Is "Allostasis" The Brain’s Essential Function?

A paper just published in Nature Human Behaviour creates some big claims about the brain. It’s called Evidence for a large-scale brain complement ancillary allostasis and interoception in humans, but how much is justification and how much is speculation?

The authors, Ian R. Kleckner and colleagues of Northeastern University, disagree that a core duty of the brain is allostasis, which they conclude as the routine by which the brain “efficiently maintains appetite law in the body”. Allostasis entails “anticipating the body’s appetite needs [and] scheming to meet those needs before they arise.” Kleckner et al. indicate to “physical movements to cold the body’s heat before it gets too hot” as one instance of allostasis.

A judgment closely associated to allostasis is interoception, the routine by which the brain receives information about the body’s inner state from feeling nerves inside the body.


While allostasis and interoception are not new ideas, Kleckner et al. go into novel domain by arguing that allostasis is a elemental routine in the brain. All kinds of ‘higher’ brain functions, such as memory, cognition, and decision-making, “rely on” allostasis, according to this paper. The authors contend that the justification for this lies in the fact that widespread neural circuits are functionally connected with brain areas concerned in allostasis:

Our work demonstrates a singular brain complement that supports not just allostasis but also a far-reaching operation of psychological phenomena (emotions, memory, decision-making, pain) that can all be explained by their faith on allostasis… a rarely connected organic garb for integrating information opposite the brain, with interoceptive and allostatic information at its core.

It’s a big idea, and we consider we need those in neuroscience, but in my view, Kleckner et al. yield little justification for the existence of a widespread “allostasis-interoception” complement in the brain, or for the suspicion that allostasis underlies other cognitive functions.

The core of Kleckner et al.’s justification is the suspicion of a structurally and functionally companion set of brain regions. These regions embody the dorsal midst and dorsal posterior insula (dmIns/dpIns), the ventral maiden insula (vaIns), the subgenual and pregenual maiden cingulate cortex ( sgACC and pACC) and maiden midcingulate (aMCC also famous as dorsal ACC). Many of these regions are also connected with subcortical structures, including the hypothalamus and several brain branch nuclei.

This connectivity is the basement for their centerpiece Figure 2, formed on anatomical studies (tract-tracing) of constructional connectors in primates, and resting state fMRI studies in humans:


My regard here is not with the figure, but with the caption. Granted that these regions are all interconnected, since should we tag them as “the one allostasis/interoceptive system”? The justification is that some of the brain areas in the network have obvious roles in interoception (e.g. the dmIns/dpIns) and in allostasis (e.g. tools of the hypothalamus and brain branch nuclei.) But we find this proof questionable.

Just since some of the brain areas in a network have a sold function, doesn’t meant the whole network is formed around that function. Take that proof to its extreme, and you could contend that the whole brain (which is all interconnected) has any duty you choose. Perhaps my brain is a network for determining my left thumb. For another thing, the justification is formed on reverse inference.

I’m further not certain about Kleckner et al.’s justification that two much-studied organic brain networks, the default mode network (DMN) and the salience network, are in fact aspects of the allostasis system. This explain is formed on the fact that their fMRI connectivity patterns overlie with that of the ‘allostais network’, but couldn’t this same proof be practical in reverse? Couldn’t we equally good contend that the ‘allostasis network’ is actually an aspect of salience network?

Where’s the justification that any of this connectivity is related to tangible discernment or behaviour? Kleckner et al. report that fMRI connectivity between two allostasis network regions, the aMCC and dpIns, predicts the grade to which people’s biased feelings of arousal conform to their design skin conductance response (SCRs) while examination emotionally evocative photos (n=41).

As the dpIns is suspicion to be an interoceptive region, this outcome is unchanging with the suspicion that the aMCC is concerned in branch interoception into unwavering romantic experience. But that’s a flattering singular result; it’s just two brain regions, and says little about the broader “unified allostasis/interoceptive system” which is the theme of the paper.

So altogether we don’t consider there’s that much justification here, positively not adequate to support Kleckner et al.’s idea that allostasis is the very purpose of all brains:

Perhaps many importantly, the allostatic–interoceptive complement has been shown to have a role in a far-reaching operation of psychological phenomena, suggesting that allostasis and interoception are elemental facilities of the shaken system… justification suggests that smarts did not develop for rationality, complacency or accurate perception… the brain developed to umpire allostasis.

