We know that humans are terrible at distinguishing real pain from faked pain, but could your smartphone just do that for you?
Building off of last week’s blog, it seems that some researchers have sought to implement the insight that computers are better at detecting genuine pain than doctors. In an exciting new study, the University of Pittsburgh’s Dr. Jeffrey Cohn sought to explore the potential of implementing machine learning techniques into an accessible software to read people’s subtle pain expressions.
Especially given current political trends and concerns as to the over prescription of opiates, doctors are under increased pressure to distinguish people who actually need medication for pain from those who are just faking in. This is helpful for doctors but also for patients as well. In fact, many patients, especially those who suffer from chronic pain, struggle to convince doctors that their pain is genuine.
If Dr. Cohn is right, however, future doctors can just pull out a smartphone and take a brief video of the patient’s face and expressions. With the app under development, a computer algorithm can match patients’ facial expressions to past video templates of people who suffer from genuine pain.
This algorithm was trained by analyzing a series of videos of people’s faces while attempting to complete manual tasks despite a shoulder injury. It tracked their winces and grimaces, creating a database of what facial expressions are most reliably caused by their feeling of pain.
Last week, we emphasized the factors that indicate deception, such as a lowered brow and raised cheeks, and Dr. Cohn’s work emphasizes that genuine pain is indicated by movement around the nose and mouth.
While these tips might help us think we can detect pain effectively, we have to recognize how fallible human efforts to detect genuine pain are, even for doctors. This fits into what we have long said about how difficult many forms of deception detection and microexpression reading are for those who are not trained in them.
However, the notion of using an app to read expressions is an exciting one, both for its practical benefits but also for the intellectual potential of driving forward our understanding of expression recognition. There is no reason why this sort of app cannot be used for other emotions.
In fact, last year we wrote about an app that used artificial intelligence to recognize our emotions. In that case, it was used to create emojis that could be sent online while still accurately representing our real facial expressions. With facial recognition technology like that, combined with a machine learning database similar to what Dr. Cohn used, the potential to use technology to enhance emotional recognition is exciting indeed.
There’s a balance, of course, in deciding how much we want to trust technology without learning these skills ourselves. While trained medical professionals may struggle at detecting genuine pain, that does not mean that training specifically in expression reading cannot help. Our time-tested and proven training programs are great evidence of your potential to learn how to do this on your own, especially because even a trained machine isn’t right all the time!