Hooks et al. note that approximately half of the studies surveyed explore relations between the “neuroendocrine stress axis” and microbiota. They criticize these studies for translating too easily between animal models and humans, and they argue that the measurement of stress via hormones such as corticosterone is potentially confounded by other factors. These criticisms are sound, but with stress as such a central concept to parts of microbiota-gut-brain research (and beyond), it is central to expand on them, because these criticisms illuminate the difficulties of exploring neurobiology vis-à-vis psychological states and behaviors such as “stress.”

Across the papers surveyed by Hooks et al., stress is first deployed with a considerable methodological and conceptual diversity. The papers report both “stress” and stress that is, for example, “chronic psychological” (Ait-Belgnaoui et al. Reference Ait-Belgnaoui, Colom, Braniste, Ramalho, Marrot, Cartier, Houdeau, Theodorou and Tompkins2014), “acute” (Crumeyrolle-Arias et al. Reference Crumeyrolle-Arias, Jaglin, Bruneau, Vancassel, Cardona, Daugé, Naudon and Rabot2014), “acute psychological” (Ait-Belgnaoui et al. Reference Ait-Belgnaoui, Durand, Cartier, Chaumaz, Eutamene, Ferrier, Houdeau, Fioramonti, Bueno and Theodorou2012), or the result of “exposure” to a “social stressor” (Bailey et al. Reference Bailey, Dowd, Galley, Hufnagle, Allen and Lyte2011). Presumably, these qualifiers specify the type of stress explored, yet they rely on no coherent theoretical paradigm and employ very different methods of inducing “stress.”

For example, one paper (Ait-Belgnaoui et al. Reference Ait-Belgnaoui, Durand, Cartier, Chaumaz, Eutamene, Ferrier, Houdeau, Fioramonti, Bueno and Theodorou2012), uses the notion of “acute psychological stress” in their paper's title, a term that surprisingly shifts into “partial restraint stress” in the paper itself. This was induced by wrapping rats in constricting tape to limit (but not prevent) their movements for 2 hours (Ait-Belgnaoui et al. Reference Ait-Belgnaoui, Durand, Cartier, Chaumaz, Eutamene, Ferrier, Houdeau, Fioramonti, Bueno and Theodorou2012, p. 1886). In another paper by the same team (Ait-Belgnaoui et al. Reference Ait-Belgnaoui, Colom, Braniste, Ramalho, Marrot, Cartier, Houdeau, Theodorou and Tompkins2014), “chronic psychological” stress is induced by placing rats on a small platform in the center of a tank filled with water, for 1 hour at a time, 4 days in a row. It is unclear, however, how qualifiers such as “psychological,” “acute,” or “chronic” actually connect to these methods. We may note, for example, that “acute stress” was induced for 2 hours, whereas “chronic stress” was induced for half that time (but repeated over 4 days). Other examples of “stress”-inducing methods include restraining mice in tubes for 1 hour (Sudo et al. Reference Sudo, Chida, Aiba, Sonoda, Oyama, Yu, Kubo and Koga2004) or putting an aggressive mouse into another mouse's cage (Bailey et al. Reference Bailey, Dowd, Galley, Hufnagle, Allen and Lyte2011), seen as a “social” stressor. One wonders why the confrontation between two mice is “social,” whereas the handling of rodents by scientists does not classify as such (especially when research has shown that olfactory exposure to men increases the stress response in mice [Sorge et al. Reference Sorge, Martin, Isbester, Sotocinal, Rosen, Tuttle, Wieskopf, Acland, Dokova, Kadoura, Leger, Mapplebeck, McPhail, Delaney, Wigerblad, Schumann, Quinn, Frasnelli, Svensson, Sternberg and Mogil2014]).

More generally, it is by no means clear if these types of stress are comparable to each other: Being physically constrained in a tube is different from being slightly restricted in movement (yet both are classified as “acute” forms of stress), which is different from being exposed to water for days in a row. Further, it is not entirely clear if whatever the rats go through is in any way comparable to stress in humans. As Hooks et al. note, most researchers attempt to avoid anthropomorphizing their rodents, hence using terms such as anxiety-like. However, no-one uses stress-like. It is simply assumed that the rats are experiencing stress comparable to that of human beings (see also Rose & Abi-Rached Reference Rose and Abi-Rached2013, p. 97). Built into rodent stress research is thus a tacit assumption that there is a direct line from the stress of a rodent to the stress of a human being. Besides such translational problems between rodents and humans, there is also a circular reasoning employed, with the a priori assumption that the experimental paradigm (e.g., restraining mice) is stress inducing, and, accordingly, a following neuroendocrine response must thus reflect the induction of stress. This common line of thinking, however, contradicts another common assumption within stress research – namely, that stress depends on the organism's perception and appraisal of the experimental stimuli as a stressor (e.g., McEwen Reference McEwen1998; McEwen & Seeman Reference McEwen and Seeman2006). Some researchers have suggested that the term stress should therefore be limited to conditions wherein organisms perceive a serious, unpredictable, and uncontrollable threat to their health (Koolhaas et al. Reference Koolhaas, Bartolomucci, Buwalda, de Boer, Flügge, Korte, Meerlo, Murison, Olivier, Palanza, Richter-Levin, Sgoifo, Steimer, Stiedl, van Dijk, Wöhr and Fuchs2011, p. 1292). Thus, tacking terms such as psychological or acute onto the “stress-inducing” experiments such as those noted previously only adds to a considerable conceptual and methodological confusion.

This also means that the measurement of hormones as indicative of “stress” is not just plagued by confounders, or circularity, but more fundamentally with the difficult problem of what it means when a rodent (or a human) experiences an increase in the activity of the hypothalamic-pituitary-adrenal (HPA) axis.

In humans, for example, one study shows how secretions of cortisol vary in experienced versus novice skydivers (the former's response is smaller and trails off more quickly), whereas their experiences of anxiety, before the jump, are similar (Meyer et al. Reference Meyer, Lee, Böttger, Leonbacher, Allison and Shirtcliff2015). If the association between subjective experience and hormonal response is equivocal, then the assumption that a hormonal response to a stress-test paradigm necessarily means that “psychological stress” has been induced is flawed.

There are, in sum, questions remaining over how we induce “stress” in animal models, of how we should analyze and interpret animal behavior, and, crucially, how this translates to the complexities of human experience and psychiatric disorders. These difficult questions go beyond microbiota-gut-brain research, relating more broadly to how we think about the “psychological” vis-à-vis the biological. Often, one will marvel at the complexity that is afforded to exploring, describing, and modeling biological pathways and mechanisms (in this case, microbiomes, guts, and brains) – yet it is peculiar that so persistently, much less effort is made to explore the psychological and behavioral beyond the most basic level of assumption and description. What is necessary are not declarations of how microbiomes “challenge our concept of self” (e.g., Rees et al. Reference Rees, Bosch and Douglas2018) but much more careful and detailed explorations of microbiomes, guts, and brains in relation to careful operationalizations of behavior and the psychological.