the metabolic cost of uncertainty

September 17, 2024 6:03 AM

"But what if brains don’t have dedicated circuits for fighting and fleeing? People clearly experience threats, but is threat detection really a primary mode of the brain with its own neural circuitry? A body of recent evidence from my lab and elsewhere suggests that we don’t go through life constantly detecting threats and reacting with flight-or-fight circuits. Rather brains operate mainly by prediction, not reaction. All brains constantly anticipate the needs of the body and attempt to meet those needs before they arise. They seek to reduce uncertainty to survive and thrive in circumstances that are only partially predictable." The ‘Fight or Flight’ Idea Misses the Beauty of What the Brain Really Does -- an essay in SciAm by Lisa Feldman Barrett (author of the amazing How Emotions Are Made).
posted by mittens (36 comments total)

Thus: ANXIETY.
posted by seanmpuckett at 6:16 AM on September 17 [5 favorites]

Isn't prediction the point behind "intrusive thoughts" as well? I don't see anything in the article about it though.
posted by techSupp0rt at 6:35 AM on September 17

Any article claiming to know what the brain "really does" sort of flies in the face of Gödel's incompleteness theorem. But of course that's just the headline, designed to get clicks. I like how it explores the interconnected complexity of the brain vs. the simplistic "emotions are in this part, rationality is in this part" approach. But it also attempts to debunk the idea that we are always in fight-or-flight mode, which seems like a straw man. My understanding of fight-or-flight was that it is a mechanism for making a quick decision when a threat appears in a low-information environment - essentially, something that could not be predicted - and that quick decision is critical to (perceived) survival. Smelling a predator is a lot different than turning a corner and being face to face with a predator that you did not smell.

I've certainly experienced something I call the "NOPE response" which caused me to actively and fully disengage from whatever I was doing. Once was when my wife paid to get me the behind the scenes tour at our zoo with the rhinoceros, and as I was brushing it, it flinched and moved back a few feet in a matter of what felt like microseconds. Suddenly fully aware of its size, power and unpredictability, I backed as far away as I could and just checked out for the rest of the tour, much to the dismay of my wife. The other time was when I was (inadvisably) mountain biking down Schwietzer Mountain in Idaho with a friend. The first part of the trail was a bunch of steep, sandy switchbacks, and much to my dismay I discovered that the brakes, however gently or forcefully applied, did not stop the bike but merely made it skid along the dirt. I made it around two corners before the "YOU ARE GOING TO DIE" klaxon starting ringing at full volume in my head and I got off, put the bike on my shoulders and walked right back up the mountain. It was mortifying to ride the ski lift back down with the bike and have everyone ask me what happened, but according to my wife the other people that went down the mountain were ghost-faced and seemed to have regretted their decision.

I will say that the one time I encountered a bear in the wild - smallish, black, in Harriman State Park - I did not fight or flee. I just said "Bear! Bear!" and stood still. I recall feeling nothing at all, just a still calm. One of my compatriots immediately turned around and started walking the other way, no doubt having a "NOPE" response. The bear looked back at us and then ran away. We stood there for another two minutes before carefully (and noisily) making our way along the trail.
posted by grumpybear69 at 6:52 AM on September 17 [7 favorites]

Any article claiming to know what the brain "really does" sort of flies in the face of Gödel's incompleteness theorem.

I don't think Gödel really practically constrains what we can know about brains, and anyway brains aren't formal systems so it doesn't directly apply
posted by BungaDunga at 7:25 AM on September 17 [19 favorites]

Any article claiming to know what the brain "really does" sort of flies in the face of Gödel's incompleteness theorem.

What Gödel's incompleteness theorem demonstrates is that any formal system powerful enough to implement arithmetic cannot achieve both consistency (the property that no two statements provable within the system are contradictory) and completeness (the property that all valid statements expressible within the system are also provable). It does this by providing a general method for constructing, within any such system, provable statements whose validity is by inspection undecidable.

Applying it to brains is metaphorical at best and horribly misleading at worst. In any case, it is glaringly obvious that brains do not and cannot exhibit anything even vaguely resembling completeness, despite the claims of assorted and usually contradictory ideologies to do so.

It's not at all obvious to me how a report on the results of experiments that examine the operation of brains under well-defined study conditions "flies in the face" of GIT even in the most lax metaphorical sense I'm willing to entertain. Care to elucidate?