There’s also no justification presented in this paper which bears on the doubt of how the brain regulates allostasis. Kleckner et al. impute to the judgment of predictive coding, and they impute to their altogether supposition as the EPIC indication (Embodied Predictive Interoception Coding). However, zero in this paper examined predictive coding. The authors contend that some-more studies to test EPIC are on the way.

Overall, this paper feels like two articles grafted together: an initial paper (the n=41 study) with interesting, but sincerely singular results; and a suppositional examination paper about an allostatic-interoceptive speculation of brain function.

As hypothesis, the ideas in this paper are positively interesting. But to contend that the information in this paper is justification for the presented supposition is a stretch, in my view.


Ian R. Kleckner and the paper’s comparison author, Lisa Feldman Barrett, pleasantly commented on a breeze of this post. Here’s what they said:

We do consider that you are maybe disagreement one critical indicate of the paper when you write:

“Just since some of the brain areas in a network have a sold function, doesn’t meant the whole network is formed around that function”

Neuroscientists have a robe of fixing a network/system in line with their own interests.  The default mode and salience networks have some-more aliases than Sherlock Holmes. In this paper, we prominence the domain ubiquitous inlet of these networks (see the Figure 5) but we also show that they interface with the physique in critical ways that are overlooked. This offers new insights and opens new questions for understanding, say, the role of interoception in memory, the role of concepts to be means to minister to allostasis, and so on. Some of these topics are discussed in Lisa’s book, How Emotions Are Made, which you can review about here.

We never claimed that default mode/salience network functions are based around any singular function. When a sold collection of neurons fires in synchrony, definition they are companion and portion a one function, this does not meant that the network serves only that function, or even primarily that function.  The whole indicate of the paper is to show that whatever else these networks are doing in a given instance (memory, judgment formation, perception, etc.) they are also controlling allostasis and representing interoception. The paper builds off other papers we have created that make a identical indicate – see 12. 3.

The paper is not claiming that allostasis and interoception are the only functions of this network, or even the primary functions of this network. Such approaches to localization are unsuitable with the bargain of neural function. Default mode and salience networks are domain ubiquitous in their function. And allostasis/interoception is partial of the repertoire of those networks. It is not that these neurons infrequently do allostasis and infrequently do memory (or another psychological function). The indicate is that while they minister to memory, to representing concepts, to language, and so on, they seem to simultaneously and always minister to allostasis (your autonomic shaken system, defence system, and neuroendocrine systems are never off).

On pages 6 and 7 of the paper, we write The fact that default mode and salience networks are parallel controlling and representing the inner milieu, while they are customarily intent in a far-reaching operation of tasks travelling cognitive, perceptual and tension domains, all of which engage value-based decision-making and actions, suggests a provocative supposition for future research: whatever other psychological functions the default mode and salience networks are behaving during any given brain state, they are concurrently progressing or attempting to revive allostasis and are integrating feeling representations of the inner feel with the rest of the brain….whatever else your brain is doing — thinking, feeling, perceiving, moving — it is also controlling your autonomic shaken system, your defence complement and your endocrine system.”

Allostatic changes always have feeling (interoceptive) consequences, and therefore interoception deduction while you are remembering, making decisions, noticing the world, and so on.  Again, from page 7 of the paper, Perhaps the many profitable aspect of the commentary is in moving over normal domain-specific or ‘modular’ views of brain structure/function relationships”.

ResearchBlogging.orgKleckner, I., Zhang, J., Touroutoglou, A., Chanes, L., Xia, C., Simmons, W., Quigley, K., Dickerson, B., Feldman Barrett, L. (2017). Evidence for a large-scale brain complement ancillary allostasis and interoception in humans Nature Human Behaviour, 1 DOI: 10.1038/s41562-017-0069

  • Efficient systems don’t survive. Robust systems survive. Robustness need redundancies and are therefore not efficient.

    So blah blah blah yak yak yak, authors. Build a indication that is predictive over a vast operation of complement parameters and test those predictions. Contemplating one’s navel is not science.

  • Neuroskeptic: You have a broken couple in the authors’ reply. The first couple is ostensible to be, but you’ve lost the “https://” in the HTML.

    Feel free to undo this criticism after you’ve bound the link.

    • Oops, thanks! Fixed.

  • Is the allostasis perspective falsifiable?

Check Also

Spotify fraud hacked site’s pay-out complement to make millions

An different operator, formed in Bulgaria, uploaded music clips for which they owned the copyright …