Friday Cephalopod: Snuggleupapus »« Grading done

Frugal to the point of vacuity

What does it take to get Carl Zimmer to review your research in the New York Times?

I suppose it helps to be at Harvard. It also helps to have a combination of subjects — evolution and the human brain — that Zimmer has written about in the past. It helps to have a paper with lots of very pretty diagrams — the authors’ hypothesis is professionally illustrated. It’s also a good idea to have a vast sweeping explanation for the exceptionalism of the human brain. In this case, they call it the Tethering Hypothesis, and it’s supposed to explain how humans evolved all these remarkable cognitive abilities.

The human cerebral cortex is vastly expanded relative to other primates and disproportionately occupied by distributed association regions. Here we offer a hypothesis about how association networks evolved their prominence and came to possess circuit properties vital to human cognition. The rapid expansion of the cortical mantle may have untethered large portions of the cortex from strong constraints of molecular gradients and early activity cascades that lead to sensory hierarchies. What fill the gaps between these hierarchies are densely interconnected networks that widely span the cortex and mature late into development. Limitations of the tethering hypothesis are discussed as well as its broad implications for understanding critical features of the human brain as a byproduct of size scaling.

You know what you don’t need? Data, or a hypothesis that makes sense.

The paper is largely a review of neuroanatomy, describing features of the human brain that we’ve known about for a long, long time…except now we can illustrate them with lovely color diagrams and fMRI scans. Here’s an illustration of the problem in human evolution:

brainevo

There are areas of the brain that we know what they do: in red, for instance, is the primary somatosensory cortex, which is a map of muscles and sensory areas on our skins, while blue is the primary visual cortex, which is where information from our eyes is processed. In between these known areas are great beige unknowns — regions of the brain called association cortex, which integrate information from various other regions in complex ways. Our primary somatosensory and visual cortices aren’t much bigger than those of a chimpanzee, which makes logical sense, since there isn’t much difference in surface area or visual acuity between us, and most of the growth has occurred in the association cortex.

All well and good. The question is, what made our association cortex expand in our evolution, and how is that expansion related to specific human intellectual capacities? Those are good questions, and I’d be curious to see them answered. Too bad this paper doesn’t.

One problem is that it is a review paper and really doesn’t test anything — it catalogs some existing knowledge about brain organization and then throws out this Tethering Hypothesis to explain it all, which it doesn’t. I do like the fact that it suggests that most of our abilities are spandrels, not explainable as adaptations, and that what it proposes is that novel abilities arose from regions of the brain that were not constrained by ancestral functional requirements. I just don’t see how their mechanism explains that.

Here’s one short paragraph from the paper that neatly summarizes their hypothesis.

The idea of some form of radiation outward from core organizing centers is appealing because the hominin cerebral cortex vastly expanded in a short time. It seems implausible that molecular gradients could emerge fast enough to specify new cortical areas, although developmental expression patterns have clearly been modified. Building from Rosa and colleagues’ ideas about visual cortex organization, we propose a more general tethering hypothesis to explain how new features of cortical organization might have emerged during the rapid evolutionary expansion of the cerebral mantle. The word ‘tether’ is used to emphasize that the expanding cortical plate is tethered to gradients that initially evolved in a cortex with a far smaller surface area. Much as taffy, being pulled apart, thins until it breaks in the middle, the expanding cortical zones far from the strong constraints of developmental gradients and sensory input may become untethered from the canonical sensory–motor hierarchies.

OK, that begs the question: why did the hominin cerebral cortex expand in the first place? They keep talking about this “expanding cortical plate”, but not why it was expanding or why it necessitates new organizing centers. The taffy metaphor is also telling; why are they talking about things being pulled apart, when expansion of the brain is not caused by external forces pulling on it, but on internal forces of growth generating more tissue between known cortical zones?

I’m also put on edge by the phrase “It seems implausible that…”, especially when applied to something that doesn’t seem implausible at all. Why balk at a timescale of several million years to evolve a use for a bit of extra brain matter?

But even worse, nervous systems growing bigger is what I study. My Ph.D. research was on connectivity in the developing spinal cord of zebrafish, for instance.

The first neurons in the zebrafish embryo emerge at about 18 hours after fertilization, at a time when the nascent spinal cord is about 2mm long, in total. Cells in the hindbrain send axons all that distance (2mm is a long way in an embryo!), and as they grow, they make a little knot of synapses every 40-50µm with cells called primary motoneurons.

That’s in the embryo. In the adult, the spinal cord is roughly 4cm long — there’s been a 20-fold expansion in size. What do you think happened to that earlier array of cells? Did the system stretch and break?

No, it grew. In the adult, the same hindbrain neurons are still present, and their axons still reach all the way back to the tailtip. And the same motoneurons are still present, they’re just spread out more to be separated by 1-1.5mm, and they still retain the same synapses.

