Maybe Your Zoloft Stopped Working Because A Liver Fluke Tried To Turn Your Nth-Great-Grandmother Into A Zombie

Or at least this is the theory proposed in Brain Evolution Through The Lens Of Parasite Manipulation by Marco del Giudice.

The paper starts with an overview of parasite manipulation of host behavior. These are the stories you hear about toxoplasma-infected rats seeking out cats instead of running away from them, or zombie ants climbing stalks of grass so predators will eat them. The parasite secretes chemicals that alter host neurochemistry in ways that make the host get eaten, helping the parasite transfer itself to a new organism.

Along with rats and ants, there is a dizzying variety of other parasite manipulation cases. They include parasitic wasps who hack spiders into forming protective webs for their pupae, parasitic flies that cause bees to journey far from their hive in order to spread fly larva more widely, and parasitic microorganisms that cause mosquitoes to draw less blood from each victim (since that forces the mosquitoes to feed on more victims, and so spread the parasite more widely). Parasitic nematodes make their ant hosts turn red, which causes (extremely stupid?) birds to mistake them for fruit and eat them. Parasitic worms make crickets seek water; as the cricket drowns, the worms escape into the pond and begin the next stage of their life cycle. Even mere viruses can alter behavior; the most famous example is rabies, which hacks dogs, bats, and other mammals into hyperaggressive moods that usually result in them biting someone and transmitting the rabies virus.

Even our friendly gut microbes might be manipulating us. People talk a lot about the “gut-brain axis” and the effect of gut microbes on behavior, as if this is some sort of beautiful symbiotic circle-of-life style thing. But scientists have found that gut microbes trying to colonize fruit flies will hack the flies’ food preferences to get a leg up – for example, a carb-metabolizing microbe will secrete hormones that make the fly want to eat more carbs than fat in order to outcompete its fat-metabolizing rivals for gut real estate; there are already papers speculating that the same processes might affect humans. Read Alcock 2014 and you will never look at food cravings the same way again.

But del Giudice thinks this is just the tip of the iceberg. Throughout evolutionary history, parasites have been trying to manipulate host behavior and hosts have been trying to avoid manipulation, resulting in an eons-long arms race. The equilibrium is what we see today: parasite manipulation is common in insects, rare in higher animals, and overall of limited importance. But in arms race dynamics, the current size of the problem tells you nothing about the amount of resources invested in preventing the problem. There is zero problem with war between Iran and Saudi Arabia right now, but both sides have invested billions of dollars in military supplies to keep their opponent from getting a leg up. In the same way, just because mammals usually avoid parasite behavior manipulation now doesn’t mean they aren’t on a constant evolutionary war footing.

So if you’re an animal at constant risk of having your behavior hijacked by parasites, what do you do?

First, you make your biological signaling cascades more complicated. You have multiple redundant systems controlling every part of behavior, and have them interact in ways too complicated for any attacker to figure out. You have them sometimes do the opposite of what it looks like they should do, just to keep enemies on their toes. This situation should sound very familiar to anyone who’s ever studied biology.

Del Giudice compares the neurosignaling of the shrimp-like gammarids (small, simple, frequently hijacked by parasites) to rats (large, complex, hard to hijack). Gammarids have very simple signaling: high serotonin means “slow down”, low serotonin means “speed up”. The helminths that parasitize gammarids secrete serotonin, and the gammarids slow down and get eaten, transferring the parasite to a new host. Biologists can replicate this process; if they inject serotonin into a gammarid, the gammarid will slow down in the same way.

Toxoplasma hijacks rats and makes them fearless enough to approach cats. Dopamine seems to be involved somehow. But researchers injecting dopamine into rats don’t get the same result; in fact, this seems to make rats avoid cats more. Maybe toxoplasma started by increasing dopamine, rats evolved a more complicated signaling code, and toxoplasma cracked the code and now increases dopamine plus other things we don’t understand yet.

Aside from the brain, the immune system is the most important target to secure, so this theory should predict that immune signaling will also be unusually inscrutable. Again, this situation should sound very familiar to anyone who’s ever studied biology.

Second, you have a bunch of feedback loops and flexibility ready to deploy at any kind of trouble. If something makes dopamine levels go up, you decrease the number of dopamine receptors, so that overall dopaminergic neurotransmission is the same as always. If something is making you calmer than normal, you have some other system ready to react by making you more anxious again.

Del Giudice makes the obvious connection to psychopharmacology. Many psychoactive drugs build tolerance quickly: for example, heroin addicts constantly need higher and higher doses to get their “hit”. Further, tolerance builds in a pattern weirdly similar to antibody response – it takes a while to build up a cocaine tolerance, and you lose it over time if you don’t use cocaine, but the body “remembers” the process and a single hit of cocaine years later is sufficient to bring you back up to the highest tolerance level you’ve ever had.

The standard explanation for tolerance is that it’s an attempt to maintain homeostasis against the sort of conditions that can cause natural variation in neurotransmitter levels. I never questioned this before. But why is the body prepared to suddenly have all its serotonin reuptake transporters inhibited? Is that something that frequently happens, out in nature? I guess maybe plant toxins could do that, but then how come the body is prepared to deal with this for months or years?

While not denying the value of these standard explanations, Del Giudice thinks defense against parasite behavior manipulation may also play a role. Remember, gammarids absolutely have parasites that try to increase their serotonin levels as a prelude to getting them killed. Is it that surprising that a lot of different animal lineages would develop a reaction of “If something other than normal cognition has started increasing your serotonin levels, it’s a trap and you need to get them back down again”? Does that explain why SSRIs don’t work for some people, or randomly stop working, or need frequent dose escalation?

Third, you encode messages in the timing of pulses. This is a central feature of neuroendocrine communication – an intense pulse of testosterone at 6 AM means something different from tonically high testosterone all day. Parasites cannot do pulses. Remember, these parasites are usually microscopic. Each parasite can only produce a miniscule quantity of neurotransmitter or hormone. Only colonies of thousands or millions of parasites can produce enough chemicals to affect host signaling. This parasites cannot communicate or coordinate with each other, so there’s no way for them to be producing lots of testosterone one minute and none at all the next. That means that when a hormone arrives in a pulse, or better yet a complicated pattern of pulses, that’s a pretty reliable sign that it’s coming from a real gland.

Fourth, you exploit your individuality. The immune system already does this; there are some genes called the major histocompatibility complex that are designed to be especially variable, such that most people (except identical twins) will have different MHCs. These help the immune system differentiate self from other. Because they have such high individual variability, pathogens can’t just evolve around the MHC; they would have to undergo an entire evolutionary process for each new host they invade.

Del Giudice wonders if parasite-host arms races created pressure for increased human variability. SSRIs will make some people less depressed. But some people will get more depressed. A few will even get suicidal. A very few will flip out and become psychotic, or improve much more quickly than the textbooks say should be possible and feel completely reborn on day 3, or have something else even weirder happen. I always assumed God just hated psychiatrists and wanted them to be miserable. But another possibility is that extreme individual variability in neurosignaling pathways is a defense against parasite manipulation. If the effects of serotonin are unpredictable for any individual, no parasite species can devise a universally valid mechanism for controlling its hosts.

Fifth, you let the parasites become part of the furniture. If everybody in your ecosystem is infected with a parasite that raises serotonin, you just evolve a tonically lower serotonin level, and then it all cancels out. This one seems a little bit weird to me – surely this isn’t the stable equilibrium? But:

A downside of preemptive strategies is evolved dependence (de Mazancourtet al. 2005): if brain physiology and behavior are designed to function optimally when the parasite is present, the absence of the parasite will lead to inappropriate or fitness-reducing behaviors (Weinersmith and Earley 2016; see also Johnson and Foster 2018).

I think this is meant to hint at the “hygiene hypothesis”, ie our immune systems are screwed up because we are not getting exposed to the parasites it was built to expect. Suppose lots of parasites try to downregulate the immune system (which sounds logical enough), and the body doesn’t know which ones it’s going to get but expects it to follow a Poisson distribution around some mean. Then it might just upregulate the activity of the immune system that same amount. If you get rid of all the parasites, then your immune system is just set too high and you get autoimmune disorders.

(in case you had the same question I did – yes, the parasitologist Kelly Weinersmith cited above is the same Kelly Weinersmith who co-wrote Soonish with Zach Weinersmith of SMBC fame.)

Sixth, you use antiparasitic drugs as neurotransmitters. This is the kind of murderous-yet-clever solution I expect of evolution, and it does not disappoint. Several neurotransmitters, including neuropeptide Y, neurokinin A, and substance P are pretty good antimicrobials. The assumption has always been that the body kills two birds with one stone, getting its signaling done and also having some antimicrobials around to take out stray bacteria. But Del Giudice proposes that this is to prevent parasites from hijacking the signal; any parasite that tried to produce or secrete an antiparasitic drug would die in the process.

Dopamine is mildly toxic. The body is usually pretty good at protecting itself, but the mechanism fails under stress; this is why too much methamphetamine rots your brain. Why would you use a toxic chemical as a neurotransmitter? For the same reason you would use antiparasitic drugs – because you want to kill anything smaller than you that tries to synthesize it.

People always talk about the body as a beautiful well-oiled machine. But sometimes the body communicates with itself by messages written with radioactive ink on asbestos-laced paper, in the hopes that it’s killing itself slightly more slowly than it’s killing anyone who tries to send it fake messages. Honestly it is a miracle anybody manages to stay alive at all.

All these features together are a pretty effective way of dealing with parasite manipulation. There are a few parasites that can manipulate human behavior – rabies definitely, toxoplasma maybe – but overall we are remarkably safe.

Del Giudice argues that a combination of factors make it easy for parasites to manipulate insects but not large vertebrates. First, insects are small, so you only need a few parasites to produce an insect-sized level of neurotransmitter. Second, insects are so simple that usually one neurotransmitter maps nicely to one behavior; they are too small to support multiple redundant systems or complicated signal cascades. Del Giudice writes:

Although parasites can evolve subtler and more indirect means of manipulation, their computational capabilities are ultimately limited by their size. As the size and complexity of the host’s brain increase relative to the parasite, the disparity may become so extreme that the host is able to “outcompute” its adversary, making complex manipulations effectively impossible. The parasite may still be able to alter the host’s behavior in nonspecific ways (e.g., sickness, brain damage), but is unable to induce the kind of coordinated pattern required for trophic transmission or bodyguard manipulation. Although this argument is admittedly speculative, it is consistent with the fact that complex behavioral manipulations have not been documented in larger, warm-blooded animals (see Lafferty and Kuris 2002).

Finally, almost nothing eats humans, so there aren’t a lot of parasites interested in using us as a vehicle to get to their definitive hosts. If parasites want anything from us, it’s probably STIs wishing we had more risky sex; accordingly, Del Giudice obliquely cites Greg Cochran’s controversial hypothesis that homosexuality may be related to parasites hijacking sexual machinery.

