Pain Masking in Prey Companion Animals: Integrating PAG-RVM Descending Modulation with Antipredator Evolutionary Theory

1. Introduction

Veterinary clinicians working with rabbits (Oryctolagus cuniculus), guinea pigs (Cavia porcellus), chinchillas (Chinchilla lanigera), and other small exotic mammals routinely encounter a frustrating diagnostic pattern: an animal with a condition known to be painful based on analogous procedures in dogs or cats nonetheless presents with subtle or absent behavioral signs of distress. The standard clinical explanation for this pattern invokes “prey instinct” — the observation that prey companion species conceal pain to avoid attracting predators. This explanation is not wrong, but it is incomplete. It names an evolutionary pressure without specifying the neural mechanism through which that pressure is actually implemented, and without generating any testable prediction beyond “expect less overt behavior in these species.” As a purely descriptive label, “prey instinct” cannot tell a clinician whether a given animal’s apparent calm reflects genuine comfort, moderate pain under partial suppression, or severe pain under near-complete suppression — a distinction with direct consequences for analgesic dosing decisions.

This review proposes that the descriptive gap can be closed by drawing on an independent, well-developed body of neuroscience concerning descending pain modulation — specifically, the periaqueductal gray (PAG) and its projection to the rostral ventromedial medulla (RVM), the circuit responsible for stress-induced analgesia (SIA). This circuitry has been characterized largely in laboratory rodent models of predator threat and fear conditioning, generally without explicit reference to companion-animal clinical practice; conversely, the veterinary literature on pain recognition in exotic companion mammals has developed largely independently of this circuit-level neuroscience, relying instead on behavioral catalogs, grimace scales, and clinical experience. This divide between ecological-neuroscience approaches to defensive behavior and clinical pain-recognition literature has been noted more generally in the comparative neuroethology literature (Kavaliers and Choleris, 2001). The two literatures describe, we argue, the same underlying phenomenon from different vantage points, and have not yet been systematically connected.

The aim of this review is to build that connection explicitly: to describe the PAG-RVM-SIA circuit and its evolutionary rationale (Section 3 and Section 4), to map documented clinical masking patterns in exotic companion species onto this circuit-level mechanism (Section 5), and to draw out the practical implications for pain assessment and analgesic protocol design, together with testable predictions that could validate the framework experimentally (Section 6). Throughout, the goal is not to displace existing behavioral assessment tools — grimace scales and ethograms remain the best available clinical instruments — but to reframe their known limitations, particularly the well-documented failure of facial-expression-based scoring in guinea pigs, as predictable consequences of a specific, identifiable neural mechanism rather than as an unexplained species idiosyncrasy.

2. Methods

This review synthesizes two literatures that have developed largely independently: the neuroscience of descending pain modulation and antipredator defensive circuitry, characterized predominantly in laboratory rodents, and the clinical veterinary literature on pain recognition in exotic companion mammals. Because the aim is to construct an explicit conceptual bridge between these literatures rather than to exhaustively catalog either one, this is a narrative rather than a systematic review; the search and selection process is nonetheless described here for transparency.

Literature was identified through PubMed, Google Scholar, and Scopus using combinations of the terms “periaqueductal gray”, “rostral ventromedial medulla”, “stress-induced analgesia”, “descending pain modulation”, “antipredator”, and “predator exposure”, combined separately with species and clinical terms including “guinea pig”, “rabbit”, “chinchilla”, “ferret”, “rat”, “mouse”, “grimace scale”, “pain assessment”, and “exotic companion mammal”, covering publications from January 2000 through August 2026, in English. Reference lists of retrieved articles were hand-searched for additional relevant sources.

