Faces of clinical pain: Inter‐individual facial activity patterns in shoulder pain patients

Abstract

Background Facial activity during pain is composed of varying combinations of a few elementary facial responses (so‐called Action Units). A previous study of experimental pain showed that these varying combinations can be clustered into distinct facial activity patterns of pain. In the present study, we examined whether comparable facial activity patterns can also be identified among people suffering from clinical pain; namely, shoulder pain. Methods Facial expressions of patients suffering from shoulder pain (N = 126) were recorded while twice undergoing a battery of passive range‐of‐motion tests to their affected limbs (UNBC‐McMaster Shoulder Pain Expression Archive Database), which elicited peaks of acute pain. Facial expressions were analysed using the Facial Action Coding System to extract facial Action Units (AUs). Hierarchical cluster analyses were used to look for characteristic combinations of these AUs. Results Cluster analyses revealed four distinct activity patterns during painful movements. Each cluster was composed of different combinations of pain‐indicative AUs, with one AU common to all clusters, namely, “narrowed eyes”. Besides these four clusters, there was a “stoic” pattern, characterized by no discernible facial action. The identified clusters were relatively stable across time and comparable to the facial activity patterns found previously for experimental heat pain. Conclusions These findings corroborate the hypothesis that facial expressions of acute pain are not uniform. Instead, they are composed of different combinations of pain‐indicative facial responses, with one omnipresent response, namely, “narrowed eyes”. Raising awareness about these inter‐individually different “faces of pain” could improve the recognition and, thereby, its diagnostic training for professionals, like nurses and physicians. Significance Similar to experimental pain, facial activity during evoked pain episodes in shoulder pain patients could be clustered into distinct faces of pain. Each cluster was composed of different combinations of single facial responses, namely: narrowed eyes, which is displayed either alone or in combination with opened mouth or wrinkled nose, or furrowed brows and closed eyes. These distinct faces of pain may inform the training of professionals and computers how to best recognize pain based on facial expressions.

Are Chronic Pain Patients with Dementia Being Undermedicated?

“… 16 It is important to realize that these facial movements are not typically displayed together when somebody is in pain, as individuals usually show different mixtures of each facial movement. 17 , 18 The most frequent response to pain is a tightening of the muscles around the eyes, paired with one or two other pain-indicative facial movements.…”

Computer Mediated Automatic Detection of Pain-Related Behavior: Prospect, Progress, Perils

“…The models and solutions that have shown promise for automatic detection are based on limited sampling. There is considerable evidence from direct facial measurement studies that facial expressions of pain involve a common core of actions (23,52), but recent findings indicate that those actions come in different clusters (73,74). This points to a need to collect further databases that sample a broader range of pain types as a way to assess the generality and generalizability of extant and novel models and solutions.…”

Faces of Pain in Dementia: Learnings From a Real-World Study Using a Technology-Enabled Pain Assessment Tool

“…Kunz et al recently reported that among participants with shoulder pain undergoing passive movement of the joint, the most common cluster of facial expression was “narrowed eyes” related to contracture of AU6/7. The facial activity was displayed in approximately 40% of participants on repeated exposure to movement of their shoulder ( 19 ). The authors also reported another cluster of facial activity, which involved AU6/7, AU4 (brow lowering), and AU43 (eye closure) in approximately 10% of participants (“narrowed eyes with furrowed brow and closed eyes”) ( 19 ).…”

“…The facial activity was displayed in approximately 40% of participants on repeated exposure to movement of their shoulder ( 19 ). The authors also reported another cluster of facial activity, which involved AU6/7, AU4 (brow lowering), and AU43 (eye closure) in approximately 10% of participants (“narrowed eyes with furrowed brow and closed eyes”) ( 19 ). This is in line with our findings that suggest that the upper face (i.e., eyes and forehead) appears to be the most expressive region of the face in moderate-severe pain.…”

“…The PainChek® pain assessment scale is a 42-item instrument, which has items distributed across six domains [Face (9 items), Voice (9 items), Movement (7 items), Behavior (7 items), Activity (4 items), and Body (6 items)]. The Face is a FACS-based domain, which has nine pain-relevant AU items representing the most common facial features that are associated with the presence of pain ( 12 , 16 , 18 , 19 ). These include upper face AUs: AU4 (brow lowering), AU6 (cheek raising), AU7 (tightening of eyelids), AU9 (wrinkling of nose), and AU43 (closing eyes); and lower face AUs: AU10 (raising of upper lip), AU12 (pulling at corner lip), AU20 (horizontal mouth stretch), and AU25 (parting lips).…”