But let’s take a step back: is any of this true?

The strongest evidence against is the dog that didn’t bark. Some systems look heavily defended against parasite manipulation, but others don’t. Amphetamines raise dopamine effectively and without significant tolerance buildup (see part IV here for a defense of this claim); antipsychotics lower dopamine equally effectively and consistently. Since dopamine is one of the most lucrative systems for parasites to hijack, it’s surprising to find it so easy to affect. And what about immune function? Externally administered corticosteroids decrease immune activity and make the body more vulnerable to infection; why don’t parasites secrete them? Why don’t we have some counter against them? These systems look consistent with an evolutionary history in which we don’t expect any threat from parasite manipulation and don’t need to defend ourselves very hard.

But also: homeostasis might be the most basic activity of all living things. Every bodily system can be modeled as a striving for homeostasis in some domain or other, even high-level cognitive functions. So it’s not clear that tolerance to psychiatric drugs needs a complicated evolutionary explanation beyond just “if you increase serotonin, your body is going to try to decrease it again, because that’s what bodies do“.

So I’m not sure how much of an effect this really had. It’s an interesting theory. But whether it explains some things, nothing, or everything, it’s too early to say.

But I like this paper because it takes the complexity of biology seriously. There’s a sense that science is stagnating, and biology is one of the worst offenders. In the 1800s and early 1900s, we were pinning down our mastery of anatomy, discovering all the major hormone systems, learning about microbes and inventing antibiotics. It seemed like the same kind of thing as physics, where you could go out into the world, observe things, and make difficult but fundamentally straightforward discoveries. But for the past fifty years, it’s been kind of a mess. Despite some amazing work by amazing people, we still don’t even understand questions as basic as what depression is. Everything seems bogged down in a million different opaque signaling cascades that fight off any effort to untangle or shift them.

Del Giudice offers a seductive explanation: the perceived perversity of the human blueprint is absolutely real. Parts of it – the parts most involved in health and disease – were sculpted by evolution to be as hard as possible to understand or affect. This makes me feel better about how often the drugs I prescribe fail in surprising ways

This entry was posted in Uncategorized and tagged , . Bookmark the permalink.

152 Responses to Maybe Your Zoloft Stopped Working Because A Liver Fluke Tried To Turn Your Nth-Great-Grandmother Into A Zombie

  1. Byrel Mitchell says:

    Suppose lots of parasites try to downregulate the immune system (which sounds logical enough), and the body doesn’t know which ones it’s going to get but expects it to follow a Poisson distribution around some mean. Then it might just upregulate the activity of the immune system that same amount. If you get rid of all the parasites, then your immune system is just set too high and you get autoimmune disorders.

    But why would parasites trying to kill you prefer to downregulate your immune system? An upregulated immune system is also a rather lethal beast after all. It’s not clear that you’d get enough of an evolutionary consensus among parasites to control with a set point adjustment.

    • Scott Alexander says:

      Parasites don’t want to kill you. They want to live inside you and eat the nutrients in your body, while you live as long and full a life as possible. Their only problem is that your immune system tries to kill them; hence their interest in downregulating it.

      • Byrel Mitchell says:

        But many of the behavior-modifications that parasites are noted to do are precisely designed to kill the host and get it eaten. Being eaten is one of the more efficient ways of host-to-host communication, and you can often spread to multiple hosts if the body is consumed by several animals.

        • cold_potato says:

          Four reasons:

          * Parasites that spread by dispersal from a living host will prefer to downregulate their host’s immune system, to keep themselves alive. This sets the evolutionary consensus, which will be followed by parasites that spread through the death of their host.

          * Upregulating your host’s immune system may kill them, but it will kill off a lot of you, the parasite, before they die.

          * Downregulating the host’s immune system kills them faster. Autoimmune diseases tend to be chronic, whereas infections can be acute and quickly fatal.

          * It’s easier to destroy a complex working system (i.e. downregulate) than to make it work better (i.e. upregulate).

          On the other hand, if downregulation is the dominant strategy, hosts will develop good defences against it, which may make upregulation preferable in some cases. Either to, as you suggest, kill the host through autoimmune disease, or to kill off other, competing parasites.

        • broblawsky says:

          Parasites that infect creatures that are low on the food chain (like rats, ants, snails, or small fish) try to get their hosts eaten in order to spread. Parasites that infect creatures that are higher up on the food chain don’t, and some parasites induce different behavior depending on which parts of their life cycle they’re in. Toxoplasma makes rats get eaten by cats, but it doesn’t seem to have any analogous effect on cats themselves. It just makes them defecate more toxoplasma.

      • deciusbrutus says:

        Parasites ‘want to’ (reproduce if they) get offspring into many other hosts. The life or death of any single host is irrelevant except so far as it impacts that result.

    • luxagenic says:

      What makes you think parasites “try” to kill their hosts?

      • Byrel Mitchell says:

        All of the examples in this post of them doing exactly that?

        • Enkidum says:

          Right, but humans are not really edible by much. Probably not a viable strategy in our case.

        • Nornagest says:

          Parasites don’t try to kill their hosts for shits and giggles. They might want to drive their hosts to certain specific deaths (“eaten by predator” is a popular one, but there’s also horsehair worms manipulating cricket hosts to seek out water, etc.), but they’re not doing it out of pure malice, they’re doing it because they need the host to die in those ways to complete their life cycle. Other deaths do nothing for them. And even that isn’t a universal: there are plenty of parasites that’re perfectly content to hitch a ride on the host nonlethally and sometimes even asymptomatically. There’s a good chance you’re harboring some in your eyelashes right now.

          A parasite downregulating its host’s immune system probably isn’t trying to kill the host: it’s not doing the host any favors, but a host that gets killed by opportunistic infection is rather unlikely to die in a way that’s convenient for the parasite. It’s probably just trying to avoid getting killed by its host’s immune system.

          • sclmlw says:

            Exactly. It’s important to remember there are lots of different strategies for survival and reproduction. The point is for an organism to get its genes into the next generation and spread them more broadly if possible. A parasite that targets insects can use the fact that they reproduce very quickly to help spread its genetic material far and wide.

            The reason this strategy doesn’t work well with humans is that we replicate orders of magnitude slower than, say, fruit flies. It’s not because humans are inherently inedible. The opposite is true. We’re very edible, it’s just we developed social behaviors that help keep us safe.

            A parasite that could hijack those social behaviors might get a few humans to wander off and get eaten by wolves, lions, or any of a number of natural human predators. It’s just not as effective to spread between slowly-replicating species like humans and lions as it is to spread between insects and rodents.

            Thus a strategy that preserves the life of the host and allows you to spread between living hosts is going to be a more successful strategy in humans. This even suggests a symbiotic relationship might be a better strategy in many (but not all) cases than a parasitic one (ergo all the viral DNA in the human genome).

    • Cerastes says:

      There are two types of parasites: direct life cycle and indirect life cycle. The former, like fleas, do everything on or in their host – born, live, reproduce, die – but this can limit gene flow at best (all parasites on a host might be descended from original colonists) or fail at worst (when a single individual of a sexually reproducing species finds a host). To overcome this, many parasites are transmitted between hosts at various lofe cycle stages – toxoplasma eggs hatch and grow in a rodent, then, once ot gets eaten by a cat, breed with other toxoplasma from other rodents aso the eggs come out in the cat feces.

      Killing the host is just a mechanism of achieving the true goal, usually getting into the gut of the next host (guts are a great way to gain access to an animal’s innards). The parasite needs the host to be eaten by a specific predator group, not just to be dead. They’re driving to a specific destination, so they want to make sure they get there and the car doesn’t fall apart first, but they also need to make sure the car doesn’t kill them on the way.

      • John Schilling says:

        If the host is an apex predator, the indirect strategy isn’t likely to work very well – or at least it’s going to require a complex multistage process. Things that eat apex predators (i.e. worms) are usually so very unlike apex predators that the same parasite is unlikely to be well-adapted to both, and “maybe I can get my apex predator eaten by a rival apex predator” is probably not going to succeed often enough to be worth the bother.

        Hence, not much of this to be expected in human parasites. If there’s any truth to it in the first place.

        • Cerastes says:

          For the vast majority of our history, we were not apex predators; humans were frequently eaten by large cats, possibly wild dogs, and definitely by crocodiles. In rural areas where these species haven’t been wiped out, they still kill a fairly substantial number of people, hundreds or thousands a year.

          Secondly, while parasitic manipulation to get eaten is the most dramatic and well studied, more subtle manipulations might increase host exposure to vectors such as ticks or mosquitoes, which could transport the parasite or its eggs.

          • acymetric says:

            Can you give some sense of what you mean by “frequently”? I’m not sure it would match how I use the term.

            Also, are you differentiating between being “eaten” vs. being “attacked”?

          • Eponymous says:

            Crocodiles are the big one these days, and they definitely eat people they kill.

            Big cats rarely attack humans, but if they get a taste they can become big killers and consume remains.

            Historically wolves were apparently a big problem — many stories. Presumably ate their victims.

            Many animals would eat human remains.

            Most shark attacks do not involve consumption.

          • acymetric says:

            But there are only about 1,000 crocodile deaths per year. Even if we triple that estimate, they are still not (in my view) common enough for some pathogen to evolve to take advantage of crocodiles eating people. Probably by several orders of magnitude.

          • John Schilling says:

            I think wolves would be the only predatory threat maybe significant enough to drive parasite evolution in human hosts, and I’m not sure they are significant enough for this purpose. That depends on how dangerous wolves were to hunter-gatherer populations, a subject I and google are both woefully short of hard data on. Definitely a threat to peasant farmers and their livestock, but even there it is mostly the livestock at risk. Human hunter-gatherers, w/re wolves, notoriously Met the Enemy and Made Him Us. Our best friend at least.

          • ll11 says:

            IIrc there are some wolves in Ethiopia, but they tend to hunt alone rather than in packs, and aren’t individually very large; otoh, there is direct evidence of predation on early humans by large cats (specifically, skulls with puncture marks that match up to leopard teeth). Lions (especially solitary males not hunting for a whole pride) certainly aren’t averse to taking a human if they find one alone. Hyenas and wild dogs probably could take humans as well, but again the prey that they actively *look for* is more on the order of horse-sized. Not that any of them object to humans as a snack, just that the direct predator-prey relationship necessary for a parasite to become dependent on it isn’t really there for most potential evolutionary-environment predators.

          • DarkTigger says:

            Not that any of them object to humans as a snack, just that the direct predator-prey relationship necessary for a parasite to become dependent on it isn’t really there for most potential evolutionary-environment predators.