If what you're actually alluding to is the old saw that if we were simple enough to understand then we'd be too simple to do the understanding, I think you'll find that a disproof by construction in the form of an engineered artificial general intelligence is only about twenty years away.

I'm quite convinced of the correctness of that prediction, given the confidence with which it's been asserted by people far more qualified than me for at least as long as I've been alive.
posted by flabdablet at 7:29 AM on September 17 [7 favorites]

It may be worth it for some to read the article in the FPP. Dr Feldman Barrett’s extensive body of work on the human brain and many recent explorations and applications of neuroscience may have more practical use than the work of Gödel.
posted by armoir from antproof case at 7:49 AM on September 17 [13 favorites]

It's not at all obvious to me how a report on the results of experiments that examine the operation of brains under well-defined study conditions "flies in the face" of GIT even in the most lax metaphorical sense I'm willing to entertain. Care to elucidate?

I believe that we are fundamentally incapable of completely understanding ourselves, QED. But my main issue is with anyone that says "this is how the brain really works!" because it has been said a billion times and has always proven to be totally wrong at worst and incomplete at best. And, like, mouse brains are not human brains. And even if AGI comes to fruition in 20 years, it is almost certainly not going to work like our brain does, even if it convincingly converses and writes limericks and solves metaphysical quandaries. So that is not de facto proof of us understanding ourselves, only of the ability to create a simulacrum based on what we know.

And I did read the article, which is why the rest of my comment was not about math theory. But, as I was drawn to the headline like a fly to a lamp, so have the other commenters to the first line of my comment. Mods, feel free to delete my comment and this follow-up if it is derailing the conversation from a more substantive discussion of the article itself.
posted by grumpybear69 at 7:54 AM on September 17 [3 favorites]

The brain is an informal system which can usually do arithmetic.
posted by Nancy Lebovitz at 8:13 AM on September 17 [2 favorites]

fMRI + mice + just-so human evolution narratives == all the components of bunk. However, this article does not seem to me to be making absurdly wild claims, but it's swimming in all the marks of bad brain science.
posted by Rhomboid at 9:06 AM on September 17

Ruben Laukkonen gave a fascinating interview about the predictive processing of the brain in regards to how it generates consciousness and a sense of time. Tracks with my meditation experience !
posted by St. Peepsburg at 9:24 AM on September 17 [1 favorite]

funny thing about bad brain science is that it purposefully uses the trappings of good brain science to make itself seem credible. fmri and mice studies are only signs of bad brain science because bad brain scientists have used those approaches to make stupid claims. they are perfectly fine tools in the hands of good brain scientists.

i also missed where the author was promoting a 'just-so' evolution narrative. to the extent the author talks about evolution, it's to debunk one of those evolution narratives.

the view that a major role of the brain involves prediction is pretty comfortably in the mainstream of neuroscience. part of predicting involves managing uncertainty, and there are many regions in cortex whose function seems to be in whole or in part related to detecting, processing and compensating for various sources of uncertainty.

uncertainty is probably not the entire story, but it is a good and interesting framework for interpreting behavior. the biggest drawback it has is that it is very non-intuitive to describe brain function in that way, especially when intelligent lay people are raised on diets of popsci 'flight-or-flight' 'left-brain/right-brain' junk
posted by logicpunk at 9:27 AM on September 17 [12 favorites]

Even if 'bad science,' I can't possibly imagine the metabolic stress of a child or an adult trying to function in an abuse situation.
posted by BlueHorse at 9:32 AM on September 17 [2 favorites]

The brain is very, very good at discovering patterns and adapting behavior to those patterns. In high school, before the advent of caller ID, I became extremely good at guessing who was calling at any given moment. To the point where I would just pick up the phone and say "Hello, [name]!" to their absolute astonishment. And I'm certain that it was because my brain picked up on when certain people called, who I had talked to recently that was likely to contact me, etc. It wasn't a foolproof algorithm, but it was successful enough to be noteworthy.

I think a lot about how time seemed really slow as a kid and is a lot faster as an adult, and I wonder how much of that is the mental effort involved in mapping out the probability trees of life and cataloging all of the new data, both of which efforts grow smaller and smaller as we age.