Sources were included if they reported empirical neuroscientific data on PAG-RVM circuitry or stress-induced analgesia, predominantly in laboratory rodent models, or empirical or clinical data on pain recognition, assessment-tool validation, or documented pain-masking behavior in the exotic companion species discussed in Section 5. A small number of non-peer-reviewed clinical resources (LafeberVet, MSPCA-Angell, Veterinary Practice News) were included where they provided clinically relevant descriptive detail not otherwise available in the peer-reviewed literature; these are explicitly identified as such in the reference list and are not treated as carrying the same evidentiary weight as peer-reviewed sources. One Turkish-language clinical text (Boztepe, 2026) is cited as an independent clinical observation of the masking phenomenon and is likewise flagged throughout as having been arrived at through accumulated case experience rather than through the neuroanatomical mechanism this review proposes.

No circuit-level neuroscientific study identified in this search was conducted directly in rabbits, guinea pigs, or chinchillas under a pain-specific, as opposed to fear-specific, experimental paradigm. This gap, and its consequences for the strength of the framework’s central claims, is addressed explicitly in Section 6.3.

3. The Descending Pain Modulation Circuit

Understanding why prey companion species conceal pain requires first understanding the neural machinery that makes concealment physiologically possible. The periaqueductal gray (PAG), a midbrain structure surrounding the cerebral aqueduct, is the principal integrator of defensive responses to threat, including both predator encounters and nociceptive input. The PAG is not a homogeneous structure: its columnar subdivisions — dorsomedial, dorsolateral, lateral, and ventrolateral — each drive distinct behavioral outputs, as demonstrated in predator-exposure paradigms in guinea pigs (Paula et al., 2022). Stimulation of the dorsal columns evokes active defensive responses such as jumping, running, and flight, whereas the ventrolateral column is associated with freezing and immobility. This functional topography matters clinically: the same structure that organizes a flight response to a predator also organizes the suppression of pain expression, because both outputs serve the same survival goal of avoiding further predation risk. We develop this observation into an explicit, literature-grounded mechanistic framework below; a similar co-occurrence of antipredator neuroanatomy and analgesic behavior has also been noted, as a clinical observation rather than a mechanistic account, in the Turkish-language veterinary literature (Boztepe, 2026).

The PAG does not act in isolation. Its principal downstream target for pain modulation is the rostral ventromedial medulla (RVM), a brainstem relay containing three physiologically distinct neuronal populations: on-cells, which facilitate nociceptive transmission when active; off-cells, which inhibit it; and neutral cells, which appear uninvolved in nociceptive control (Pagliusi and Gomes, 2023). The PAG-RVM projection allows descending control over spinal nociceptive processing in both directions — facilitation and inhibition — depending on behavioral context (Pagliusi and Gomes, 2023; Shah et al., 2025). This bidirectionality is the mechanistic core of stress-induced analgesia (SIA): under acute threat, PAG activation biases RVM output toward off-cell dominance, producing a systemic, context-dependent attenuation of nociceptive sensitivity that is independent of any change in the underlying tissue pathology (Pagliusi and Gomes, 2023; Shah et al., 2025). Critically, SIA is not an absence of pain but an active suppression of its behavioral and perhaps subjective expression — the nociceptive input persists, but its downstream consequences are centrally gated. A mechanistically distinct but anatomically related RVM phenomenon, diffuse noxious inhibitory control, by which pain at one body site suppresses pain at another, is mediated by an overlapping but non-identical population of descending inhibitory RVM neurons (Ganley et al., 2026); this review’s argument concerns threat-triggered SIA specifically and should not be conflated with that separate inter-site inhibitory mechanism.

This distinction has direct clinical relevance. A prey animal undergoing PAG-RVM-mediated suppression is not “hiding” pain in a voluntary or cognitive sense; rather, its capacity to display pain-related behavior is being actively down-regulated by a conserved brainstem circuit that evolved for an entirely different, though related, purpose: reducing conspicuous behavior in the presence of threat.