            Well just because non of the predators alive today hunt humans primarily, does not mean that non in the past did.
            We know that the introduction of Homo Sapiens to an environment is strongly corelated with extinction events for mega fauna in that environment. Which is even true for Africa shortly after behavioral modern humans developed.
            I would not be surprised if every predator that primarily hunted for the genus Homo had nasty case of getting extinct, in the last 300k years.

  2. paulbali says:

    Going thru the Del Guidice now. Fascinating!

    From the Del Guidice, quoting Read & Braithwaite [2012]:

    “The other way is to counter the manipulation itself, either by making behavior control systems less vulnerable to attack, or by recalibrating things to accommodate the manipulation.”

    Interesting to note that Life has, even vis a vis parasites, an ahimsic option, a Do-Not-Kill immune response to even this most intimate of intrusions into the Self: we simply defend ourselves from their attack, by adjusting our own selves.

    _______
    Hypotheses aside, I do condemn the continuation of predator-exposure experiments on my rodent friends. There, my ceterum censeo.

  3. amirlb says:

    This kind of speculation is related to a thought I have for some time now: how comes allergy to condoms/latex isn’t more common? These kinds of allergies exist, and there has been documented use of condoms for centuries. Should be enough time for natural selection to do its thing with a strong incentive, isn’t it?

    • metacelsus says:

      there has been documented use of condoms for centuries

      Early ones were sheep-gut, though, not latex. Latex ones started becoming prevalent in the early 1900s.

      Also, latex allergy would need to be highly heritable for selection to operate meaningfully over a short time. I’m not sure that it is.

      However, I would expect there to be strong selection pressure for being too impulsive to use condoms (or other birth control) . . .

      Sufficiently strong selection by STIs in the opposite direction might cancel things out, though.

    • noyann says:

      Latex allergy mostly comes through cross-reactivity within a (probably pollen-acquired) allergy.

      “Approximately 30-50% of individuals who are allergic to natural rubber latex (NRL) show an associated hypersensitivity to some plant-derived foods, especially freshly consumed fruits. This association of latex allergy and allergy to plant-derived foods is called latex-fruit syndrome. An increasing number of plant sources, such as avocado, banana, chestnut, kiwi, peach, tomato, potato and bell pepper, have been associated with this syndrome. The prevailing hypothesis is that allergen cross-reactivity is due to IgE antibodies that recognize structurally similar epitopes on different proteins that are phylogenetically closely related or represent evolutionarily conserved structures. “ https://www.ncbi.nlm.nih.gov/pubmed/12440950

      “In our study, tomato, potato, and latex showed a common band of 44-46 kDa probably corresponding to patatin. This protein could be implicated in the high cross-reactivity between tomato, latex, and potato observed in the immunoblot and CAP inhibition.” https://www.ncbi.nlm.nih.gov/pubmed/11736750

      ETA: Compare the times pollen resides in the respiratory pathways, and food in the intestines, with the usual condom contact times. Also, the nature of the involved body surfaces differs strongly.

  4. Two objections:

    1. I don’t think that your argument that “the current size of the problem tells you nothing about the amount of resources invested in preventing the problem” makes sense. Evolution does not work without selective pressure. If parasites attempting to manipulate behavior barely affect an organism we should expect that species to remain at the same level of complexity until some parasite does start affecting its behavior. A hypothetical war between Saudi Arabia and Iran is different because humans are intelligent and can anticipate a potential risk without it ever happening, unlike natural selection. It’s possible that parasites have a small effect on behavior that nonetheless produces enough selective pressure to ensure that organisms continue becoming more complex just to counteract this. This seems like something we should be able to detect if we look for it, although I’m not sure what the current state of the evidence is on the matter. Also, it raises the question of why mammals have it so much easier to confuse parasites than insects that they require less selective pressure to balance the arms race.

    2. I think there’s already an adequate explanation for why biology is so complex. Simply put, there’s no selective pressure to make things simple. If some sort of system behaves simply with one choice of parameters but is complicated and hard to predict with another choice of parameters, then all else being equal humans would prefer the simpler system until they can understand the more complicated system, but evolution would be indifferent and choose the more complicated system if it produces the slightest advantage. And even human-designed systems have a tendency to become more complex over time.

    There may be some circumstances where the body deliberately creates a hard-to-fake signal to avoid being coopted by parasites. On the whole, however, I believe that more complex organisms being harder to affect by parasites is a side effect rather than a cause for their complexity.

    • teageegeepea says:

      In the absence of pressure from selection, de novo mutations will accumulate, breaking complex adaptations that had evolved previously.

      We have evolved to be “simple” in some ways that biological constraints forced us to economize on, and “complex” in other ways we could afford to be. Robin Hanson compares our brains specifically to human-written software along those lines in “The Age of Em”.

      • You’re right that there are also ways in which evolution tends towards simplicity. They are not directed specifically towards being human-comprehensible the way that human designs are. I think the way that biological organisms are complex seem compatible with the tendencies of evolution towards simplicity and complexity. I do want to emphasize that as I understand the way people got a feel for the manner in which organisms are complex is mostly by observation are than reasoning through what evolution should create, and my statement about seeming compatible is a fairly weak one that could be said for many other ways which organisms could have appeared.

      • viVI_IViv says:

        I’d say there are two different kinds of complexity: “AND complexity” where you need many different parts which must be all present in order for the system to function, and “OR complexity” where you have multiple (partially or completely) redundant systems.

        Human engineers prefer simplicity, tolerate “AND complexity” when they can’t have simplicity, and try to avoid “OR complexity” except in the cases where there are strong redundancy requirements.

        Evolution instead tends to avoid “AND complexity” at the level of genes (it tolerates it at phenotype level) and accrues “OR complexity” essentially for free as long as each system of a redundant set individually provides even a small advantage.

        This means that for human-designed artifacts redundancy is in a trade off with design and manufacturing costs, in addition to space and energy trade offs, while in evolved organisms redundancy has only space and energy trade offs, while in terms of “design” cost it’s almost free.

    • Eponymous says:

      I don’t think that your argument that “the current size of the problem tells you nothing about the amount of resources invested in preventing the problem” makes sense. Evolution does not work without selective pressure. If parasites attempting to manipulate behavior barely affect an organism we should expect that species to remain at the same level of complexity until some parasite does start affecting its behavior. A hypothetical war between Saudi Arabia and Iran is different because humans are intelligent and can anticipate a potential risk without it ever happening, unlike natural selection.

      Came here to make the same comment. In fact, even positing high levels of parasitic interference in the past isn’t sufficient (unless it was the *very* recent past evolutionarily speaking) since complex biological machinery degrades fairly quickly in the absence of ongoing selective pressure.

      This is actually a good example to illustrate the advantages of intelligence over natural selection as an optimization process. Human brains can simulate the outcome of wars, and then abstract general rules about when war might be a good idea and unite this with other abstract reasoning. Evolution investigates hypotheticals by making organisms and seeing how many kids they have.

      Indeed there’s fairly simple math for some of this, to the degree that we can estimate the historical fitness costs of certain diseases by the frequency of costly defenses (like sickle cell).

  5. m1el says:

    Are there other systems that are intentionally complex/harmful so that they’re hard to fake?

    The first thing that comes to mind is social signalling.

    • Nornagest says:

      Address space randomization is the first thing that comes to mind.

      • Michael Watts says:

        ASLR means memory layout is intentionally secret, but it’s got nothing to do with being hard to fake. The point is to prevent the attacker from knowing two things:

        1. Where is the code I want to exploit?

        2. Where is the code I supplied to this process (in hopes of getting the process to call it)?

        Your bog-standard memory exploit works like this: the computer has a register devoted to storing the current line† number in the program it’s executing. (The “program counter”.) You want to write a new value in there so that the computer will suddenly start executing code from a different source, usually code that you’ve managed to write to memory yourself. ASLR is designed to prevent you from knowing what value to put in the program counter.

        †Not really a line.

        • Nornagest says:

          I was thinking of spoofing the program counter as analogous to spoofing a neurochemical signal. But it’s not a perfect analogy.

    • eric23 says:

      Cicadas emerge every 7, 13 or 17 years (always a prime number) to make it difficult for predators to match their cycle.

  6. anon1 says:

    Parasites cannot do pulses.

    Malaria is a famous counterexample.

    • janrandom says:

      Can you provide a reference please?

      • anon1 says:

        The cyclic quality of malaria is its most distinctive feature that sets it apart from other fevers, and this has been common knowledge since antiquity.

        * From https://en.wikipedia.org/wiki/Malaria#Signs_and_symptoms:

        The classic symptom of malaria is paroxysm—a cyclical occurrence of sudden coldness followed by shivering and then fever and sweating, occurring every two days (tertian fever) in P. vivax and P. ovale infections, and every three days (quartan fever) for P. malariae. P. falciparum infection can cause recurrent fever every 36–48 hours, or a less pronounced and almost continuous fever.

        * The commenter 1 minute after me made the same point in more detail (Plasmodium is the genus of parasites that cause malaria):

        Plasmodium famously has cycles in 24-hour multiples (1-3 days, depending on the species), which I believe is based on the parasite’s own circadian rhythm. But also, a parasite wouldn’t need their own clocks if they can clue into the host’s.

        * From https://en.wikipedia.org/wiki/Plasmodium_malariae:

        While found worldwide, [P. malariae] is a so-called “benign malaria” and is not nearly as dangerous as that produced by P. falciparum or P. vivax. It causes fevers that recur at approximately three-day intervals (a quartan fever), longer than the two-day (tertian) intervals of the other malarial parasites

        * In Of the Epidemics, Hippocrates clearly describes periodic fevers in which the cycle length relates to the severity of the disease. E.g:

        The least dangerous of all, and the mildest and most protracted, is the quartan, for it is not only such from itself, but it also carries off other great diseases.

        That last bit is a neat little detail I didn’t expect him to have known, and further raises my confidence that he’s talking about malaria: the high fever caused by malaria is an effective treatment for certain other diseases that are less resistant to heat. (LATE SYPHILIS—TREATMENT WITH A SINGLE STRAIN OF MALARIA) I have not heard of any other infection being used in this way.

        • cold_potato says:

          Okay, malaria parasites can co-ordinate to effect cycles with periods of a day or more. But discharging adrenaline into your bloodstream gives you a response within seconds. That’s much harder for a parasite to achieve.

          Going a step further, nerve impulses are driven by ion flows on timescales of milliseconds. Sure, we need that speed in order to achieve sub-second gross physical reaction times, but as a side effect it makes our axon-level nervous system impossible for a parasite to hijack.

          • deciusbrutus says:

            If you’re going to hijack a self-driving car and make it turn into oncoming traffic, you don’t try to add hydraulic fluid to one side of the power steering piston, you alter the computer or the computer’s input signals.