What's interesting is that sometimes the brain rewards or prioritizes uncertainty. I absolutely love a day where I've no idea what's going to happen and I go on adventures with people. To be freed from the mental weight of decision-making, to just be a ride-along.
posted by grumpybear69 at 9:56 AM on September 17 [2 favorites]

Dr. Barrett is author of over 250 scholarly papers on psychology and neuroscience. Here is a small selection.

And

Six Northeastern Professors Named to 2019 List of ‘Highly Cited Researchers’ Around the Globe

posted by jerome powell buys his sweatbands in bulk only at 10:00 AM on September 17 [2 favorites]

Ever since coming across the idea of prediction, it has made a sort of sense of weird phenomena like pet-pareidolia; seeing the dog, maybe even starting to say hello to the dog, then realizing it's not the dog, just a backpack on the floor I caught out of the corner of my eye. How do you know how far to move your mouse, to put your cursor in the comment field? You don't sit around and think it through, you don't do math about it, clearly...but the idea that you're making some very minor, low-effort prediction on where your hand should be, has a basic sort of sense to it. Just always making little models of the world and where your body is in it and what will happen next.

And it's understandable that, if that's the way we process the world, then larger uncertainties, unresolvable ones, require more effort. The crazy back-and-forth of a mind under stress, picturing everything that could go wrong--and trying to picture what could go right--I mean, it's no wonder it's exhausting (and often seems to require a lot of calories? is this why we stress-eat?).

By the way, the author is really aware of bad science in the field, and talks about it in the above-noted book, in an accessible and entertaining way. Highly recommended if you're interested in brains and what they do.
posted by mittens at 10:21 AM on September 17 [2 favorites]

Metabolic expense, if it drags on for long enough, can feel unpleasant. When your brain attempts to learn in uncertainty, it releases chemicals that increase your arousal level, and you may feel worked up and agitated....
Uncertainty is a normal condition of life, but these days, with social media and round-the-clock news coverage, it sometimes bombards us. At every moment, there’s a crisis somewhere in the world: war, political chaos, climate-induced fires and floods and school shootings, not to mention the occasional pandemic. Too much uncertainty is metabolically draining and can leave you feeling distraught and worn out. But these feelings don’t emerge from mythical, overtaxed fight-or-flight circuits. They may just mean, in an ever changing and only partly predictable world, that you’re doing something really hard.

Yeah, that's how I feel these days.

I went to bed before 8 p.m. last night-- and I'm a night owl, I have to be feeling pretty shitty to be able to do that--and it took 10 hours to try to get 8 hours of sleep because I woke up for about 2 hours in the middle of that. Admittedly I've been on an exhausting schedule for weeks and presumably I'm still recovering from that, but I feel tired every day even after 8 hours, and 8 hours is hard to come by in the first place, much less consistently 8 hours every night.

I'm always waiting around for some other anvil to land on me, even if I have a lot less potential anvils than I did before.
posted by jenfullmoon at 10:33 AM on September 17 [5 favorites]

I really enjoy novelty, such as trying a new restaurant. This article is making me wonder if there's a stress-cost to novelty. Would my overall stress level go down if I stuck to my three favorite restaurants?
posted by tofu_crouton at 10:41 AM on September 17

Like maybe I should increase predictably where it's easy to do so since there are so many places that I can't.
posted by tofu_crouton at 10:42 AM on September 17 [3 favorites]

MFK Fisher described the perfect seducible state as being totally unsure what would happen but totally secure that it would be pleasant. If you’re good enough at choosing and enjoying new restaurants, tofu_crouton, maybe you’re giving yourself that?
posted by clew at 10:55 AM on September 17 [12 favorites]

Would my overall stress level go down if I stuck to my three favorite restaurants?

