A grimace is a reflexive tightening and distortion of the face that most commonly signals pain, but it also surfaces during intense physical effort, disgust, and moments of acute emotional distress. The expression is anchored by a specific cluster of muscle movements: the brows draw together and down, the eyes squeeze shut, the nose wrinkles, the upper lip lifts, and the mouth opens. These movements are so consistent across people that researchers have catalogued them into a standardized set of facial “action units,” and the same basic pattern appears in newborns, across cultures, and even in other species.
The Muscles Behind the Grimace
Facial expression research relies on the Facial Action Coding System, which maps individual muscle movements to numbered action units (AUs). A systematic review of studies on pain-related facial movements identified a core set that recur across experiments and clinical settings: lowering the brows (AU4), tightening the area around the eyes with a cheek raise and lid compression (AUs 6 and 7), wrinkling the nose and raising the upper lip (AUs 9 and 10), and opening the mouth (AUs 25, 26, and 27).1Pain. Facial muscle movements encoding pain—a systematic review If you picture someone who just stubbed a toe, you are probably imagining exactly this combination. The fact that these action units show up so reliably is part of what makes the grimace useful to clinicians, caregivers, and even automated detection systems: the face is doing something measurable and predictable, not just vaguely contorting.
Not all parts of the grimace carry the same information. Research using hypnotic suggestion to independently manipulate the sensory intensity and emotional unpleasantness of pain found that movements around the eyes track more closely with how intense the pain feels, while brow and upper-lip movements are more tied to how distressing it is.2PubMed Central. Are both the sensory and the affective dimensions of pain encoded in the face? In other words, a grimace is not one undifferentiated signal. Different regions of the face encode different dimensions of the experience, which is one reason the expression can look subtly different depending on the situation.
Pain as the Primary Trigger
Pain is the context most people associate with grimacing, and for good reason. From a biological standpoint, the pain grimace appears to be hardwired rather than learned. Newborns who have never observed another person’s face produce a recognizable pain expression in response to a heel-stick blood draw: the mouth opens, the brows draw together, and the eyes close. A study of 57 neonates from three ethnic backgrounds found this “primal face of pain” was consistent across sex and ethnicity, and it appeared even in infants whose mothers had received epidural anesthesia during delivery.3PubMed. Neonatal pain facial expression: evaluating the primal face of pain Premature infants also produce pain-related facial actions, though their expressions are less fully developed than those of full-term newborns, suggesting the system matures over the final weeks of gestation.4Pain. Developmental changes in pain expression in premature, full-term, two- and four-month-old infants
The consistency of the pain grimace across age and background is striking, and it holds across cultures as well. A cross-cultural study comparing Western and East Asian observers found that mental representations of pain-related facial movements were similar across the two groups. Interestingly, the same study found that expressions of pleasure (specifically orgasm, which superficially resembles pain) showed distinct “cultural accents,” but pain did not.5PubMed Central. Distinct facial expressions represent pain and pleasure across cultures Pain grimaces, it seems, are close to a universal language.
Grimacing During Effort and Exertion
You do not need to be in pain to grimace. Anyone who has watched a weightlifter at the top of a heavy squat or a cyclist grinding up a hill has seen the effort grimace: furrowed brows, clenched jaw, tightened face. This is not just theatre. Electromyography (EMG) studies have measured the electrical activity of frowning muscles during physical tasks and found that activity in the corrugator supercilii (the muscle that pulls the brows together) increases as task difficulty rises. When researchers induced additional muscle fatigue in participants performing leg extensions, frowning activity climbed further, tracking closely with both how hard the legs were working and how effortful the task felt to the person.6Biological Psychology. The face of effort: Frowning muscle activity reflects effort during a physical task
Jaw clenching adds another layer. During incremental cycling tests that push heart rate and perceived exertion higher and higher, masseter (jaw muscle) activity ramps up alongside heart rate and leg muscle output. The correlation is strong enough that some researchers have suggested jaw-clenching patterns could serve as a supplementary marker for how hard someone perceives they are working.7PubMed Central. Frowning and jaw clenching muscle activity reflects the perception of effort during incremental workload cycling The effort grimace overlaps with the pain grimace in the brow region, but tends to emphasize the jaw and lack the nose-wrinkling and upper-lip raising that characterize acute pain. If you are trying to tell whether someone is struggling or suffering, the area around the nose and mouth is often where the difference lives.