            Producing meaningful motion in a human takes something nearly as complicated as a motor cortex; simulating conditions that result in aggressive behavior is much simpler.

          • Phigment says:

            You probably can’t usefully control a human body with a parasite that’s much less complicated than brain.

            A parasite isn’t going to be nearly large or coordinated enough to directly hijack your nervous sytem and force you to go jump in a lake for its benefit. What it’s going to do is provide incentives to you that make you really want to go jump in the lake of what you perceive as your own free will.

            See Guinea Worms, for instance.

  7. thoramboinensis says:

    Parasites cannot do pulses.

    This parasites cannot communicate or coordinate with each other, so there’s no way for them to be producing lots of testosterone one minute and none at all the next.

    Color me skeptical on this particular argument.

    Plasmodium famously has cycles in 24-hour multiples (1-3 days, depending on the species), which I believe is based on the parasite’s own circadian rhythm. But also, a parasite wouldn’t need their own clocks if they can clue into the host’s.

    Also, quorum sensing a super cool and rapidly developing field of study in microbiology. Things as simple as bacteria absolutely do communicate and coordinate among themselves, essentially to figure out their density to adjust gene regulation accordingly. Not infrequently this takes the form of coordinated activation of some sort of collective behavior (everything from bioluminescence to virulence factor production). Often there are very important clinical outcomes that result. (There was even a paper a couple years ago showing that cancer cells quorum sense and that’s one of the triggers for metastisis.)

    Very interesting post!

    • eric23 says:

      It’s possible, but it’s an additional level of complexity the parasite needs to figure out.

      • spencer says:

        It’s not much complexity. Quorum sensing minimally needs only two genes (a transcription factor and an autoinducer), and working systems are common and easily adapted to fit the host.

        Plus, bacteria can “communicate” by all triggering of the same host events.

    • acymetric says:

      I could be wrong, but I think the pulses Scott was referring to are much faster than a 1-3 day cycle. On the scale of days, there would certainly be time for the parasites to do whatever they’re doing, detect that the environment has change sufficiently, and start doing the other thing that causes the other phase of the cycle. I wouldn’t call those “pulses” though.

      • Douglas Knight says:

        Malaria has a cycle of 2 or more days. Since the cycle is maintained, not smeared out, the resolution of coordination must be much smaller. The cold stage lasts 15-60 minutes, which probably means that they coordinate to a window of 60 minutes. It suggests that the more competent ones can coordinate a pulse of 15 minutes. That’s an order order of magnitude more than the hypothetical 1 minute, but it’s getting there.

        • acymetric says:

          This also assumes that the changes in the presentation of malaria is the result of changes in the “behavior” of the malaria and not changes in the way the body is responding to malaria doing the same thing the whole time. I have no idea which is the case, but it seems equally plausible that malaria is just doing its malaria thing, and the body responds by going “ok, time to get the chills…wait no that’s not right, do the other thing now”.

          • Michael Watts says:

            but it seems equally plausible that malaria is just doing its malaria thing, and the body responds by going “ok, time to get the chills…wait no that’s not right, do the other thing now”.

            The plausibility of that scenario is severely undermined by the fact that malaria has been especially distinctive in producing this effect for the last few thousand years.

            We already know, based on the rest of the world, that regular fever pulses don’t occur by chance. We also know that they don’t do much about the malaria. If they’re not an unplanned response and they’re not a planned response, you’re left with the option that they’re deliberately provoked.

  8. eric23 says:

    “Honestly it is a miracle anybody manages to stay alive at all.”

    “Some systems look heavily defended against parasite manipulation, but others don’t.”

    Maybe the ones that don’t look heavily defended are the ones where, for whatever complex reason, the defense would not be compatible with life? (Or at least too costly in terms of life)

    • DarkTigger says:

      Or were ther wasn’t an attack in the last couple of hundered million years.

      And every deadly parasite that high jacked that system died out with it’s terribly undefended host.

  9. fion says:

    There is zero problem with war between Iran and Saudi Arabia right now

    Somewhat unfortunate wording. When I first read this I thought you were saying war was acceptable, when what you actually meant was that war is very improbable.

    ETA: Typo: “This parasites cannot communicate or coordinate with each other” -> “These parasites”?

    “our immune systems are screwed up because we are not getting exposed to the parasites it was built to expect” also seems to switch from plural to singular

  10. viVI_IViv says:

    Del Giudice offers a seductive explanation: the perceived perversity of the human blueprint is absolutely real. Parts of it – the parts most involved in health and disease – were sculpted by evolution to be as hard as possible to understand or affect. This makes me feel better about how often the drugs I prescribe fail in surprising ways

    Possibly, but the simplest explanations for why your drugs often fail in surprising ways is that psychiatry is like trying to debug a computer program by taking out the magnetron of your microwave oven and waving it in front of your CPU.

    A serotonin spike on neuron X at time T has a different than more serotonin on neuron Y at time T+0.1s, for the same reason why a 1 on logic gate X at time T has a different effect than a 1 on logic gate Y at time T+0.1s. But drugs can only manipulate average neurotransmitter levels on the whole brain on a span of hours to days. Given how blunt the instruments of psychiatry are, it’s a miracle that it works at all.

    • noyann says:

      Given how blunt the instruments of psychiatry are, it’s a miracle that it works at all.
      They work more akin to under- or overvolting: “By applying a higher voltage to the devices in a circuit, … resulting in faster operation of the circuit and allowing for higher frequency operation.”

  11. gray says:

    …and if the mind is a new kind of parasite then… I’m not sure what follows except that all that evolutionary dodging and weaving failed spectacularly for us humans.

    It isn’t that hard for me to imagine. Our neural ‘platform’ seems rich territory for colonisation. The types of cunning strategies claimed for toxoplasmosis seem everywhere evident in ego self protection mechanisms. For example, we start an argument with ourselves or another, to enflame our egos, when life gets too calm. These parasitic, contagious, thought complexes have an interest in preventing us catching sight of the madness of our self-talk. Lest that we might then, through calm observation, start to dissolve those tangled knots of auto cannabilistic thinking.

    The weird part of this idea is that the thing having the idea may itself be that parasite.

    • noyann says:

      You’re hinting at Buddhism as a cure-all, except for itself.

    • FeepingCreature says:

      Right, so because you have indexical uncertainty you should behave in ways that both support the thing being colonized and the thing doing the colonizing. Not sure how that cashes out in practice though.

  12. benf says:

    “The strongest evidence against is the dog that didn’t bark. Some systems look heavily defended against parasite manipulation, but others don’t. Amphetamines raise dopamine effectively and without significant tolerance buildup (see part IV here for a defense of this claim); antipsychotics lower dopamine equally effectively and consistently. Since dopamine is one of the most lucrative systems for parasites to hijack, it’s surprising to find it so easy to affect.”

    This seems backwards to me. Neither of these groups of chemicals are abundant in nature – they’re specifically synthesized by human beings, and have been selected out of groups of hundreds of thousands of useless or totally toxic chemicals to serve that specific purpose BECAUSE they’re so much better than basically anything else ever discovered. This is like saying the skull isn’t very good at protecting the brain because a .45 slug can kill you – it took a long time and a lot of ingenuity to build that tool for that specific purpose.

    • viVI_IViv says:

      Amphetamines are chemically similar to the neurotransmitter phenethylamine, which according to Wikipedia is anti-microbial against certain bacteria, which may explain why parasites (at least bacteria) don’t synthesize it.

      Antipsychotics have a variety of side effects, but as far as I can’t tell nothing that could increase your probability of being eaten or engage in risky sexual behaviors (on the contrary, they tend to cause sexual dysfunction), hence they would probably not help parasites very much.

      Corticosteroids are a mystery though: cortisol is a natural hormone with a relatively simple structure, something that intuitively shouldn’t be too complicated for a parasite to synthesize, and it suppresses the immune system and increase blood glucose levels, which would naively both help parasites. Yet parasites don’t seem to have figured out this simple hack.

      • DarkTigger says:

        Have you ever seen the list of side effects of Cortison? Constantly high cortisol levels might be a good way to loose a host early.

        • deciusbrutus says:

          Having the result of a hormone hijacking be ‘died without being eaten by anything’ could very well be part of the pro-survival trait ‘lives in a population that cannot have cortisol systems hijacked by parasites to spread’.

      • viVI_IViv says:

        Corticosteroids are a mystery though: cortisol is a natural hormone with a relatively simple structure, something that intuitively shouldn’t be too complicated for a parasite to synthesize

        Now that I think about it, why don’t parasites synthesize testosterone? Like cortisol it’s a simple steroid, it increases aggression and libido, which should help parasites to spread.

        Maybe it’s because both cortisol and testosterone, and probably most hormones, are regulated by negative feedback loops: increasing the blood levels a bit just causes the body to compensate by reducing natural production, and parasites don’t have enough resources to produce enough hormones to overwhelm homeostasis.

    • Simon_Jester says:

      Indirectly secreting amphetamines in hopes that your host will start behaving recklessly isn’t a winning proposition anyway because the downside of “host gets killed by side effects of having meth-producing microbes in his gut all the time” is likely to outweigh the upside of “host behaves recklessly and has more risky sex and exposes himself to predators more often.” Indirectly secreting antipsychotics to reduce your host’s dopamine only helps if there’s some way for the parasite to benefit from the effects of reduced dopamine.

      I’m not sure how that would benefit a parasite.

      In general, an evolved system’s not necessarily unhackable, it’s just difficult enough to hack that no known parasite species has figured out an angle yet.

      • Garrett says:

        I seem to recall reading (but am currently able to source, so am perhaps mis-remembering) that people who were dying from AIDS during the AIDS crisis were seen to become more promiscuous just prior to death. There was speculation that this was part of the reproductive strategy of the virus.

  13. onyomi says:

    I’ve infected myself with necator americanus the past several years in an attempt to eliminate a gluten sensitivity. It didn’t do that but it seemed to reduce my overall tendency to inflammation (skin and eyes seem less prone to be red and puffy, for example) and also to have had a mildly positive effect on my mental health (my subjective appraisal is that inflammation and a tendency I have to anxiety and OCD are linked).

    • noyann says:

      Worms are known to put a damper on inflammatory reactions, and modify allergies.

      skin and eyes … red and puffy That could be allergic instead of / together with gluten sensitivity.

      I have noted (N=1, high placebo confounder!) that my mood, anxiety, depressive emptiness were much better the day after taking NSAIDs. Also, swallowing mometasone soon after / together with an allergic food also improved them. As it is metabolized in the liver (1st pass effect), the inflammatory reaction must be in the upper digestive tract, but its effects reach across the blood-brain barrier.

