A new development will make it possible to reveal "liars" by movements of the facial muscles
Lies may not make our noses longer, but they do make the facial muscles work uncontrollably. For the first time in the world, researchers from Tel Aviv University were able to identify 73% of lies based on the contractions of the facial muscles when telling the lie. In addition, the researchers were able to identify two groups of "liars": those whose lies make their cheek muscles jump and those who lie above their eyebrows. According to the researchers, the new study has many implications regarding the detection of lies in all areas of life, such as security and crime for example.
The study was conducted by a team of experts from Tel Aviv University, led by Prof. Yael Hanin from the School of Electrical Engineering and Prof. Dino Levy From the Faculty of Management by Kohler, with the participation of Dr. Anastasia Shuster, Dr. Lilach Islandberg, Dr. Uri Osami and PhD student Liz Isaacson. The study was published in the prestigious journal Brain and Behavior.
How can you detect a lie anyway?
"Many studies have shown that it is impossible to really detect a lie, and all the people who claim that they know how to detect a 'poker face' - are deluding themselves," explains Prof. Levy. "There are experts, for example police investigators, who are a little more successful, but only a little. Whereas the existing truth detector technology is so problematic that it is not even admissible in court. You can always learn to control your pulse and outwit the machine. From all this it follows that a more reliable and accurate technology is needed to detect lies. One basic premise of the research is that facial muscles twitch when we lie, but until now the electrodes simply weren't sensitive enough to measure these twitches."
The new research was made possible thanks to an innovative and groundbreaking development from the laboratory of Prof. Yael Hanin: stickers printed on soft surfaces and containing special electrodes, which allow monitoring and measuring the activity of muscles and nerves. The development, which is already commercialized through the company X-trodes, has many applications such as sleep monitoring from home, early detection of neurological diseases and rehabilitation, but this time the researchers from Tel Aviv University decided to use it in a different direction: lie detection.
As part of the experiment, the researchers stuck the special electrodes on two groups of facial muscles: the cheek muscles next to the lips and the muscles above the eyebrows. The experimenters were asked to sit facing each other, wearing headphones that played the words "line" or "tree". When one experimenter heard "line" and said "tree", or heard "tree" and said "line", he obviously lied - and the person sitting in front of him had to try and identify the lie.
In the second phase, the experimenters switched, and the guesser was asked to tell a truth or a lie. As expected, the participants in the experiment were not able to identify if they were lied to with statistical significance, but the electrical signals from their faces allowed the researchers to reach an unprecedented result of detecting the lie in 73% of the cases.
So how do you lie - with the cheek or with the forehead?
"This is preliminary research, so the lie itself was simple," says Prof. Levy. "Usually, when we lie, we tell a story longer than a 'line' and a 'tree', with elements of truth and elements of lies. But the research advantage here is that we knew what was being said in the headphones, that is, we knew when a lie was being told and when the truth was being told, and so we trained the software using sophisticated machine learning to detect lies based on the EMG signals in the electrodes, and we reached an accuracy of 73% - not perfect, but much better than any technology exists. Another interesting result was that different people lie using other facial muscles: some lie with the cheek muscles and some with the muscles above the eyebrows."
These results may have dramatic implications for many aspects of our lives, as basic research into the physiology of the face when lying can eliminate the need for electrodes and train video software to detect lies based on the muscle movements themselves. "In the bank, in the interrogation room, at the airport or just in a Zoom job interview, high-resolution cameras that have been trained to detect the movements of the facial muscles will know when we are telling the truth and when we are lying," Prof. Levy concludes. "As soon as we pass the experimental phase, trust the software and eliminate the need for electrodes, the applications are many and varied."
"Many studies have shown that it is impossible to really detect a lie, and all the people who claim that they know how to detect a 'poker face' - are deluding themselves," explains Prof. Levy. "There are experts, for example police investigators, who are a little more successful, but only a little. Whereas the existing truth detector technology is so problematic that it is not even admissible in court. You can always learn to control your pulse and outwit the machine. From all this it follows that a more reliable and accurate technology is needed to detect lies. One basic premise of the research is that facial muscles twitch when we lie, but until now the electrodes simply weren't sensitive enough to measure these twitches."
More of the topic in Hayadan:
