To create the system, Ekman and his team created a map of the muscles activated during a wide variety of facial expressions. The FACS labels specific facial movements in Action Units, or AUs: raising the eyebrows is AU 1, for example, while narrowing the eyes is AU 6. To avoid bias, the system only uses numbers, not names, for each expression, and it can detect whether an expression is fake or genuine, voluntary or involuntary, and intentional or spontaneous. Studies have shown that the system can spot a liar with 80% accuracy, and may also detect other important characteristics like heart problems and suicide risk.
From psychics to lie detectors, humans have always looked for ways to identify when someone is telling the truth. Researcher Paul Ekman may have developed one such method. It's called the Facial Action Coding System, or FACS, and it's surprisingly accurate.