      There is definitely a connection, not just to depression/anxiety. “A knock to the immune system in early life might trigger a life-long increased immune reactivity, and infections and autoimmune disorders are now known to be risk factors for development of schizophrenia and [major depression]. “ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5442364/

      Causation may go both ways: early trauma keeps the body, too, permanently prepared for the worst, ie, on high defense / repair alert. Other things do, too: “A range of factors appear to increase the risk for the development of depression, and seem to be associated with systemic inflammation; these include psychosocial stressors, poor diet, physical inactivity, obesity, smoking, altered gut permeability, atopy, dental cares, sleep and vitamin D deficiency.” https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3846682/

    • FLWAB says:

      A bit off topic, but I find helminthic therapy to be fascinating, and I have a lot of questions about it. I don’t know if you would know the answer, but do you know why hookworms seem to be the favored treatment over pinworms? From what I can tell hookworms can potentially cause iron deficiency and fatigue, while pinworms at their worst cause itching. So why not pinworms? Are they not bad enough dudes to temper an overactive immune system?

      • throwaway534321 says:

        From what I can tell hookworms can potentially cause iron deficiency and fatigue, while pinworms at their worst cause itching.

        (Throwaway account for obvious reasons.) I had multiple pinworm infections as a child and you are severely underestimating how much the itching affects quality of life. I would 1000% prefer hookworms and anemia/fatigue/dizziness/muscle weakness.

        It follows a circadian cycle and is relatively calm during the day, but towards evening it takes intense effort not to be scratching your ass every couple of minutes. This wasn’t just unpleasant, it had social consequences. The itching would continue to worsen until midnight or so, and could easily have been costing me an hour of sleep every night though I wasn’t tracking it at the time. So you don’t even avoid fatigue with pinworms, it just happens through a different, probably far more unpleasant, route.

      • onyomi says:

        I’m happy to answer any questions about my personal experiences with helminthic therapy (have only used necator americanus aka New World hookworms). There is also an extensive wiki with lots of info and other personal experiences.

        As for the question below, I don’t know whether pinworms would have a similar immune modulating effect, but I can say that the side effects of hookworms for me are largely limited to the initial inoculation rash, which amounts to a couple of weeks of localized itching at worst. Other than that, I experience no negative side effects and a number of positive effects.

        I think NA are popular due to their favorable systemic benefit vs potential risk profile. People with colon-localized problems like IB also sometimes use whipworm, which I’ve considered trying, but apparently their effect on systemic inflammation is less. My Eurasian ancestors probably had not NA but a larger, scarier hookworm called ancylostoma duodenale, but the necator americanus is smaller and apparently unable to migrate elsewhere in the body, which I believe the AD can sometimes do. It also has the benefit of lasting a while in people who don’t have very over-active immune systems but also being unable to complete a reproductive life cycle within the human body, meaning there’s no risk of a colony growing out of control (assuming you don’t poop in your own garden and then walk around barefoot in it).

        I also suspect that, as a first world man (in other words, no periods) who eats red meat, having a bit of blood sucked out of me on a regular basis may be, if anything, beneficial.

  14. TJIC says:

    Excellent essay, as usual.

    In a bit of synchronicity, I delivered (well, second hand) a speech on the topic of parasitism and an organism’s defenses against it just yesterday (referencing this blog, no less!), and it may be of interest to your readers.

    http://lfs.org/blog/travis-corcoran-accepts-prometheus-award/

  15. Radu Floricica says:

    Everything seems bogged down in a million different opaque signaling cascades that fight off any effort to untangle or shift them.

    A book I’m reading now blames this on ethics. Apparently it’s ethically acceptable to use drugs to change behavior, but absolutely unacceptable to touch neurons directly. Change that, and it takes less than a decade to both understand and cure half of psychiatric conditions. It kinda makes sense. I do get it’s scary, but I think it makes sense.

    • JPNunez says:

      Is this really like that? It’s not like psychiatry has shied away from just blatantly cutting the brain of people in half, trying to stop, say, seizures. Elon Musk is trying to get brain/computers interfaces going, and some of the objections are about how hard this is to do, which tells me people have tried to do this before and encountered problems. Laboratories are trying to grow human neuron cells artificially.

      I don’t disagree that there are ethical concerns, but it seems to me people are doing bold experiments with brain matter and we aren’t really minding it too much.

      • Radu Floricica says:

        The specific context was that we know enough about brain architecture to cure obesity with direct intervention. I don’t see that kind of thing going through an ethics committee any time soon.

        • JPNunez says:

          Aaah.

          Ok I can see the point.

          But if at some point, say, Elon Musk develops a more successful brain/computer interface, maybe the risk of directly touching the neurons that are driving, say, eating disorders may look less risky.

          • Radu Floricica says:

            Not sure which is more scary – cutting open the brain to physically touch a neuron. Or building a headset which allow anybody with a laptop experiment with his own brain – or others’.

            But I don’t think trying to keep the genie in the bottle is the correct solution either way, especially when it’s such an useful genie.

      • John Schilling says:

        It’s not like psychiatry has shied away from just blatantly cutting the brain of people in half, trying to stop, say, seizures.

        I’m pretty sure psychiatry has shied away from this, for every definition of “shied away from” this side of “100.00% absolutely stopped”.

        • aiju says:

          Are you thinking of lobotomies? Because he’s talking about hemispherectomies which I think are still a thing in certain cases of severe, otherwise untreatable epilepsy.

          • John Schilling says:

            Yes. They are a thing that psychiatrists have not 100.00% absolutely stopped doing. They are also a thing that psychiatrists have shied away from doing. If you say “we still do X in certain severe cases”, that’s pretty much the definition of “shying away from” X.

          • JPNunez says:

            They shied away from hemispherectomies because of unfortunate side effects and dangers, not because of any phantom ethic system about “not wanting to touch the neurons”.

            The procedure has been refined and perfected. It’s pretty rare, yes, but it is still accepted.

            Lobotomies I will give you, but I was yeah, talking about epilepsy and cutting an hemisphere off.

          • Radu Floricica says:

            @JPNunes

            phantom ethic system about “not wanting to touch the neurons”.

            I don’t really know how we could settle this, but intuitively, to me at least, this kind of ethic system seems to be obvious – the percentage of cases that end up with actually touching a neuron has to be absolutely insignificant. Everybody knows somebody that went to therapy, if not yet everybody went to therapy themselves, but doing anything non-drug related to the brain is so rare as to appear in the domain of fantasy. Even if we’re generous and include less direct means like electric or magnetic stimulation.

            Anything that could potentially touch just a cluster of neurons is… no, I take it back, it is the actual domain of fantasy. We only read about it in science-fiction and scary reporting about chinese research.

          • aiju says:

            Yes. They are a thing that psychiatrists have not 100.00% absolutely stopped doing. They are also a thing that psychiatrists have shied away from doing. If you say “we still do X in certain severe cases”, that’s pretty much the definition of “shying away from” X.

            Sorry, I had problems parsing the second half of your sentence.

            Everybody knows somebody that went to therapy, if not yet everybody went to therapy themselves, but doing anything non-drug related to the brain is so rare as to appear in the domain of fantasy. Even if we’re generous and include less direct means like electric or magnetic stimulation.

            The obvious thing to mention is ECT which is not exactly common but I know a few people that had it done.

            Deep brain stimulation is also a thing which exists.
            I think while people have obvious reasons for shying away from brain surgery on healthy people, once you are sufficiently ill, there are no deep ethical concerns holding people back from doing whatever is possible to your brain.

    • broblawsky says:

      I feel like having strong ethical limitations on radical, experimental brain surgery is probably a net positive?

      • Radu Floricica says:

        Don’t have a strong opinion either way, but I have to comment that while having those strong ethical limitations, “radical, experimental” brain surgery is going to stay radical and experimental instead of becoming routine like in any other area.

    • eyeballfrog says:

      I feel like the ethical problem here is with killing patients or leaving them with severe brain damage, not with touching the neurons per se. Brain surgery to remove a tumor (which can cause all sorts of neurological diseases) is considered 100% fine even though it necessitates touching neurons. It’s just that until recently, removing a brain tumor left you with a dead patient.

      As far as experimental procedures, this is all in line with medical practice in other areas. Vivisection would give a lot of excellent info about anatomy, but we still don’t do it for the obvious reasons.

    • Protagoras says:

      I have a friend who’s a neuroscientist; his dissertation involved research using surgically implanted sensors to directly monitor what was going on in chimp brains. This was fairly recent, and that fairly recent research definitely seemed to me to be more about trying to figure out how to get any useful results at all by this method, as opposed to discovering anything specific. I think you are underestimating how much effort is going into touching neurons direction, and drastically overestimating how quickly such research could produce results (if it was that easy, they would have already done it, as they are in fact trying to do it).

    • Doctor Mist says:

      Radu Floricica-

      Pointer to the book?

  16. Loris says:

    There’s a sense that science is stagnating, and biology is one of the worst offenders. […] But for the past fifty years, it’s been kind of a mess. Despite some amazing work by amazing people, we still don’t even understand questions as basic as what depression is.

    You’re kind of lumping all of biology in with psychiatry’s failings here.
    Due to improvements in technology and technique, most of biology has been undergoing major development over the last few decades.
    That difference is probably at least in part down to the difficulty of doing even simple non-invasive experiments.

  17. noyann says:

    Only colonies of thousands or millions of parasites can produce enough chemicals to affect host signaling. This parasites cannot communicate or coordinate with each other, so there’s no way for them to be producing lots of testosterone one minute and none at all the next.

    Maybe not on a timescale of minutes — but gut flora adapts from one meal to the next, if necessary.

    Two other points to consider:
    Many (most?) gut flora bacteria are yet unknown, because they are so highly specialized (requiring such a uniquely tailored environment) that they simply won’t reproduce when isolated from it and won’t show up in a Petri dish. It’s not very likely that they won’t exert some control over this environment.

    And they do communicate: Biofilms” are not just bacterial slime layers but biological systems; the bacteria organize themselves into a coordinated functional community. “ (emphasis mine) (…Sounds like the blurb on a paperback describing the backdrop-cum-protagonist of a bio-sci-fi thriller, Peter Watts style.)
    “Various signalling events including two-component signalling, extra cytoplasmic function and quorum sensing are involved in the formation of biofilms. “ https://www.ncbi.nlm.nih.gov/pubmed/26377585
    Pack them tighter, and the sewer sludge bug lump will have enough internal signal speed and bandwidth to gain consciousness to reflect on its condition.

  18. boylermaker says:

    Great overview.

    A couple of technical points:

    Del Giudice wonders if parasite-host arms races created pressure for increased human variability.

    You have to be careful with this one. “Variability” is a population-level trait, and selection (~)only works on individual level traits. Now you can have selection on “being a different genotype than the majority”, but that doesn’t necessarily lead to highly variable traits at the level of the MHC. I would want to see some modelling before I was too convinced by this point.