4. Evolutionary Logic: Why Suppress Pain Expression?

The adaptive rationale for pain masking follows directly from predator-prey dynamics. Any behavior that increases an animal’s conspicuousness to a predator — limping, vocalizing, reduced vigilance during self-grooming, or prolonged immobility outside of a defensive freeze — increases predation risk. For a prey companion species, the marginal survival cost of displaying pain can exceed the marginal benefit of eliciting conspecific or caregiver attention. This asymmetry may be reduced, though not eliminated, in predator species: domestic cats, for instance, are themselves well documented to mask pain extensively despite being carnivores, and ferrets — discussed further in Section 5 — complicate any strict predator/prey dichotomy. The asymmetry is nonetheless likely to be strongest in obligate prey species with few defenses beyond flight, freezing, and concealment. Over evolutionary time, this asymmetry would be expected to select for neural mechanisms that decouple internal nociceptive state from external behavioral display — precisely the function that the PAG-RVM axis and SIA provide.

Empirical support for this framework comes from experimental predator-exposure paradigms. Guinea pigs confronted with a live Boa constrictor show measurable, column-specific PAG activation (via Fos protein expression) that differs between unconditioned and conditioned fear situations, with the lateral PAG column showing preferential activation after direct predator exposure (Paula et al., 2022). In rats, direct single-unit recordings during an “approach food-avoid predator” paradigm show that PAG neurons responsive to a looming robotic predator also drive downstream amygdala activity, and that optogenetic PAG stimulation is sufficient to trigger fleeing behavior — demonstrating that this circuit operates on naturalistic, ethologically relevant threat, not only on artificial conditioning stimuli (Kim et al., 2024). Both paradigms elicited acute fear responses rather than directly measuring pain-behavior suppression; the extension from fear-induced PAG activation to acute pain-display masking specifically represents an extrapolation from acute fear dynamics, not a finding either study reports directly. With that caveat, these findings support treating SIA not as a laboratory curiosity but as a conserved, ecologically deployed antipredator adaptation.

The clinical consequence of this evolutionary logic is a genuine measurement problem, not merely an observational inconvenience. Guinea pigs illustrate this starkly: despite the broad success of facial grimace scales across numerous mammalian species (Cohen and Beths, 2020; Reijgwart et al., 2017), validation studies converge on the conclusion that grimace-based facial expression scoring is unreliable in guinea pigs specifically (Benedetti et al., 2024; Cohen and Beths, 2020; Mota-Rojas et al., 2025) — the species-typical facial anatomy and behavior appear to resist the very signal that grimace scales are designed to detect elsewhere. Guinea pig-specific composite pain scales have instead been built from postural, functional, physiological, and behavioral descriptors rather than facial action units (Benedetti et al., 2024; Dunbar et al., 2016), and clinicians report persistent uncertainty about baseline pain recognition and analgesic dosing in this species (Evans and Benato, 2025). Read through the framework proposed here, this pattern may plausibly reflect more than a failure of grimace-scale methodology in isolation; we hypothesize that selection pressure for pain-display suppression may have acted with disproportionate strength on the facial musculature that grimace scales rely upon in this species — though this evolutionary attribution is this review’s hypothesis rather than a conclusion drawn by the cited validation studies themselves, which report the empirical failure without proposing this mechanism. A circuit-informed perspective therefore reframes the clinical problem as a hypothesis worth testing: the goal is not simply to look harder for behavioral signs, but to consider that the signs most likely to be suppressed may be exactly the most visible ones.

5. Clinical Manifestation in Exotic Companion Species

The circuit-level framework predicts graded, not binary, differences in pain expression across exotic companion species, and the clinical literature broadly supports this. Rabbits and rodents — obligate prey species with few natural defenses beyond flight, freezing, and concealment — show the most consistent masking: reduced overt vocalization, altered but subtle postural changes, and a tendency toward immobility that can be misread as calm rather than as guarded behavior (Mayer, 2007; LafeberVet, n.d.; Miller and Leach, 2023). Guinea pigs are frequently singled out in the clinical literature as a particularly difficult case, becoming markedly still and suppressing normal behaviors such as foraging and social interaction rather than displaying overt distress (Mayer, 2007; MSPCA-Angell, 2026). Owners frequently misinterpret this stillness as calmness rather than pain-induced guarding behavior, and vocalization is typically reported to appear only once pain is already severe (MSPCA-Angell, 2026).

