Facial and bodily expression of pain: into the future with automation
8 Jun 2015 | Print
Words are the most obvious mechanism by which we express pain – whether through a simple “ouch” or through rather stronger language! However, pain also can be communicated in a wide variety of non-verbal ways, including facial expressions (eg, eye squeezing or teeth clenching), non-verbal vocalizations (eg, crying or groaning) and bodily movements.
These verbal and non-verbal actions have been called ‘pain behaviors’ and they play a key role in communicating that pain is being experienced. Some serve primarily to modify the pain itself (eg, rubbing of the site), while others may be specifically adapted for social communication.
Impact of pain expression on others
Expressions of pain can have a major impact on how we are viewed by those around us. For example, in a recent study, participants watched video footages of patients with chronic back pain performing a physically demanding lifting task. Unsurprisingly, those patients who displayed pain behaviors – whether communicative (eg, facial expressions) or protective (eg, guarding) – were perceived as having significantly more pain than those who did not display pain behavior. However, rather more worryingly, patients that showed protective pain behaviors were also perceived as being less likeable, less dependable and less ready to work than patients displaying other forms of pain behavior.
In a second study, participants viewed two video clips of human figures exercising. The videos were created by a motion tracking system and showed only dots placed at various points on the body, so body motion was the only visible cue. In one of the clips, the figure displayed pain behaviors (such as rubbing, holding and hesitating); in the other clip, the figure did not. Again, participants judged the person displaying pain behaviors to be in more pain. However, participants also perceived them as being less warm, less competent, more negative in mood and less physically fit than the person who did not display pain behaviors.
The implication of these social evaluations is clear: when patients with pain say that they feel stigmatized, they are probably right. Society may be making adverse judgments about them as human beings.
Even healthcare professionals (HCPs) frequently misjudge their patients’ pain.[4,5] Indeed, studies have shown not only that HCPs typically underestimate pain, but also that their assessment is influenced by a wide variety of demographic factors, including patients’ race, ethnicity, culture, gender and socioeconomic status.
Understanding pain behaviors
Various methods have been developed for analyzing pain behaviors. Many use self-reports via questionnaires or checklists, or observations from trained observers. Usually, these methods are based on facial expressions, although a recent study identified a set of body postures that communicate pain.
However, these techniques are typically subjective and often provide inadequate timing information with regard to pain.
New developments in computer-based technologies have the potential to offer more sophisticated tools for assessing pain behavior. For example, in a recent video-based study, so-called ‘facial action units’ were used to obtain an objective measure of pain on a frame-by-frame basis. This system can be used to automatically detect the frames in which a patient is in pain.
Furthermore, novel sensing technologies and affective computing methods – in which the device detects and responds appropriately to the user’s emotions and other stimuli – could, one day, be harnessed to automatically recognize, interpret and act upon affective states like pain.
For example, the ‘Emotion & Pain Project’ is involved in the design and development an intelligent system for monitoring and assessing pain-related mood and movements. The specific aims of the project are to:
- Develop methods for automatically recognizing audiovisual cues and behavioral patterns that relate to pain (specifically of the lower back), as well as affective states that influence pain; and
- Integrate these methods into a system that will provide feedback and prompts to the patient, based on their behavior during physical therapy sessions.
However, the key question for this type of work is: which specific behaviors should the automatic system be based upon? This remains a work in progress, but the focus so far has largely been on simple non-interactive behaviors:
- Hesitation – during continuous movement
- Guarding – stiff, interrupted or rigid movement
- Support – using support for movements that could be done without it
- Altered gait – irregular cadence, stride, timing and/or weight-bearing
- Rubbing/stimulating – massaging or touching an affected body part with another; superfluous hand movements while stationary
So far, the most promising of these behaviors, with regard to automation, appears to be guarding. Much work remains to be done before these technologies can be used in clinical practice, but there is no doubt that they offer substantial potential benefits.
Eventually, they could meaningfully improve the management of chronic pain – major step forward for pain patients.
- 1. Prkachin KM. Assessing pain by facial expression: facial expression as nexus. Pain Res Manag 2009;14:53-58.
- 2. Martel MO, Wideman TH, Sullivan MJ. Patients who display protective pain behaviors are viewed as less likable, less dependable, and less likely to return to work. Pain 2012;153:843-849.
- 3. Ashton-James CE, Richardson DC, de C Williams AC, Bianchi-Berthouze N, Dekker PH. Impact of pain behaviors on evaluations of warmth and competence. Pain 2014;155:2656-2661.
- 4. Kappesser J, Williams AC, Prkachin KM. Testing two accounts of pain underestimation. Pain 2006;124:109-116.
- 5. American Society of Anesthesiologists Task Force on Acute Pain Management. Practice guidelines for acute pain management in the perioperative setting: an updated report by the Anesthesiology 2012;116:248-273.
- 6. Walsh J, Eccleston C, Keogh E. Pain communication through body posture: the development and validation of a stimulus set. Pain 2014;155:2282-2290.
- 7. Lucey P, Cohn JF, Matthews I, et al. Automatically detecting pain in video through facial action units. IEEE Trans Syst Man Cybern B Cybern 2011;41:664-674.
- 8. The Emotion & Pain Project. Available at: www.emo-pain.ac.uk. Accessed March 2015.
- 9. Aung MSH, Bianchi-Berthouze N, Watson P, Williams AC de C. Automatic recognition of fear-avoidance behavior in chronic pain physical rehabilitation. Presented at the Eight International Conference on Pervasive Computing Technologies for Healthcare. 2014. Available at: www.emo-pain.ac.uk/papers/PervasiveHealth2014.pdf. Accessed March 2015.
With Professor Sam AhmedzaiIndividualizing cancer pain therapy
With Dr Judith PaiceCancer pain: Are we making progress?
With Dr Karen H SimpsonAcupuncture for chronic pain and pain vulnerability
With Dr Panagiotis Zis and Professor Kallol Ray ChaudhuriPain in Parkinson’s disease: classification and management