Voluntary Versus Involuntary Grimacing
The grimaces that matter most in everyday life tend to be spontaneous: you do not decide to grimace when you burn your hand on a stove. But people can also produce facial expressions deliberately, and these two kinds of expression rely on different brain circuits. Clinical reports of stroke patients illustrate this neatly: some patients lose the ability to smile or grimace on command but can still produce spontaneous expressions in response to emotion or pain. Others show the reverse, able to perform a voluntary smile but unable to produce a genuine one. Research in both humans and non-human primates supports the idea that voluntary facial expressions depend more heavily on cortical pathways, while spontaneous expressions run through faster, more direct circuits connecting brainstem sensory and motor nuclei to subcortical regions like the amygdala.8PubMed Central. Sensorimotor regulation of facial expression – An untouched frontier
This dual-pathway architecture helps explain why a genuine grimace can be so hard to fake convincingly. The brainstem-driven circuit fires quickly and coordinates the face in characteristic temporal patterns. When someone deliberately performs a pain face, the timing and coordination shift in detectable ways.
Spotting a Fake Grimace
Humans are generally bad at telling real pain expressions from faked ones. In experimental settings, untrained observers perform roughly at chance. But the expressions themselves are not identical, and both human coders and algorithms can pick up on the differences when they know what to look for.
Faked pain expressions tend to include a greater number of both pain-related and unrelated facial actions, last longer at peak intensity, and stretch out in overall duration compared to genuine expressions. The individual facial movements in a faked grimace also tend to be less temporally contiguous, meaning they do not flow together as smoothly.9PubMed. Detecting deception in pain expressions: the structure of genuine and deceptive facial displays It is as if fakers overshoot on the individual ingredients and undershoot on the coordination.
Automated analysis sharpens this picture further. When researchers used computational decoding to compare genuine and faked pain expressions, they found that mouth openings in faked expressions were shorter on average, the intervals between them were shorter, and the variability in timing was roughly half that of genuine expressions.10Current Biology. Automatic Decoding of Facial Movements Reveals Deceptive Pain Expressions Real pain produces a messier, more variable signal. Fake pain is too regular, too rhythmic. The face is doing what the person thinks pain should look like rather than what pain actually makes the face do.
Why Suppression Does Not Fully Work
Sometimes the goal is the opposite of faking: people try to hide a grimace. Chronic pain patients, athletes playing through injuries, and anyone who wants to appear stoic may deliberately attempt to suppress their pain expression. Research on patients with chronic low back pain found that attempts to suppress the grimace during a painful range-of-motion exercise were not entirely successful. Residual facial activity persisted, particularly in the core action units around the brows and eyes.11Pain. Genuine, suppressed and faked facial behavior during exacerbation of chronic low back pain The subcortical circuits driving a genuine pain grimace are fast and partly involuntary, so while a person can dampen the expression, they rarely eliminate it entirely. This is useful information for clinicians assessing patients who may underreport pain: the face often leaks what the patient will not say.
What Happens in Your Brain When You See a Grimace
Watching someone grimace triggers a broad network of brain regions, many of which overlap with areas active during your own pain experience. Functional imaging studies show that viewing facial expressions of pain activates the anterior cingulate cortex and the anterior insula, both of which are central to processing the unpleasantness of first-person pain. At the same time, areas involved in action observation and motor mirroring, particularly the inferior frontal gyrus and the inferior parietal lobule, also light up.12NeuroImage. Brain responses to facial expressions of pain: Emotional or motor mirroring?
A separate study compared brain responses when people viewed their own previously recorded facial expressions of pain versus those of strangers. The neural network activated was largely the same in both cases, spanning temporal, frontal, and cingulate regions.13PubMed Central. Pain Mirrors: Neural Correlates of Observing Self or Others’ Facial Expressions of Pain Your brain processes someone else’s grimace by partially simulating it, which is likely the neural basis for the instinctive wince you feel when you see another person hurt themselves. The research also found that viewing pain expressions activated medial prefrontal regions associated with social cognition and emotional understanding, suggesting the response is not purely mimicry but involves some active interpretation of what the other person is going through.
Grimace Scales in Animals
Animals cannot report pain verbally, which makes facial expression one of the most valuable tools veterinary researchers have for assessing their welfare. Grimace scales have been developed for mice, rats, rabbits, horses, and several other species, each adapted to that animal’s facial anatomy but built on the same principle: specific, scorable changes in facial features that correlate with known painful conditions.
