The development and use of facial grimace scales for pain measurement in animals
Section snippets
Pain assessment and treatment in a veterinary context
Most of the understanding about veterinarian attitudes and concerns around pain recognition and the provision of analgesics stems from surveys about domestic pets. These surveys have shown that analgesics are underused in acutely painful conditions even in pet dogs and cats (e.g., Capner et al., 1999; Dohoo and Dohoo, 1996; Joubert, 2001; Watson et al., 1996). In the 1990s, rates of analgesic use (opioids and non-steroidal anti-inflammatories) for elective surgery (e.g., ovariohysterectomy and
Pain measurement in a biomedical research context
Success in translating basic science knowledge into novel clinical treatments for humans has been limited for many decades now (Scannell et al., 2017), and the pain research field is no exception. In fact, the only major new class of analgesics to be approved in recent years are the calcitonin gene-related peptide (or receptor) antibodies for the treatment of migraine; failed analgesic targets (or, at the very least, failed clinical trials) include: calcium channel blockers, cannabinoid
Facial expressions of pain in humans
Pain assessment in humans has typically relied on self-report measures (e.g., 0–10 numeric ratings scales or visual analogue scales). Facial expression emerged as a method to measure pain in individuals who are unable to express themselves verbally (e.g., neonates, young children, those with verbal or cognitive impairments) (Hadjistavropoulos et al., 2011; Prkachin, 2009). Interest in the facial expression underlying human emotions first arose in the 1970s. During this period, Ekman and Friesen
The development of the mouse grimace scale and the rat grimace scale
The deficiencies in the status quo of pain measurement in both the veterinary and biomedical research contexts, and the success of pain measurement via facial expression in non-verbal humans, provide the background context for the development of animal grimace scales. However, the specific impetus to develop the Mouse Grimace Scale (MGS) (Langford et al., 2010a) was not direct, but rather was first conceived as one of many follow-up experiments to the discovery in 2006 by the first author and
What are grimace scales measuring?
Pain in humans is a multidimensional experience, including at the very least sensory-discriminative, motivational-affective, and cognitive components (Melzack and Casey, 1968). Others have made convincing cases for viewing pain as a homeostatic emotion or drive state (Craig, 2003), and as having a fundamental social communication role (Craig, 2009). The insula lesioning study reported in Langford et al. (2010a) suggested that grimacing might reflect the affective component of pain. It remains
The duration of grimacing
In the original grimace scale development studies in mice and rats (Langford et al., 2010a; Sotocinal et al., 2011), no statistically significant elevation of grimacing scores over baseline levels were observed in assays in which pain duration was less than ≤10 min or more than 24 h post-injury. We concluded that grimacing in animals occurred only in this time range. This pattern is somewhat different to that observed in humans. Grimacing to acute noxious stimuli definitely occurs in people,
Psychometric principles of grimace scales
A key concept in the development of health assessment scales, such as grimace scales, is applying psychometric principles to provide indications of validity and reliability (Streiner and Norman, 2008). Validity can be simply defined as, “does a scale measure what it claims to measure?” That is: 1) are the items making up a scale necessary and important (content validity), 2) how does a scale compare to a criterion standard should one exist (criterion validity), and 3) whether a scale can
Confounding and modulatory factors of grimace scales
Any scale should ideally be specific to the variable of interest (e.g., pain) and minimally affected by unrelated variables (e.g., anxiety). This assessment, known as divergent validity, has received minimal attention in the grimace scale literature. This is part due to a sparsity of validated instruments for comparison. Using simpler, less direct approaches, grimace scales have nonetheless been demonstrated to be affected by a number of factors in the research or clinical environment,
The relationship between facial grimacing and other measures of pain
Grimacing as a measure of pain is only one of many new measures that have been developed in recent years in response to published calls for same (Mao, 2002; Mogil and Crager, 2004; Vierck et al., 2008). The reader is directed to a comprehensive review of these new measures (Mogil, 2019a). Briefly, they can be categorized as follows: 1) biomarkers, 2) pain-stimulated behavior (including grimacing), 3) pain-depressed behavior, 4) measures of pain-related functional disability, and 5) methods
Do grimace scales work in real time?
A drawback of the original grimace scale methodology is the dependence on generating still images from video recordings. While providing a temporal record of behavior and allowing confirmation that confounding behaviors (e.g., grooming, sleeping) are not present at the time of image capture, time spent video recording following by image selection and processing adds considerably to the time and labour required to generate data. In a research setting this is inconvenient but surmountable,
What have we learned from grimace scales?
In a veterinary and welfare context, grimace scales have been applied to re-evaluate analgesic efficacy and the potential for pain associated with common husbandry procedures and invasive models.
Future developments
The development and use of grimace scales has the potential to transform both preclinical pain research and veterinary pain research and management. It will never do so with its original implementation, requiring the laborious steps of image collection from video (although this can be automated with Rodent Face Finder® software; (Sotocinal et al., 2011) and manual scoring. However, standard machine learning (artificial intelligence) algorithms can be taught to recognize grimacing. This has
Acknowledgments
This work was supported by the Canadian Institutes for Health Research (Foundation Grant FRN154281 to J.S.M.), the Natural Sciences and Engineering Research Council of Canada (Discovery Grant 77922 to J.S.M.; Discovery Grant 42402 to D.J.P.), Fondation Levesque (to D.J.P.), and the Louise and Alan Edwards Foundation (to J.S.M.). The authors thank Dr. Kenneth Craig for valuable input on facial expressions in humans.
References (170)
- et al.
The grimace scale reliably assesses chronic pain in a rodent model of trigeminal neuropathic pain
Neurobiol. Pain
(2017) - et al.
The role of TRPA1 in muscle pain and mechanical hypersensitivity under inflammatory conditions in rats
Neuroscience
(2015) - et al.
Neuropathic pain symptoms relation to overall pain rating
J. Pain
(2004) The pain of pain: challenges of animal behavior models
Eur. J. Pharmacol.
(2015)What is spontaneous pain and who has it?
J. Pain
(2012)- et al.
Facial correlates of emotional behaviour in the domestic cat (Felis catus)
Behav. Processes
(2017) Pain-like behaviours in animals - how human are they?
Trends Pharmacol. Sci.
(2004)- et al.
Behavioral treatment of chronic pain: the spouse as a discriminative cue for pain behavior
Pain
(1980) - et al.
The influence of demeanor on scores from two validated feline pain assessment scales during the perioperative period
Vet. Anaesth. Analg.
(2017) - et al.
Postingestive alliesthesia: the rat tells the same story
Physiol. Behav.
(1990)