    Fifth, you let the parasites become part of the furniture. If everybody in your ecosystem is infected with a parasite that raises serotonin, you just evolve a tonically lower serotonin level, and then it all cancels out. This one seems a little bit weird to me – surely this isn’t the stable equilibrium?

    Check out Wolbachia, an intracellular bacterial parasite. The most famous and most weird example is in pillbugs. The bacteria causes genetically male pillbugs to develop into females (Wolbachia is only transmitted via eggs, so its host developing into a male does it no good), but some pillbugs have managed to domesticate this process. They are all genetically male, but have evolved the ability to control which of their offspring have a Wolbachia infection. Infected offspring become females; uninfected ones males. In effect, the bacterium is the sex chromosome. https://evolbiol.peercommunityin.org/public/rec?id=42

    • Randy M says:

      That’s fascinating.

    • baconbits9 says:

      To risk a semantic discussion

      You have to be careful with this one. “Variability” is a population-level trait, and selection (~)only works on individual level traits. Now you can have selection on “being a different genotype than the majority”, but that doesn’t necessarily lead to highly variable traits at the level of the MHC. I would want to see some modelling before I was too convinced by this point.

      Sexual reproduction is typically viewed as being beneficial due to its ability to increase variation, so you can clearly have strategies that rely on variability but also effect individual level traits.

      • boylermaker says:

        Sure. But the important thing to remember is that an individual needs to benefit at each step of the way. Returning to the OP example, this doesn’t work, because individuals are paying a cost which only the population benefits from:

        Scenario 1: Everybody has a serotonin response that is very useful in neuroregulation, but hackable. Some of the population evolves a different, worse serotonin response, lowering their fitness. However, everybody wins, because now parasites much evolve some sort of costly mechanism to detect and handle different serotonin responses, and this lowers the total parasitism pressure on everybody.

        The population may be protected, but the new serotonin response is harmful to the individuals expressing it, and so we expect it to go extinct regardless of how much everyone else is benefiting from its presence in the population.

        Whereas this scenario works, evolutionarily:

        Scenario 2: Everybody has a serotonin response that is very useful in neuroregulation, but hackable. Some of the population evolve a different, worse serotonin response, lowering their fitness in the absence of the parasite. However, they are also protected from the parasite, which increases their fitness more than their unusual serotonin response decreases it.

        So if Scott/Guidice are proposing a Scenario 2, that check out in terms of natural selection. But it doesn’t obviously lead to a variable population in the long term. If the new serotonin response is a fitness advantage, then we might expect it to spread through the population until everyone has it (back to zero variation). Or we might expect that it will spread until it is so common that the parasites evolve countermeasures, pushing its prevalence back down. That’s a situation that will lead to increased population variation.

        The problem is that it’s not at all obvious from a handwavy thought experiment like this whether Scenario 2 will lead to more long-term variation or not. You need to have a pretty good model that includes some (maybe a lot of) information about the dynamics of the parasite and host.

        So while it’s fairly straightforward to say things like “Individual humans have [some trait] because parasites can’t attack people with [some trait]”, it’s not at all straightforward to claim “Human populations show variation in [some trait] because parasites can’t attack people with [some trait] and/or can’t attack populations with variation in [some trait].”

        • Nornagest says:

          But it doesn’t obviously lead to a variable population in the long term. If the new serotonin response is a fitness advantage, then we might expect it to spread through the population until everyone has it (back to zero variation). Or we might expect that it will spread until it is so common that the parasites evolve countermeasures, pushing its prevalence back down.

          The fitness advantage from parasite resistance isn’t a constant. It depends on the prevalence of parasites in the environment, and that prevalence also depends on how many non-resistant individuals they have to parasitize (especially if they’re specialized to a single host species, which parasites often are). This gives you a feedback system that looks a lot like predator/prey population dynamics — depending on details there may or may not be a stable equilibrium, but we wouldn’t expect it to go straight to zero.

          • boylermaker says:

            Well, that’s ONE highly-parameterized model! You’re right that if we have the right parameters for our model, we could get dynamics where there is a stable equilibrium that isn’t 0 or 1. But it’s not at all inevitable, which is my point: there are plenty of scenarios where we DO expect variation to be rapidly eliminated by selection.

        • syrrim says:

          In scenario 2, a worse serotonin response aids the individual because it is uncommon. As it becomes more common, it helps the marginal individual less and less. At a certain distribution within the population, it would neither help nor hurt an individual to adopt it, at which distribution we would expect it to remain.

          • boylermaker says:

            Maybe, maybe not. How are you arriving at the idea that it helps the marginal individual less and less? Counter-evolution by the parasite? Reduction in transmission probabilities as the anti-parasite genotype increases in the population?

            These are all plausible, but definitely not given, which was my point–not that you can’t get increased variation at the population level, but that there are a lot of scenarios in both ways, so I’d want to see some boots-on-the-ground research to come up with specific models and model parameters before I thought that it was likely (as opposed to just plausible) that parasite pressure promotes variability in human populations.

  19. Deiseach says:

    I come from a country background so it’s entirely possible nearer ancestors than Nth great-grandmother were exposed to liver fluke; farming/working as farm labour/being exposed to farm yards is a great way to be exposed to all kinds of fun diseases and parasites: brucellosis (which could leave you with recurrent fevers and general debility), bovine TB, Q fever, farmer’s lung and many others.

    But why is the body prepared to suddenly have all its serotonin reuptake transporters inhibited? Is that something that frequently happens, out in nature?

    Well, think of faith healers – the traditional way, in Irish folklore, to tell if that seventh son of a seventh son can really cure is to see if they can cure ringworm. People said to have “the cure”, that was a hereditary ability in certain families, for specific ailments often involved rubbing their blood onto the affected area/person. Worm infestations (roundworms, threadworms, tapeworms) used to be quite serious (and are still around today, if not so much because of improved hygiene). I’d say that up to fifty years ago nearly everyone had a good chance of coming into contact with some source of infection, and that even today in certain circumstances that possibility remains even in Western First World countries.

    • broblawsky says:

      Is it possible to pass on antibodies through a blood transfusion? Because if true, that lends an interesting dimension to this type of folk cures.

      • Lambert says:

        You don’t want to pass on the antibodies, but rather the antigens that the antibodies bind to.
        Which is basically how vaccines work.

  20. deciusbrutus says:

    It seems on priors to be plausible to me that e.g. serotonin, or some step in serotonin production, itself has serotonin-blocking effects over time, and there is some other process that controls the rate of serotonin unblocking. Any parasite attempting to manipulate serotonin would need to develop both mechanisms simultaneously, and variation between people could even deny parasites a victory condition at all.

    Modern pharmacology would suffer exactly the same problems as parasites would, for exactly the same reasons.

  21. Randy M says:

    Scott, fascinating post, thanks for bringing these ideas to our attention.
    But, you say

    This parasites cannot communicate or coordinate with each other, so there’s no way for them to be producing lots of testosterone one minute and none at all the next.

    This in a work about how a parasite can control a host, you can’t think of a way for them to coordinate?
    Plus, since I doubt a single instance of this parasite can secrete enough neurotransmitter to significantly alter host behavior, we can assume they are already evolved to ‘work together’ in a sense. Is it so hard to conceive of feed back mechanisms they have to coordinate surges of chemical release? Say, a low level continual release, then if a threshold is reached, drastically increase production. If another threshold is reached, go dormant for a period of time.

    Now I grant you that an organism with more complex organization will be able to have much more complicated systems of communication and it’s signals harder to mimic precisely. I buy the arms race analogy. But an organism that demonstrably communicates with the host ought to be able to communicate among itself to, at some simple level.

    edit: Not exactly an original point.

    • sclmlw says:

      I read that part of the original post and thought mostly the same thing. It would have made my old microbial physiology professor cringe. He would often say things like, “so-called lesser organisms” when talking about microbes.

      Yes, microbes do coordinate with each other. Yes, they have complicated interactions, where they signal together to manipulate their environments. Just because they’re all the same genetically doesn’t mean they’re all the same within the population. It’s not just multi-cellular organisms that coordinate behavior, and making that assumption ignores a lot of what any one single-celled organism is doing. The argument that microbes can’t coordinate pulsed signals is unconvincing to me.

      • Simon_Jester says:

        Remember that the point here is to create a security system highly likely to make microbes not evolve the tools necessary to break it.

        If the only way to hack a given species’ brain is with not just neurotransmitter-secreting parasites, but coordinated, pulsed secreting parasites, then for such parasites to evolve, they first have to evolve an ineffectual neurotransmitter-secreting mechanism, then separately evolve a mechanism to make it pulse.

        This provides, at least in relatively short evolutionary timescales, considerable security. Eventually something might crack a way past it, sure, but then on those timescales larger organisms can evolve in response too, and come up with even weirder and more perverse neurotransmitter responses as necessary.

        • sclmlw says:

          Yes, but you don’t get to evolve in isolation. While you’re busy crafting that difficult to crack system, it’s being actively thwarted. This is one of the fundamental problems with stepwise evolutionary theory at a molecular level. How do you create a complex system like Scott describes in a competitive environment?

  22. alephnaut says:

    You may want to know that there’s a missing period at the very end of the article.

    Great post by the way.

  23. sclmlw says:

    The description of MHC was highly inaccurate. Every person doesn’t have a completely unique combination of MHC genes, just because its unlikely they have the same MHC as their sibling. This is kind of like names, where few people have truly unique names, but it’s unlikely for two randomly selected individuals to share the same name.

    The rest of this comment is for people interested in knowing why MHC is variable, and as well as why it’s not true that this variability is not the same as everyone having a completely unique MHC.

    Most people know that antibodies are able to detect pathogens in the blood and target them for destruction. MHC is part of a system that not only detects things outside of cells, but can reach into cells as well. It works in two different ways, with two different types of MHC, but their function is similar in both parts of the system.

    MHC II is used by macrophages and dendritic cells. They go around eating everything in sight. They then chop all that up and present the pieces on their cell surface. MHC II is the molecule they use to present with, as it acts like a hand holding out the various chunks of whatever the dendritic cell happened to be eating that day. Different T-cells come around looking for matches to their specific T-cell receptor, checking the loaded-up MHC II molecules. When they find a match the dendritic cell tells the macrophage what context it found that piece of protein in. For example, it could be, “this is just garbage, don’t worry about it”, or it could be, “there were a lot of viral signals around, and a bunch of cell death – if you see this again, take it out before it does damage”.

