Is artificial intelligence able to help with pain assessment?

Abstract

Most people are able to communicate in an expressive manner when they are in pain. But what about people who are not able to report their pain experience or whose expression of pain is hard to interpret? Examples include individuals suffering from dementia, patients developing delirium, and newborns. Techniques that provide a reliable assessment of pain experience are a prerequisite for effective pain therapy. Due to recent advances in the automated detection and analysis of behavioral cues, the question arises of whether these techniques may help assess pain-related states in a reliable manner. Physical pain is closely related to emotional states that may modulate the experience of pain and vice versa. Furthermore, principles and techniques from affective computing provide a solid basis for the automated analysis of pain-related states. Thus, it comes as no surprise that research on pain has increasingly attracted the interest of the affective computing community.