The Mouse Grimace Scale, developed over a decade ago, scores changes in five facial features, and a scoping review of its use in biomedical research confirmed its widespread adoption as a pain assessment method.14PubMed Central. Methods Used and Application of the Mouse Grimace Scale in Biomedical Research 10 Years on: A Scoping Review In practical validation studies, mice that received analgesics after surgery showed lower grimace scores than those given saline, and the scores correlated well with independently coded pain behaviors like guarding and flinching.15PLoS ONE. The Assessment of Post-Vasectomy Pain in Mice Using Behaviour and the Mouse Grimace Scale
The Horse Grimace Scale works on a similar principle, scoring six facial action units to assess equine pain. Recent work has proposed anatomical updates and refined descriptors to improve reliability across observers.16PubMed. Enhancing the Horse Grimace Scale (HGS): Proposed updates and anatomical descriptors for pain assessment One of the practical limitations of these scales is that they depend on a trained human observer who may not be available around the clock. This has pushed researchers toward automation: a machine-learning model trained to score the Horse Grimace Scale from video was developed to evaluate pain levels in horses without requiring a human observer to be physically present or continuously watching.17PubMed Central. Pain assessment in horses using automatic facial expression recognition through deep learning-based modeling A separate study found that pain stimulation in horses also triggered measurable drops in parasympathetic nervous system activity, offering a physiological signal that could complement facial scoring.18PubMed Central. Parasympathetic tone activity, heart rate, and grimace scale in conscious horses of 3 breeds before, during, and after nociceptive mechanical stimulation
The broader point is that grimacing appears to be deeply conserved across mammals. It is not a culturally learned behavior or a human invention. When a mouse in pain narrows its eyes and pulls its ears back, it is doing something functionally analogous to what you do when you stub your toe, and the similarity is robust enough to build clinical tools around.
AI-Powered Grimace Detection in Medicine
The same consistency that makes grimace scales work in animals has attracted interest in automating pain detection in humans. AI systems trained on facial expression data have shown promise as tools for identifying pain-related facial movements, with the goal of creating objective, standardized pain assessments that do not depend on patient self-report.19PubMed Central. Using AI to Detect Pain through Facial Expressions: A Review This matters most for populations who cannot reliably communicate their pain: infants, patients with dementia, people under sedation, or individuals with severe intellectual disabilities.
A study evaluating AI-based facial expression analysis for detecting postoperative pain found that the technology could assist physicians in identifying severe pain, though performance still needed improvement through further training on larger datasets. The researchers described the results as a step toward a fully automated, rapid, and standardized method for measuring pain from facial expressions.20PubMed. Artificial intelligence to evaluate postoperative pain based on facial expression recognition The technology is not yet replacing clinicians, but the trajectory is clear: because grimaces follow measurable, predictable patterns, they are well suited to computational analysis.
When Grimacing Becomes a Symptom
Most grimacing is transient and tied to a stimulus. But involuntary grimacing that persists without an obvious trigger can signal a neurological problem. Tardive dyskinesia is one well-known example: a movement disorder that develops after prolonged use of dopamine-blocking medications, often antipsychotics. It produces repetitive, involuntary movements of the face, including grimacing, lip smacking, and tongue protrusion. The condition can be disabling and is sometimes permanent, which makes early identification important.21PubMed. Identification and management of tardive dyskinesia: A case series and literature review
Other conditions that can produce involuntary grimacing include dystonia (sustained muscle contractions causing twisting postures), Tourette syndrome (in which facial tics can include grimace-like movements), and certain types of seizure activity. In these cases, the grimace has detached from its usual role as a pain or effort signal and instead reflects abnormal motor circuit activity. Recognizing this distinction matters clinically: a person with tardive dyskinesia who is grimacing is not in pain, and treating them as if they were would miss the actual problem entirely.
Evolutionary Roots in Primate Displays
The grimace has deep evolutionary roots. In non-human primates, the closest analogue is the bared-teeth display, in which the lips retract to expose the teeth. Research on captive chimpanzees found that this display functions as a social signal whose meaning varies with context. Silent bared-teeth displays and vocalized bared-teeth displays appear to carry different communicative meanings, and the overall pattern aligns with predictions from models of social dominance in moderately hierarchical species. The researchers concluded that the human smile likely originated from the primate bared-teeth display, based on both the morphological similarity (the same muscles are involved) and the functional parallels.22PubMed Central. The Association Between the Bared-Teeth Display and Social Dominance in Captive Chimpanzees (Pan troglodytes)
This might seem like an odd connection. A grimace and a smile use overlapping musculature but signal very different things. The evolutionary story suggests that the bared-teeth display originally served as a submission or appeasement signal, and over time its meaning diversified. In humans, the same basic facial architecture now supports a range of expressions from warm smiling to pain grimacing, depending on which muscles fire, how intensely, and in what temporal pattern. The grimace, in this view, is not a corruption of the smile or vice versa. Both are descendants of the same ancestral display, shaped by different social and physiological pressures into expressions that serve very different purposes while sharing a common muscular toolkit.