Counterpoint: seeking novelty and one response to exceptionally missed predictions may give rise to the very essence of humor. What if there is evolutionary selection value (with all the caveats that it's not just about procreation) in both seeking novelty and finding it? Both As an adaptation against constant anxiety and the reduction in evolutionary viability that might result in a fear of any action or the way senses of humor align among individuals with complementary or compatible tolerances for anxiety and unpredictability.
posted by Lenie Clarke at 11:12 AM on September 17 [1 favorite]

tofu_crouton, I think it is pointing the arrow the wrong way to look at your own preferences through the lens of this article and then assuming that the broad conclusion of "uncertainty == stress" applies to you. The tolerance people have for ambiguity varies widely, so what may stress one person out may make another person happy. If you like going to new restaurants, keep doing it!
posted by grumpybear69 at 11:14 AM on September 17 [1 favorite]

Excellent essay. Feldman Barrett does an excellent job explaining and making the case for the new picture of neural processing around prediction and uncertainty that many researchers, including her, have been carefully discovering over the past 20 years or so. She references and debunks a lot of bad old ideas about the brain that have stuck around in the popular thinking despite being clearly wrong: I would take seriously every criticism and correction that she makes in this essay.
posted by biogeo at 11:38 AM on September 17 [6 favorites]

I've always found the 'fight or flight' hypothesis sort of weird if you pay attention to most other animals - of course there are skittish ones, but those at the top of the food chain are obviously not constantly queued up on that threat, and IMO humans haven't been for thousands if not tens of thousands of years. So I (of course) find this article fascinating and a far more accurate presentation of behavior. I think it sort of overstates things a bit for discrete events - but I think it kind of loses it a bit at the end: "uncertainty, like all things that are metabolically costly, needs to be managed in a metabolically efficient way." .

That sentence deserves an entire article, and about 5 definitions to start with. Maybe it means (some people) are bored by predictability and by base predictions and more interested in threats, and seek them out? Like 'fight or flight' I think it's way too broad a generalization to be useful.
posted by The_Vegetables at 2:28 PM on September 17

but I think it kind of loses it a bit at the end: "uncertainty, like all things that are metabolically costly, needs to be managed in a metabolically efficient way." .

That sentence deserves an entire article, and about 5 definitions to start with.

not to be glib, but there *are* articles, vast tracts of academic verbiage, unpacking why uncertainty is metabolically costly and what managing it in an efficient way means, none of which really fits in an article written for popular consumption. like, the article talks about uncertainty as if it were a single thing, but you could subdivide uncertainty into subtypes. like how confident you are in your sensory input or which response is best given your (uncertain) perception, or what the actual outcome of your response is given all the other levels of uncertainty and the stochastic nature of the world itself. there is known uncertainty and unknown uncertainty, and uncertainty about uncertainty which is a different thing, and then you can get all meta on it and talk about the uncertainty of the uncertainty of your uncertainty.

a kind of background context to the article is karl friston's free energy principle/active inference, which remains in the background for good reason, namely because it sounds fucking bonkers, a fact which is not helped by friston himself being kind of an opaque writer and frequently saying shit like 'self-organizing systems maximise the evidence for their own existence' which sounds like straight-up feces-smearing madness, but the math more-or-less checks out.

so you have this whole research direction in neuroscience that is almost impossible to explain to people outside the field, and even many of the people in the field have some *visceral* reactions about it, and the kind of irony of the whole thing is that people are resistant to hearing a story about the function of the brain being uncertainty reduction because brains are so great at reducing uncertainty that you don't even notice you're doing it. it is, as the article points out, so well-learned as to be almost automatic. and the reason it has to be well-learned and automatic is because otherwise it would be too metabolically demanding given all the horrible uncertainty one has to deal with on a day-to-day basis.

anyway
posted by logicpunk at 7:03 PM on September 17 [7 favorites]

You don't sit around and think it through, you don't do math about it, clearly

I dunno, I'd argue that you clearly do do maths about it, even if not on a conscious level. It's the same kind of doing maths about it that you do when working out a trajectory to catch an incoming ball, for example - sure you aren't consciously running through an equation in your head, but a trajectory calculation is nonetheless happening.
posted by Dysk at 9:08 PM on September 17 [1 favorite]

The interplay between recall, prediction and awareness is something that practising the drums (I won't dignify any of the horrendous noises I make by claiming to play the drums) frequently brings to my attention. A question that keeps recurring while but mostly after doing that is: how long is now? Seriously. Estimates in milliseconds, if you please.
posted by flabdablet at 12:18 AM on September 18 [1 favorite]

Also, related: how big is here? Estimates in millimetres appreciated.
posted by flabdablet at 12:20 AM on September 18

how long is now? Seriously. Estimates in milliseconds, if you please.

