How to Distinguish Real Pain Expressions From Deceptive Ones

Sometimes it may be necessary to learn how to distinguish real pain expressions from deceptive ones. Not only does this aid in deception detection, but it can also help further our understanding of the pain expression, itself!

In order to address this question, a team of researchers from the University of California, San Diego, and the University of Toronto ran two experiments trying to see if observers could reliably distinguish faked pain from real pain, including one that implemented this analysis with computer learning software.

Initially, the treatment conditions had to be determined. This involved recruiting an admittedly small number of 26 participants who were asked to be recorded under three conditions. In addition to a baseline where no pain expression was solicited, they were also induced to present a genuine or faked pain expression.

This was stimulated by having each participant submerge their arm in water, either at a lukewarm or freezing cold temperature. When the water was lukewarm, they were asked to act as though it was cold and painful, while the freezing cold water reliably generated the desired expression of genuine pain. These images were evaluated first by a new set of 170 volunteers but also by an automated system which tried to analyze facial expressions without human interference.

In the first of these, the new volunteers were exposed to video clips of the genuine and fake pain expressions and asked to distinguish them. Interestingly, their success was almost entirely random, at about 52 percent accuracy. Notably, they had no prior experience making these distinctions.

While this first study indicates that people without training can’t seem to distinguish real pain from deception, they also cite previous research finding that even clinicians do not perform much better.

However, machine algorithms may make up for our human failures in this regard. Each clip was broken down into a set of discrete frames, allowing a computer to look at individual stills extracted from the videos. These were then paired with the previously discussed FACS software which codes images based on a preexisting series of templates representing typical facial expressions, in order to see if the software could also distinguish between faked and genuine pain expressions.

While the software was able to reliably distinguish with an accuracy of over 70 percent, what is more interesting is exactly what components of the facial expression were most useful in doing so. For instance, faked expressions relied on a lowered brow which was not present in genuine instances of pain. Similarly, raised cheeks and fearful brows were also reliable indicators of deception.

This study shows us a couple of pretty interesting things. First, human efforts alone in distinguishing real and fake expressions were prohibitively inaccurate, at least for untrained participants. Second, computer methods were much more accurate. This is perhaps the most interesting finding, given the potential to feed any videotaped facial expression into this software to determine whether the subject is lying.

Importantly, this blog is written as part of Pain Awareness Month. While many people, especially those with chronic illnesses, are frequently accused of faking their pain, this is not an effort to exacerbate that. Instead, it shows how unreliably doctors determine whether patients are telling the truth, often to the patients’ detriment, and it helps show the way forward to a more reliable approach.

Moreover, we have much to learn in how to accurately detect deception. There is no reason that similar approaches cannot be applied to fear or sadness, and while you might be inaccurate untrained, there is a reason we offer a robust training program in deception detection!