How the Polygraph Actually Measures Physiological Arousal Instead of Deception

The American Polygraph Association maintains that a properly conducted polygraph test can detect deception with up to 89% accuracy. Law enforcement agencies and federal departments rely on that figure to screen applicants and interrogate suspects, treating the machine as a definitive arbiter of truth. But the instrument itself does not measure honesty. According to the National Research Council and the American Psychological Association, the polygraph strictly records sympathetic nervous system arousal. Because the human body produces the exact same biological signals when experiencing fear, anxiety, or anger as it does when lying, the scientific consensus is that a unique physiological signature for deception simply does not exist.[1][3][4]

The device, first developed in 1921 by a law enforcement officer, is essentially a collection of medical sensors repurposed for interrogation. It does not read minds or analyze cognitive load. Instead, it tracks three specific biological metrics that spike when the body enters a "fight or flight" state. Examiners use these metrics to infer deception, a leap that bridges hard physiological data with subjective psychological interpretation.

The first signal is cardiovascular activity. Examiners wrap a standard sphygmomanometer—a blood pressure cuff—around the subject's arm to track continuous changes in blood pressure and heart rate. When a person feels threatened or stressed, their heart pumps faster and harder to deliver oxygen to the muscles. The polygraph records these microscopic cardiovascular fluctuations as a continuous line on a graph.[4][5]

The second signal is respiratory rate. Two pneumographs, which are rubber tubes filled with air, are strapped around the subject's chest and abdomen. These tubes expand and contract as the subject breathes, measuring both the depth and the frequency of respiration. A sudden intake of breath, a suppressed sigh, or a quickened breathing pattern all register as deviations from the subject's baseline respiratory rhythm.[4][5]

The third and most sensitive signal is electrodermal activity, historically known as the galvanic skin response. Examiners attach metal electrodes to two of the subject's fingers. These sensors measure the electrical conductivity of the skin, which changes based on sweat gland activity. Even a microscopic, imperceptible increase in perspiration—triggered by the sympathetic nervous system during a moment of stress—will cause the skin's electrical resistance to drop, sending a sharp spike across the polygraph monitor.[4][5]

To translate these three signals into a verdict of truth or deception, examiners typically rely on the Comparison Question Test (CQT). During a CQT, the examiner asks "relevant" questions directly tied to the investigation, such as, "Did you steal the money from the safe?" They intersperse these with "control" questions designed to provoke a baseline stress response, such as, "Have you ever lied to get out of trouble?"[5]

The entire methodology rests on a single psychological assumption: an innocent person will be more anxious about the broad control questions, while a guilty person will show a sharper physiological spike on the relevant questions. If the blood pressure, respiration, and sweat metrics jump higher on the relevant questions, the examiner scores the test as deceptive.[5]

That assumption is where the scientific community draws a hard line. In 2003, the National Research Council published a landmark, 398-page evaluation of polygraph science. The committee found that while the machine accurately records biological changes, the link between those changes and lying is fundamentally flawed. An innocent suspect terrified of going to prison will often show massive physiological spikes on relevant questions simply because those questions carry life-altering consequences.[1]

The American Psychological Association echoes this conclusion. In their 2025 review of the technology, the APA noted that "scientific evidence suggests that polygraph tests are not a reliable and valid way to detect deception." The biological markers of lying are indistinguishable from the biological markers of being interrogated.[4]

The gap between what the machine measures and what it claims to detect creates a severe mathematical problem when polygraphs are used for security screening. The American Polygraph Association claims an 85% accuracy rate for multi-issue screening tests. However, applying that 15% error rate to a large population demonstrates the base-rate fallacy. If a federal agency screens 10,000 employees to find 10 actual spies, the 15% error rate will flag approximately 1,500 innocent people as deceptive. The machine produces a mountain of false positives while statistically missing at least one of the actual threats.[1][3]

Because of this unreliability, the U.S. legal system largely rejects polygraph evidence. According to the Legal Information Institute at Cornell Law School, "The test cannot actually test for honesty, however. Instead, it relies on the polygraph operator's analysis of the tested person's responses, which can be inaccurate." Consequently, roughly half of all U.S. states ban polygraph results from criminal trials entirely, while the rest generally require both the prosecution and the defense to stipulate to their admissibility before the test is even administered.[5]

Despite the scientific and legal consensus, federal agencies continue to administer thousands of polygraphs annually. A report by the Congressional Research Service examining the Department of Energy's polygraph program highlights the real reason the practice endures: it works as an interrogation prop. The machine does not need to be a scientifically valid lie detector if the person sitting in the chair believes it is. The stress of the examination and the fear of the sensors routinely prompt applicants to volunteer disqualifying information or confess to crimes.[2]

The polygraph remains a fixture in law enforcement not because it decodes deception, but because it leverages human psychology. The blood pressure, breathing, and sweat data it prints out are entirely real. The interpretation of that data as a lie is a subjective judgment that the scientific evidence simply cannot support.[1][4]