I also did research on the earliest neurons to differentiate in the grasshopper nervous system. I studied Q1, a neuron that established one of the commissures in the grasshopper ganglion. That story is a little different: Q1 doesn’t seem to have any function in the adult, and in fact looks to be abandoned and gone. But what it does is send the first slender thread across on a specific pathway; it pioneers a route across the nervous system, and then other axons pile on and follow it across. It’s like sending a kite string across a chasm, then using the string to pull a rope across, and then using the rope to pull a cable across, and pretty soon you’ve got a bridge — and it’s doing this as the chasm is widening, because like the zebrafish, the grasshopper is also growing substantially during these events.

Growth is an integral process in the development of the nervous system. Without specific evidence that these developmental mechanisms break down during growth (which seems implausible to me…), why would you postulate that a failure of developmental processes was an essential element of human evolution? A tripling of brain size from the human-chimpanzee common ancestor to the modern human seems like a small shift relative to the much larger expansion of the human fetal brain to the adult brain.

What I’d like to see, and did not find in this paper, are comparative developmental studies. When these various cortical regions of the brain are specified or establishing connectivity, how far apart are they in different species? Compare mouse and rat, for instance, or rhesus monkey and human. I suspect that in early embryos, the distances, and their relative differences, will be minuscule, and that signaling centers will be close enough that it will seem silly to argue that broad patterns of connectivity would be unable to form in the biggest brains, leaving gaps that need to be filled in by novel mechanisms and structures.

But again, the paper doesn’t look at any of that at all. I found one paragraph that briefly discusses other observations that association cortex matures later than other regions of the brain, and that’s about it — it is definitely not sufficient information to argue that association cortex is out of reach of intrinsic signaling gradients in the early brain.

At least the first subtitle in the paper is “A Speculative Hypothesis,” which is entirely accurate. I don’t see how it justifies the praise it was given in Carl Zimmer’s article.

Dr. Sherwood, the George Washington University expert, praised the hypothesis for being “fairly frugal.” The emergence of the human mind might not have been a result of a vast number of mutations that altered the fine structure of the brain. Instead, a simple increase in the growth of neurons could have untethered them from their evolutionary anchors, creating the opportunity for the human mind to emerge.

Oh, wait. When the best thing you can say about a hypothesis is that it is “fairly frugal”, that’s not much praise at all.


Buckner RL, Krienen FM (2013) The evolution of distributed association networks in the human brain. Trends Cogn Sci 17(12):648-65.

Comments

  1. Nice Ogress says

    This research paper might be a hot mess, but it’s had a good effect – people who are knowledgeable on the subject are all riled up – and that might be better, in the long run, than if their paper was tepidly right and everyone could ignore it.

    It’s like the old story of someone trying to get directions at a bar. All the locals ignore her, until one guy gives directions that are so cockamamie awful everyone else chimes in to correct him. The end result (deliberate, of course) is that the traveler gets good directions after all.

    S’important to remember that spectacular failure can be just as useful as success, if it’s properly managed.

  2. DonDueed says

    It seems to me that “frugal” can indeed be high praise, when one considers Occam’s Razor.

  3. ragdish says

    “In between these known areas are great beige unknowns — regions of the brain called association cortex, which integrate information from various other regions in complex ways”

    Great beige unknowns? Damage to an association area such as the fusiform gyrus results in results in prosopagnosia. That area is involved with visual integration towards facial recognition. Indeed the work by cognitive neuroscientists such as Antonio Damasio reveal that those regions are likely neuronal ensembles that have binding codes coordinating the neural activity downstream in the primary visual cortices to allow you to recognize your Aunt Betty. He refers to these neural ensembles as “convergence zones” (Convergence and divergence in a neural architecture for recognition and memory. Meyer K., Damasio A. Trends Neurosci. 2009 Jul 32(7): 376-382). Others such as Marsel Mesulam have proposed similar binding theories. Albeit this work on the association areas is far from complete, would you really dismiss such evidence as “great beige unknowns”?

  4. stevem says

    “Fairly frugal”, is that some kind of euphemism implying Occam’s Razor? I.E. the maxim that the simplest explanation is usually the correct one, thus a “frugal” explanation is pretty simple without too many ‘other’ requirements. The whole paper is way beyond my capabilities, just wondering about the use of that phrase. Is it just a “subtle” implication that this explanation fulfills Occam’s requirement for a “good theory”? I can’t go any further, way beyond me, just curious, back to listening…

  5. David Marjanović says

    What’s particularly appalling is the tree above. It’s completely wrong.

    Rodents and primates are more closely related to each other than to anything else in the tree.
    “Ungulates”. There are no “ungulates”. Why not just write artiodactyls? That’s what sheep are. Anyway, artiodactyls and carnivorans do not have a long common history not shared with bats (chiropterans). If it’s not a hard polytomy (a genuine multifurcation), bats and carnivorans are probably a bit closer to each other than either is to artiodactyls.
    These three together form a branch that is closer to the hedgehogs (lipotyphlans – there are no “insectivores”) than to rodents + us.
    How can you reconstruct the mammalian common ancestor if you only look at placentals? You plainly can’t, so show us the marsupials and monotremes you used already!
    This is pure nomenclature as far as this tree is concerned, but there’s no such thing as a “prosimian” either: the tarsiers are closer to us than to the strepsirrhines (lemurs + lorises).

