in a polygraph test are very important because they can directly trigger psychological conditions in the subject which, in turn, according to the underlying assumptions of the polygraph, produce physiological responses that are detected by the polygraph machine.

During the actual test, examiner and subject are in a room alone, with the subject connected to the polygraph machine. The examiner asks about 10 or 11 questions, of which only 3 or 4 are “relevant” questions. The other questions are known as control questions, such as “Is your name x?” “Do you live in city y?” etc. They yield non-lying physiological data and provide reference points for comparisons to the relevant questions [3].

After the test, the examiner analyzes the collected physiological data and looks for irregularities. The assumption behind this evaluation is that questions “relevant” to the liar trigger emotional responses and cause significant fluctuations in physiological conditions. For example, if the subject’s skin resistance, heart rate, and blood pressure all increased and his or her respiration halted when a relevant question was asked, examiners might decide the subject was being deceptive. However, in most cases, the physiological data recorded is not so well-defined. In more ambiguous situations, the subject”s truthfulness is based on the subjectivity and experience of the polygraph examiner. This inherent subjectivity has led to the debate over the accuracy of polygraph tests.

Despite the subjectivity and perhaps spurious assumptions associated with the polygraph, it is still widely used throughout the nation. The range of error in which the polygraph operates is at the root of the debate over whether polygraph tests should be administered at all. U.S. Attorney General John Ashcroft concedes that polygraphs used in federal agencies have an error rate of about 15% [1]. Some critics even assert that credible scientific research has found that commonly used polygraphs have error rates of 40% or more, only slightly better than flipping a coin to decide if a subject is lying [1].

These errors are not usually in the physical measurements made by the machine, but rather are embedded in the assumption that physiological conditions can indicate the psychological state of lying. Critics question the theory by which the polygraph operates, rather than the quantitative measurements it provides. It may be generally true that when people lie, they exhibit certain physiological phenomena. However, there may be other reasons for an honest subject to demonstrate these same physiological signs. For example, a truth-teller might recognize that a question has significance in the investigation and exhibit the same increased cardiovascular and sweat activity as a liar. This innocent subject may be wrongly accused of lying, with very serious and unjust consequences.

Some people, such as spies, are trained to disguise their feelings, can lie very naturally, and can easily pass polygraph tests. Furthermore, since polygraphs rely heavily on interpretation by polygraph examiners, human error and bias can create inaccuracy in the results. Why, then, is the polygraph still used today? First, the willingness to undergo a polygraph test often provides proof that the subject has nothing to conceal. Second, the response to certain questions in the polygraph tests can sometimes lead investigators to new focuses [3]. Finally, fear of the test can prompt the guilty to confess [6].

There are new technologies that can be used for lie detection. One is the facial thermal imaging, a technology that maps facial blood flow. When a person lies, he or she often becomes anxious and excessive blood flows to areas around the eyes. This blood flow can be detected by a thermal imaging screener [7]. Lasers have been developed to detect muscular, circulatory, and other bodily changes assumed to be associated with the anxiety of lying [8]. Some computer programs claim the ability to detect lies by analyzing the voice and tone of a speaker. According to their inventors, “when a person lies, an involuntary interference of the nerves causes the vocal cords to produce a distorted sound wave, namely a frequency level which is different from the one produced by the same person when telling the truth” [9]. One company has even developed a lie-detecting keyboard, which is claimed to be able to detect lies when a person types into a computer by analyzing typing patterns, sensing moisture in fingertips, recording body heat, and monitoring how fast the fingers were moving when they hit the keyboard [10].

Some of these methods may be more accurate than modern polygraphs due to the replacement of subjective judgment by an examiner with quantitative analysis by computers. However, all of these techniques are still based on the assumption that lying is associated with certain physiological changes. These technologies were geared towards the identification of changes in the physiological conditions that might or might not be a direct result of lying.

Very recently, researchers have discovered that certain regions of the brain exhibit unique activity during lying [11]. Interestingly, one of these regions, the anterior cingulate cortex, is a region that has been linked to conflict monitoring as well as attention and response inhibition [12]. This coincides with the idea that lying causes a conflict within the brain between the lie and the truth. The increased activity can be detected by functional magnetic resonance imaging (fMRI), which records brain activity by identifying changes in brain blood flow and metabolic rate [12].

Researchers of this technology warn that they “do not claim to have identified the signature of deception” [12]. However, their discovery is a step closer to developing a lie detector that does not depend on nonspecific physiological vectors that can be induced by conditions other than lying.

The polygraph is a device that seeks to accurately and reliably detect lies – an ability that humans have sought to develop throughout the centuries. However, due to its inherently subjective nature, the use of the polygraph is surrounded with controversy. Regardless, the polygraph does serve as an instructive tool for government authorities and law enforcement agencies. With technological advancements, humans will be able to better correlate the psychological state of lying with physiological responses.