Ferrets (Mustela putorius furo) complicate a simple predator-versus-prey dichotomy. Although ferrets are small carnivores and are frequently cited in the veterinary literature as showing more overt pain behavior than rabbits or rodents (Reijgwart et al., 2017; Schoemaker, 2023), ferrets are themselves subject to predation by larger carnivores and raptors in the wild, and clinical sources note that ferrets — like other small mustelids — still frequently conceal pain and rely on indirect behavioral and physiologic parameters for assessment (Schoemaker, 2023; MSPCA-Angell, 2026). This parallels the point raised in Section 4 regarding domestic cats: taxonomic predator status does not straightforwardly predict the degree of pain masking. This suggests that the relevant selective pressure is not strictly “predator versus prey” taxonomic status but relative predation risk across an animal’s ecological niche — a distinction that refines, rather than undermines, the evolutionary framework proposed here.

A second, methodologically important confound is the observer effect. Clinical sources note that the mere presence of a human observer can suppress pain-related behavior in rodents and rabbits (Mayer, 2007; Veterinary Practice News, 2024), consistent with the possibility that a nearby human is itself processed, at least partly, through the same threat-detection circuitry that responds to a natural predator. This means that the standard clinical encounter — approaching, restraining, and directly observing the animal — may itself trigger the very PAG-RVM-mediated suppression the clinician is trying to see through, compounding the underlying biological masking with an assessment-induced artifact.

Existing behavioral pain-assessment tools partially address this problem but do not resolve it at the mechanistic level. Grimace scales, validated across numerous mammalian species, quantify facial action units such as orbital tightening and ear position as pain indicators (Cohen and Beths, 2020; Reijgwart et al., 2017), and composite behavioral ethograms combine postural, functional, and physiological descriptors into a single score (Dunbar et al., 2016). Both approaches, however, still depend on the animal producing an externally visible signal — precisely the output that SIA is positioned to suppress. Guinea pigs again illustrate the limit case: validation work converges on the conclusion that grimace-based scoring is unreliable in this species specifically (Benedetti et al., 2024; Cohen and Beths, 2020; Mota-Rojas et al., 2025), and clinicians have instead had to construct composite scales from non-facial descriptors (Benedetti et al., 2024; Dunbar et al., 2016), alongside candid acknowledgment that current analgesic dosing in guinea pigs is frequently extrapolated from other species in the absence of validated values (Evans and Benato, 2025). Read against the framework proposed here, this may not simply be a gap in tool development; it may be, at least in part, the outcome of applying a facial-expression-dependent instrument to a species in which facial suppression is, we propose, under particularly strong selective pressure — a hypothesis this review raises rather than one established by the cited validation studies.

Comparable clinical patterns have also been documented independently in the Turkish-language veterinary literature (Boztepe, 2026), arrived at through accumulated case observation rather than through the neuroanatomical mechanism proposed here.

6. Discussion

6.1. Clinical Implications of the Framework

The clinical implication of this framework is not merely that pain is “hard to see” in prey companion species, but that existing behavioral tools may systematically underestimate the true magnitude of pain even when they do detect its presence — if, as we propose, SIA operates as a graded, context-dependent gain control on nociceptive expression rather than an all-or-nothing switch. Under that proposal, a low grimace or ethogram score in a rabbit, guinea pig, or small rodent would be better interpreted as a floor on likely pain severity rather than a ceiling, though this interpretation itself awaits direct experimental test (Section 6.2). This has potential consequences for multimodal analgesia protocols in exotic companion practice: dosing decisions calibrated strictly to visible behavioral change risk under-treatment in precisely the species where suppression may be evolutionarily strongest. This is broadly consistent with existing clinical guidance recommending that analgesia in rabbits, rodents, and guinea pigs be provided as indicated by the underlying condition or procedure rather than withheld pending unambiguous behavioral confirmation (LafeberVet, n.d.; Love et al., 2014); we would emphasize that clinicians should weigh this guidance more heavily, not less, when a behavioral score is low but a painful condition or procedure is present, rather than interpreting a low score as reassurance of comfort.