    There’s no excuse for this. The name Afrosoricida was only coined after practically all of the above was well established.

    Aaaaaaaaand of course the tree has to be depicted with us on top. *sigh* *headshake*

  6. ibbica says

    Human brains? Bah. I’m still waiting for someone to explain the elephantfish brain to me…

    (Incidentally, does anyone know where I can get a legit digital or paper copy of the comparative brain poster on that page, from The Central Nervous System of Vertebrates? Or if Springer’s telling the truth when they claim that book will be “available soon” as an e-Book? Probably don’t need the whole textbook for intro A&P, but that’s kind of an amazing poster.)

  7. says

    #7, David Marjanović:
    I didn’t even look at the tree! But now that you mention it…that’s freaky.

    #5, ragdish:
    I know. I’m aware of those bits we understand, at least as far as effects from stroke or epilepsy. I’m echoing the paper, which treats those big beige areas as a problem.

  8. ChasCPeterson says

    No, it doesn’t

    I was going to point that out too, but in this case it’s not so egregious. The argument in question can be formulated something like this:
    1. The human neocortex is exceptionally large. Why?
    2. Because of all this association cortex. Where did it come from?
    3. It evolved because these areas were developmentally untethered from signalling gradients. Why?
    4. Because the human neocortex is so exceptionally large.
    Which, if accurate, is at least similar to assuming the conclusion.

  9. ChasCPeterson says

    ibbica @#8: thanks for a fascinating link.

    in the electric fish Gnathonemus petersii, the brain is responsible for approximately 60% of body O2 consumption, a figure three times higher than that for any other vertebrate studied so far, including human.

    wow!!

    (And that is a nice-looking figure; I can’t read it well enough to judge if the illustrated tree is phylogenetically accurate though.)

  10. David Marjanović says

    ibbica @#8: thanks for a fascinating link.

    From there:
    “This rostral protrusion of the cerebellum is only present in actinopterygian fish.”
    I knew the whole-genome duplication had to have had some lasting effects! :-)

    (And that is a nice-looking figure; I can’t read it well enough to judge if the illustrated tree is phylogenetically accurate though.)

    Nobody can read that figure well enough. Most letters are only one pixel tall, and there doesn’t seem to be a bigger version on the website. *grumble*

  11. unclefrogy says

    are they saying that the human brain and therefore human intellectual abilities developed by some other mechanism then all the other animals because our brain got too big?
    Are they implying that our thinking in some way is separated (untethered?) from the rest of the nervous system?
    or is it like Chas suggested another complicated circular argument?
    uncle frogy

  12. rwiess says

    All the questions about what drove the human brain to become larger almost sound like questions in search of a divinity. To widen the field, start with the cooking hypothesis – that cooking made great quantities of energy available to support an enlarged brain – the standard drive to gigantism when it is a feasible alternative – and from there we found new things to do with the new tools.

  13. peterh says

    “Instead, a simple increase in the growth of neurons could have untethered them from their evolutionary anchors, creating the opportunity for the human mind to emerge.”

    I’m certainly not a biologist (other than what I pick up in these threads) but that sounds like pure bullshit.

  14. ChasCPeterson says

    the standard drive to gigantism when it is a feasible alternative

    the what, now?

    My guess is brain expansion was triggered by a black monolith.

    ha!
    There’s evidence! I think I remember seeing a documentary!

  15. chigau (違う) says

    I prefer The Rwiess Cooking Hypothesis.
    The need to bake the perfect soufflé could easily account for much technological innovation.

  16. David Marjanović says

    the what, now?

    Cope’s Rule, which isn’t a rule (and wasn’t apparently ever stated by Cope; Depéret stated it in extremely strict terms some 30 years later).

    There’s always the Jaw theory.

    …which fails to explain why we have much stronger bites (and much thicker enamel) than chimps.

  17. David Marjanović says

    …True. There is, however, of all things, “Marsh’s Law” which states that the brain always grows! :-D

  18. brucegee1962 says

    It seems pretty likely to me that humans got their big brains the same way that peacocks got their big tails — some form of sexual selection. Adaptive selection doesn’t seem to offer enough of an advantage for the energy costs of having a big brain to your average hominid. But if the chicks were diggin’ the smart guys, and the guys were diggin’ the smart women (or more likely some proxy for smartness, like artistic ability or jewelry-making skills) then that might easily explain the rapid brain expansion over a relatively short time.

  19. Pierce R. Butler says

    brucegee1962 @ # 26: But if the chicks were diggin’ the smart guys, and the guys were diggin’ the smart women (or more likely some proxy for smartness, like artistic ability or jewelry-making skills) then that might easily explain the rapid brain expansion over a relatively short time.

    So what went wrong between then and now?

  20. gillyc says

    the expanding cortical zones far from the strong constraints of developmental gradients and sensory input may become untethered from the canonical sensory–motor hierarchies

    I’m surely not the only one who reads that and thinks it sounds, well, disabling at the very least? Can’t imagine that being selected for.