In the 1930s there was an Umwelt (self-world) studies research institute in Germany. They did all sorts of interesting research, and one was to determine the temporal "now" of a snail. They put a clamp on a snail's shell to hold it in place, and placed it on a ball which would rotate around, so that the snail could move while nevertheless being fixed in place. And then they started poking the snail, who would of course retreat into its shell. But they discovered that once you poke the snail at around four times a second (I think, this is from memory), it would think it is a stable platform and attempted to crawl onto the poking stick. So the the now of a snail is around two tenths of a second or thereabouts. As far as reacting to being poked is concerned, anyway.
posted by Pyrogenesis at 1:13 AM on September 18 [3 favorites]

Thanks for that! I've found an account of the snail experiment in A Foray into the Worlds of Animals and Humans, the whole of which I'll soon be reading with great interest. In particular, the idea of the passage of time as an experience generated by processes that model reality snapshot-wise is food for thought.

My main motivation for practising drums is enjoyment of and fascination with the unusual states of awareness that become accessible by doing so; a lot of the material in Taft and Laukkonen's conversation linked above resonates strongly with me.

One metaphor I've recently started using when thinking about what happens behind the kit is based on the existence of the Fourier transform, a mathematical process allowing time-based information to be recast as fully equivalent information in the frequency domain and vice versa. The feature of that transform that I find most interesting is the way consistency of representation in one domain manifests as simplicity of representation in the other.

If while playing my usual improvisational mash I make a conscious effort not to care about trying to keep the timing of my strokes precise, but instead put my attention on playing in such a way as to hear what comes back as a collection of overlapping rhythms that's simple enough to be both comprehensible and manipulable, then two interesting things happen: the precision of what I'm playing hugely improves, and I find my way into a state of awareness where the passage of time just goes away. There is no now, there is only this, and how this feels is, at its best, fractal and intricate and gorgeous - but even when it doesn't rise to those heights, it always feels spatial.

Having found my way to that place I can usually stay there for as long as I want, doubtless to the chagrin of my neighbours (although only one of them has ever actually complained).
posted by flabdablet at 5:38 AM on September 18 [2 favorites]

Now do it with a group of people, some with drums as big as a car: taiko.
posted by seanmpuckett at 5:43 AM on September 18 [1 favorite]

One complaining neighbour is quite enough for me :-)
posted by flabdablet at 5:49 AM on September 18 [1 favorite]

In the 1930s there was an Umwelt (self-world) studies research institute in Germany. They did all sorts of interesting research

ummmmmmmmmmmmmm
posted by grumpybear69 at 12:44 PM on September 18

um indeed
posted by BungaDunga at 12:50 PM on September 18 [2 favorites]

how long is now? Seriously. Estimates in milliseconds, if you please.

I recently attended the laureate lecture of Doris Ying Tsao for her Kavli Prize in neuroscience, where she presented some pretty interesting evidence -- to this complete neuroscience amateur at least -- that we actually experience the world consciously in short bursts of maybe a few hundred milliseconds, depending on the situation. So there's an estimate!

The work is apparently still unpublished, but the talk I attended is supposed to go up on the Kavli Prize's official Youtube channel at some point, and she also talks about it a bit in her recent interview on Sean Carroll's podcast Mindscape.
posted by Spiegel at 4:20 PM on September 18 [3 favorites]

It's the same kind of doing maths about it that you do when working out a trajectory to catch an incoming ball, for example - sure you aren't consciously running through an equation in your head, but a trajectory calculation is nonetheless happening.

I started to do a lengthy quote from her book about catching a ball because I think the idea is really marvelous--it's not quite like creating a million frames of a movie focused on the ball and your hand, and it's not quite like creating a million sequential universes where the only things that exist are the ball and your hand, and yeah, it is definitely a kind of unspoken math, a visual kind, and astonishingly beautiful. (there's a bit of the idea from this link she has online as part of a footnote from that example in the book)
posted by mittens at 5:46 PM on September 18 [1 favorite]

that we actually experience the world consciously in short bursts of maybe a few hundred milliseconds

I read a lot about meditation and practice it in great detail and this is generally the consensus - our brain creates a new “moment” about every 200ms or so. Anything deeper than that and you switch into the pure sensory information (ie light flashing) which is a whole different experience in and of itself.
posted by St. Peepsburg at 10:54 AM on September 19