6.2. Testable Predictions

The framework also generates testable predictions that could move this from a conceptual synthesis to an empirically validated model. If PAG-RVM-mediated SIA is responsible for behavioral masking, then experimentally manipulating perceived predation risk — for example, comparing grimace or ethogram scores obtained with a visible human observer present versus scores obtained via remote video monitoring (Mayer, 2007; Veterinary Practice News, 2024), or comparing scores collected in the presence versus absence of simulated predator olfactory or acoustic cues — should produce measurable, reversible shifts in apparent pain severity, independent of any change in the underlying nociceptive stimulus. A demonstration of such context-dependent variability, particularly in guinea pigs where grimace-based tools have already failed for reasons that remain only descriptively characterized (Benedetti et al., 2024; Cohen and Beths, 2020; Mota-Rojas et al., 2025), would provide direct evidence that the mechanism proposed here, rather than simple species-typical stoicism, underlies the clinical assessment problem. Until such direct tests are conducted, the framework should be regarded as a plausible, literature-consistent hypothesis rather than an established mechanism.

6.3. Limitations

This synthesis has several limitations that bear directly on the strength of its conclusions. First, the neuroscientific evidence for the PAG-RVM-SIA circuit described in Section 3 comes almost entirely from laboratory rats and mice; direct circuit-level evidence in the target clinical species discussed in Section 5 — rabbits, guinea pigs, and chinchillas — is limited to a single guinea pig study, which measured predator-induced fear responses rather than pain-behavior suppression directly (Paula et al., 2022). The extension from that fear paradigm to pain masking specifically is an inference drawn by this review, not a finding reported by the cited study. Second, the evolutionary attribution offered for the guinea pig grimace-scale failure (Section 4) is this review’s own hypothesis; the cited validation studies report the empirical failure without endorsing this or any other causal mechanism. Third, this review relies substantially on narrative synthesis of clinical sources rather than a systematic search, and the strength of evidence therefore varies considerably across the specific claims summarized in Section 5. Readers should treat the framework as a set of literature-grounded hypotheses that motivate the experimental tests proposed in Section 6.2, rather than as an established causal account.

6.4. Conclusions

The tendency of prey companion species to mask overt pain is frequently described in the clinical literature, including in Turkish-language sources (Boztepe, 2026), in terms of “prey instinct,” a phrase that names the phenomenon without explaining it. Pain masking in rabbits, rodents, guinea pigs, and related exotic companion species may be best explained by the active operation of a conserved descending pain-modulation circuit — the PAG-RVM axis and the stress-induced analgesia it produces — that evolved to reduce conspicuous behavior under predation risk, and may now operate continuously in captive clinical settings that share none of the original selective context. As detailed in Section 6.3, this remains a literature-grounded hypothesis rather than a directly demonstrated mechanism. Reframing “hard to read” as “possibly actively suppressed” nonetheless carries a practical implication worth clinical attention: clinicians may reasonably treat the absence of overt pain behavior in these species with somewhat more caution than the behavioral score alone would suggest, particularly when a painful condition or procedure is independently known to be present, rather than treating a low score as straightforward reassurance of comfort. Future work developing pain-assessment tools for prey companion species — particularly for guinea pigs, where existing facial-expression-based instruments have already been shown to fail — would benefit from testing the context-dependent, circuit-informed predictions proposed in Section 6.2, alongside continued refinement of descriptive behavioral catalogs.

Artificial Intelligence Use Statement: An AI-assisted research tool was used to support literature searches and verify bibliographic information. All information provided by the AI tool was independently reviewed by the author. The selection and accuracy of sources, as well as all scientific content, interpretations, evaluations, and conclusions presented in this work, remain the sole responsibility of the author.