    MHC I is found on all cells. Every cell is constantly breaking down the proteins and other stuff inside it in a process of continual turnover. Some of this cellular garbage attaches to MHC I, which is displayed out on the cell surface and forms a kind of catalog of what’s inside the cell. A T-cell can then wander by, like a cop checking cars at a roadblock for suspicious material. Say the T-cell from the example above (that was told a certain piece of protein is viral) stumbles across that same protein again on the surface of a normal-looking cell. Knowing this is a viral protein, it will kill the cell before the virus has a chance to make a bunch of copies of itself.

    You might assume the best strategy for a parasite/virus would be to simply downregulate MHC or change it so it can’t pick up proteins, in order to mask its presence. But there’s another cell called a Natural Killer cell that goes around checking to see if any cells don’t have your own body’s specific MHC on its surface. Any cell with altered or absent MHC is immediately killed.

    This is why organ transplant matching is so difficult. If you take an organ from someone with one kind of MHC and put it in the organ of someone with a different kind, the Natural Killer cells will target it and destroy it. The probability that you match a sibling is not high, since there are so many genes that contribute to your specific MHC, but it’s higher than the general population. It’s like taking two people with average-length names and creating a new name by randomly selecting letters from the mix. The likelihood you’ll end up with the same name for two siblings is low, but not impossible. Do this over a few million iterations across society, and you have high variability, even though lots of people are going to have the same combinations.

    How does this all relate to variability? Not every MHC molecule can pick up every piece of protein. Some short peptides will be picked up by one type of MHC, and others will be picked up by another. Humans have a variety of different types of MHC, just like we have a variety of versions of lots of different genes, but the point isn’t variability for its own sake. You want a broad array of MHC so you can pick up a larger variety of peptides to present on the cell surface. In order to ensure you can present protein from each pathogen you encounter you need more than one MHC gene (HLA), so each person carries a small collection around with them. Even so, any one person’s complement of MHC won’t be able to pick up all possible peptides, so some people will be susceptible to infection of certain pathogens that others wouldn’t be and vice-versa. At a population level every peptide is covered. This ensures that pathogens that can’t be detected by people with one collection of MHC will be detected by people with a different type, so no one pathogen can spread through the population and wipe us all out.

    There’s a hypothesis that a lack of MHC variability is why so many American Indians died off (95-98%) when exposed to European diseases, but not as many Europeans died off in the exchange. Europeans had a much higher diversity of MHC, so new diseases that killed one person might be fought off by another person (whose immune system could easily recognize the pathogen). Because the Americans mostly came from a small founder population, their MHC were much more similar to one another, and any disease that was fatal to one person was very likely to kill everyone else.

    • Scott Alexander says:

      Thanks, I’ve slightly edited the post.

      • sclmlw says:

        Thanks! It was probably fine for a layperson’s understanding. I’m still not sure MHC is what you’re going for here, though. The point of MHC variability isn’t to protect your body from stray foreign human cells that might infect you, since organ transplantation isn’t an evolutionary pressure. The point is to have a set of genes capable of picking up different peptides. The transplant headache is just an accidental side effect. I guess you could call it a variability strategy against pathogens, but I’d probably use a different example. (Don’t change the post again, though.)

        I’m not arguing against human variability here. I think biological mechanisms drive toward genetic diversity as a robust survival strategy. We like things orderly, but biology likes things messy – just in case.

        Interestingly, there’s a sub-type of T-cell called a gamma-delta T-cell. It doesn’t require activation by recognizing peptides bound to MHC, and is believed to recognize pathogenic lipid motifs, among other things. (There’s always another subset of T-cells.)

  24. Another Throw says:

    Key takeaway: bring back leeching.

  25. uau says:

    Del Giudice offers a seductive explanation: the perceived perversity of the human blueprint is absolutely real. Parts of it – the parts most involved in health and disease – were sculpted by evolution to be as hard as possible to understand or affect.

    I think the “to understand” part is wrong. Parasites do not work by conscious understanding and analysis. It could be that the kind of features that make it hard for parasite evolution to find ways to manipulate a biological system also tend to make it harder for human intelligence to analyze, but that is not a direct effect.

    • acymetric says:

      I’m pretty sure it means “to understand” in a biological or evolutionary sense, as shorthand for the longer explanation that you just gave which seems reasonable and easy to understand to me.

      • uau says:

        I disagree. It essentially says “Why is this hard to understand? Because it evolved to be hard to understand!”. The first “understand” clearly refers to human understanding. If you now say that the second refers to a very different thing, the argument turns into a misleading pun at best.

        • Doctor Mist says:

          Yeah, but the pun is pretty fundamental. The astonishing thing about evolution, and part of what feeds resistance to the idea even now, is that it grossly really does seem to mimic a designer. Dennett expresses this by saying something to the effect that evolution displays creativity and invention — just not conscious creativity and invention, and folding “understanding” into that formulation doesn’t seem inappropriate.

          Yes, the understanding displayed by evolution (and by proxy the parasites) is not identical to the understanding Scott strives for as a doctor, but could well be legitimately analogous enough that we shouldn’t be surprised if obstacles to one are obstacles to the other.

  26. Douglas Knight says:

    in case you had the same question I did

    Whether this del Giudice is the same guy you wrote about before?

  27. Douglas Knight says:

    Del Giudice obliquely cites Greg Cochran’s controversial hypothesis that homosexuality may be related to parasites hijacking sexual machinery.

    Maybe it’s oblique because del Giudice isn’t actually doing it? The Cochran-Ewald-Cochran paper talks about a lot of things, not just homosexuality. In particular, it talks about something very relevant to the citation: why do STIs cause infertility? Maybe it’s because that’s where they enter; they do what they must to get in. But they propose that infertility is way of encouraging breakup and thus new partners.

    In any event, that isn’t Cochran’s hypothesis. His hypothesis is that homosexuality and schizophrenia rare side effects of common infection, not the main course of the disease and not hijacking, not selection on the disease. Narcolepsy is caused by infection, often Epstein-Barr, but 90% of people get infected by Epstein-Barr without getting narcolepsy, or even mono. If EB can cause narcolepsy without even trying, then it (or something else) could just as easily cause schizophrenia or homosexuality.

    • Scott Alexander says:

      Maybe, but why wouldn’t STIs cause a condition that leads to more sexual partners and riskier methods of sex? Seems like a pretty natural match for the theory.

      • Douglas Knight says:

        First of all, you should distinguish the object-level argument from the claim about who endorses and insinuates which theories. Cochran makes an interesting argument which is drowned out by people rounding off his argument to a different one. And then people declare that dangerous and suppress it and then start seeing dog whistles. Also, ignoring that his paper claims lots of other things. And by seeing a dog whistle in del Giudice, you obscure what he’s saying.

        ———

        Your new question is the converse. In fact, we don’t see STIs causing promiscuity, except through the small, long-term effect of sterility. You may predict that gonorrhea causes promiscuity, but we don’t observe it, not to an easily measurable degree. Maybe it tries, but it’s not easy for the parasite to achieve. We do see rabies causing promiscuity. We do see it in some diseases in other animals.

        On the contrary, as discussed in the CEC paper and, I think, in earlier papers by Ewald, gonorrhea and other STIs seem evolved for low virulence, long time between partners, not for promiscuity. We see syphilis evolve toward low virulence when it crosses from skin disease to STI around 1500 (which is the consensus, regardless of whether it’s from the Old World or New).

        If there’s already a pool of promiscuous gay men, then an infection causing homosexuality helps it spread. But if the main cause of homosexuality is the infection, restricting partners to the already infected makes it harder to spread to new hosts! (Also, note that bi is rarer than gay.) Moreover, since gay men always knew their orientation, they were probably infected by age 5. You could posit that it spread by child rape, but that is a very different hypothesis from adult promiscuity. This hypothetical is very far from reality. Very few gay men report being molested that young and most of their preferences don’t point in this direction. Much more likely is that it’s something that endemic at age 5, like EB or the common cold.

  28. broblawsky says:

    There’s a sense that science is stagnating, and biology is one of the worst offenders. In the 1800s and early 1900s, we were pinning down our mastery of anatomy, discovering all the major hormone systems, learning about microbes and inventing antibiotics. It seemed like the same kind of thing as physics, where you could go out into the world, observe things, and make difficult but fundamentally straightforward discoveries. But for the past fifty years, it’s been kind of a mess. Despite some amazing work by amazing people, we still don’t even understand questions as basic as what depression is. Everything seems bogged down in a million different opaque signaling cascades that fight off any effort to untangle or shift them.

    I guess the question with hypotheses like Del Giudice’s is, how can we test them? Without a testable hypothesis, world-views like this are no more than “just-so” stories like evolutionary psychology – frameworks that can justify any given phenomenon, but not in a way that’s amenable to scientific investigation.

  29. zzzzort says:

    Finally, almost nothing eats humans

    I’d always assumed that burial was a social adaptation to reduce the spread of disease in a community, but never realized that by preventing other organisms in the same biosphere from eating the corpse could also have important effects. As humans are generally very committed to burying/burning dead people, a putative parasite would need to get the host killed in such a way that no one could find them, but where the parasite’s offspring could still find its way back to humans.

    I wonder if cultures with sky burial traditions (some tibetans and zoroastrians) had any issues with parasite manipulation. Or maybe a high prevalence of vultures that can act as parasite sinks.

  30. If the brain and associated hormone systems have a lot of unnecessary complexity built into them to guard against hijacking, then couldn’t it be the case that artificial general intelligence will turn out to be much easier than foreseen, so long as we continue doing it our own way, rather than through copying?

    This is unfortunate, because copying the human brain (and associated systems) is probably the much safer way to get general intelligence than fiddling around with algorithms that evolve based on feedback. Tech companies are going to beat things like the Human Brain Project, and that’s bad, and if we can find a consistent way to work out which parts of the system are there to flummox microbes we could change that.

    • broblawsky says:

      That assumes that general intelligence isn’t a byproduct of complexity, rather than a casualty of it. I don’t think we know enough to be sure either way yet.

  31. John Schilling says:

    If parasites want anything from us, it’s probably STIs wishing we had more risky sex;

    Consequently, to the extent that this hypothesis is true, both love potions and “female Viagra” are going to be very hard problems. Male Viagra only had to deal with the relatively simple biomechanical problem of facilitating an erection.

    • Scott Alexander says:

      Seems like another failed prediction; doesn’t testosterone alone make men hornier?

      Actually, I know it’s supposed to do this for testosterone-deficient people, but I don’t know whether huge overdoses of testosterone would make people arbitrarily horny. Sounds like someone should test this for science.

      • Psycicle says:

        I don’t know whether testosterone does that, but anecdotal evidence confirms that huge overdoses of progesterone make people arbitrarily horny (and also incredibly sleepy).

