We've seen it in movies many times before … the reluctant witness who looks right at the mug shot of the murderer, and falsely claims that they don't know him. Soon, however, police could know if such people are lying – by watching their eyes.

Already in Japan, authorities utilize an eye-tracking process known as the concealed information test (CIT) in order to see if someone possesses "inside information" regarding a crime. This info could be something that only the guilty party would know, such as the location or type of weapon used in a murder. Until now, though, CIT had never been used to assess people's reactions to faces.

As part of the ConFace project, a team led by Dr. Ailsa Millen – from Scotland's University of Stirling – decided to change that.

They did so by getting test subjects to view a dataset of facial photographs, all but one of which were of people they didn't know. With all the photos, they were told to indicate that they didn't know the person, by pressing a button and saying "no." Additionally, they were instructed to make a point of not moving their eyes any differently when seeing the shot of the person they knew.

It turns out that even though the volunteers tried to control their actions, eye-tracking hardware revealed that when most of them viewed the familiar face, they displayed fewer and longer eye fixations (these are the locations upon which a person's gaze rests). In fact, the harder that they tried to conceal the fact that they knew the person in the photo, the more pronounced their telltale eye movements were.

"Some witnesses are honest – but many are hostile and intentionally conceal knowledge of known identities," says Millen. "For example, criminal networks – such as terrorist groups – might deny knowledge to protect one another, or a victim might be too afraid to identify their attacker … The main aim [of our study] was to determine if liars could conceal recognition by following instructions to look at every familiar and unfamiliar face with the same sequence of eye fixations – in short, they could not."

A paper on the research was recently published in the journal Cognitive Research: Principles and Implications.