      • John Schilling says:

        There’s an asymmetry in that the default state for men is horny enough that “insufficient horniness” usually isn’t the major roadblock to their having more sex. And it’s going to take a very clever parasite to significantly enhance the host’s Game (or stamina). You’re right that there might be a niche for upping the sex drive of low-testosterone men, but insofar as we are defining “low testosterone” relative to something presumably close to the evolutionary optimum, any parasite targeting that niche would more properly be a symbiont and there would be no evolutionary pressure for defenses against it.

        With women, higher sex drive translates more readily to more sex. If it were as simple as “hormone X makes women horny”, then an STI that produces that hormone would have a substantial evolutionary advantage and we’d expect human evolution to make human female brains resistant to it. And so we find that, while testosterone may somewhat boost female libido the effect is relatively small. If psycicle is right, progesterone may substantially boost female libido but make them too sleepy to do anything about it. Hardly proof, but consistent with the prediction that women will have preferentially evolved defenses against a simple chemical hack greatly increasing their sex drive.

  32. Kuiperdolin says:

    Since everyone’s MHC is different, pathogens can’t just evolve to mimic the MHC; they would have to undergo an entire evolutionary process for each new host they invade.

    Can’t they, though ? Some bacteria have a dozen generations a day or more. If a population sticks around in a human host for a week or so (not a particularly tall order), it sounds plenty enough to start evolving adaptations to the local environment, including the host’s MHC.

    Of course the point still stands for parasites that have to jump from host to host each generation.

  33. arch1 says:

    The OP seems to make a lot of avoiding scrutability. But isn’t it really “controllability by an evolved parasite in ways that enhance its genes’ reproduction” (where.the “evolved” part gets unpacked to talk about a sequence of feasibly-small random variations of the parasite’s genome each of which achieves that same enhancement thing) which is to be avoided? If so, that seems different from (though maybe correlated with) scrutability.

    • beleester says:

      Controllability correlates with scrutability because you need controllability to run an experiment. If you get a different result every time you run the same test, you won’t learn much. That’s true whether “running the test” means injecting a chemical into the cell, or inserting a gene into a parasite that secretes that chemical.

  34. nick1c says:

    n the same vein, one thing I have always found odd is the vulnerable nature of the thyroid gland.
    It controls the entire metabolic rate of the body. However, it is extremely vulnerable to auto-immune disease, Hashimoto thyroiditis is by far the most common endocrinopathy. In addition, it requires Iodine, leaving thyroid hormone production dependent on exogenous intake and making congenital iodine deficiency syndrome (Cretinism) a far too common cause of mental retardation.

    Prima facie, there is no need for this.
    Thyroid hormone could simply be a short chain polypeptide like almost all hormones in the body and avoid (almost) both of these problems.

    He mentions that the human body avoided simple on-off switches as they are vulnerable to manipulation, but the thyroid hormone is such a switch. Hyperthyroidism can lead to a restless, manic and agitated state. A parasite that could produce T3-T4 could affect human behavior with a single molecule. The lack of available Iodine would prevent a parasite from producing too much thyroid hormone,

    In addition, hypothyroidism makes people lethargic and even comatose. One could imagine a parasitic life cycle that destroyed the thyroid gland, making the human easy prey for a tiger (akin to the rat-cat life cycle for toxoplasmosis). Again, the particularity of the gland seems to prevent this. For one, Iodine is an anti-septic and would prevent any parasite from getting too close. For another, the active molecules T3, with a half life of one day, is peripherally converted by the body from T4, with a half life of one week. So even if the gland got instantly destroyed, it would take a good month before stores ran out.

    Any thoughts?

    • stopandgo says:

      Smartest comment on here.

      It’s no coincidence iodine is extraordinary anti-septic and kills just about anything in even tiny doses, and very bizarre the body relies on a system subject to so much breakage and a very historically common nutrient deficiency as its master regulator of movement levels (i.e. most able to produce rapid movement or total lack of it similar to the animal parasitism examples).

      It would be very difficult to get enough iodine to over-rev the system from food sources, and any high consumption of it would naturally be very rough on the parasite.

      As to dopamine: Scott fails to mention it’s pretty tough to get access to the brain, or at least takes some time, and the sophisticated viruses and bacteria that do get in may by that point have a much more complex strategy than turn the person crazy. There’s a reason people have a real instinctual fear of psychotic or unstable people–evolutionarily speaking this is a defense against brain parasites affecting these systems.

  35. eyeballfrog says:

    This is the kind of murderous-yet-clever solution I expect of evolution

    What are some other good examples of this? I do love seeing the crazy, murderous arms race that is evolution.

  36. Jan_Rzymkowski says:

    Actually, microbes can efficently coordinate their actions — see quorum sensing.

  37. Jan_Rzymkowski says:

    Also, did anyone look into the possibility of drug tolerance being mediated by immune system? It’s not far fetched to think that the antibodies might recognize drug molecules and induce a reaction leading to formation more antibodies that would bind the drug molecules and effectively decrease the drug plasma levels or that would bind to organism receptors nullifying the effect of the drug.

    While it seems very hypothetical, it would be able to explain how people can have tolerances over periods of years while not experiencing any abnormal homeostasis.

    • Garrett says:

      Antibodies can be detected in the blood. And though not routinely checked for, I would imagine that enough people with reason to be screened in detail (eg. infected with mysterious virus) would have this kind of antibody or immunoglobulin present to have been detected accidentally by now.

  38. Eponymous says:

    Meh. I don’t think we particularly *need* an explanation for why human signalling pathways are complex. We know roughly how evolution works, and it stands to reason that this massive ball of spaghetti code formed by random errors is going to produce some pretty complicated and strange ways of implementing things. We’re highly complex organisms, so DNA has to code for machinery that implements complex conditional behaviors without the benefit of structure that isolates the effects of one part of the code from the rest of the program’s state. In programming terms, most evolved processes are going to involve manipulation of the same set of global variables, which is going to screw up every other process. If you change one thing, you’ll affect all kinds of other things.

    Plus this has to work (well enough) even with a base level of mutational load, and in a variety of environments. So you need plenty of redundancy. See genetic canalization.

    So I can believe that parasitic manipulation is one small piece of what makes the problem of writing our instruction set so difficult. But only one small piece. I think it’s much more likely that the causality runs from our complexity to our (relative) lack of parasitic manipulation than the reverse.

    A more succinct version of the argument: it seems that behavioral complexity requires signalling complexity; and if the evolutionary pressure was on signalling complexity specifically, this wouldn’t automatically create behavioral complexity (which is an added constraint); so the existence of signalling complexity and behavioral complexity together suggests the evolutionary pressure was on behavioral complexity, not signalling complexity; this would have produced the observed data (parasitic manipulation of complex animals difficult and thus rare) without needing to posit an additional process; thus it’s the simpler explanation.

  39. User_Riottt says:

    IIRC one of the down sides of MDMA was that the serotonin overload knocked out some of the serotonin neurotransmitters (especially when dehydrated and overheated). Which eventually lead to “loss of joy” for hardcore ravers where MDMA wouldn’t work anymore and they were rather numb.

  40. Walliserops says:

    I think vertebrate parasites are doing a good enough job of manipulating us, but maybe they’re not using direct ways because we evolved to make these ways harder. Ribeiroia can’t brainwash frogs to make them seek out birds, but it deforms their limbs and makes them less likely to escape. Guinea worms can’t make people drown themselves, but they can produce a burning pain so that their host tries to soothe the wound in water. Pinworms can’t make you force others to swallow them, but they can leave the intestine and crawl around to create an itch, and you spread eggs through your fingers after you scratch. Whooping cough has an even better handle on the “make the host do the thing that spreads more of you” business.

    Also, drastic behavior changes are mostly in the “need to get host eaten” category. It’s possible that vertebrate parasites just don’t have that need, because they’re already likely to get eaten by something relevant. If a thornhead worm wants you to be eaten by a duck, it better take extra measures because when you’re an amphipod there are lots of other things to be eaten by. If you die before doing that, there’s also a big chance you’ll just decay before anything large and final-host-like scavenges you. But if Echinococcus wants you to be eaten by wolves, it doesn’t need to make you seek some, it can just rupture its cyst and count on the wolves to come. And if it really wants extra insurance, it’s probably easier to crack the “can function in other big scavenging carnivores” code than the “can mess with brain machinery” code.

    I’d be interested in knowing the situation in plants. They also have lots of manipulators that work at the molecular scale (gallmakers, root-knot nematodes, endophytic fungi, parasites that interface with mycorrhizal communities, herbivores that inject hormones during feeding to avoid plant defenses, there’s even a wasp that makes its gall secrete sugar to attract guardian ants), and they have rough analogues to insects and higher animals in herbaceous and tree forms (I’m assuming, likely wrongly, that the main distinction is biomass/K and r tactics). Did trees develop the “gain lots of complex signals to deter parasites” strategy while herbs took the “maintain fewer defenses but reproduce faster to compensate” route? Trees and herbs can evolve into one another, so what happens to say, secondarily woody trees?

  41. Dinnty says:

    I find this explanation for Biological complexity delicious. Explains my failure to predict the results of my carefully designed experiments. That parasites that attempt to manipulate these systems exist is factual, so the selection pressure is there. The clearest case can be made for the immune system and infection.
    But that’s not all! This is a general argument for complex systems defeating manipulation by increasing complexity. So, Human Psychology, Economic systems, certainly the stock market adapts in this way. Certainly the unpredictability is what we do see. A system which is too easy to manipulate is doomed by predators and parasites.
    The increasing complexity of Computer operating systems and the internet are driven by this need for resistance. But they are limited because they must remain predictable.

  42. The idea that our gut bacteria are trying to manipulate how we eat and we might be evolved to expect that makes their relationship to mental health more interesting. If we’re going to engage in heavy genetic engineering in the future it might make sense to jus tnot have gut bacteria to simplify things. That’ll come with it’s own problems but might be worth it. Come to think of it that sounds just like the Shapers in Bruce Sterling’s Schizmatrix books.

    • stopandgo says:

      Good luck with that. (The immense complexity of the genomes and actions of millions of co-evolved bacteria are not easy to reproduce.)

  43. dlr says:

    Tuberculosis famously increases peoples sex drive. Which is no doubt a good mechanism to increase it’s transmission rate.

  44. jackjohnson says:

    I’m coming late to this thread, and I’m surprised that AFICT no one has yet mentioned
    Shane Carruth’s film

    https://en.wikipedia.org/wiki/Upstream_Color .

  45. Charlie Lima says:

    The strongest evidence against is the dog that didn’t bark. Some systems look heavily defended against parasite manipulation, but others don’t. Amphetamines raise dopamine effectively and without significant tolerance buildup (see part IV here for a defense of this claim); antipsychotics lower dopamine equally effectively and consistently. Since dopamine is one of the most lucrative systems for parasites to hijack, it’s surprising to find it so easy to affect. And what about immune function? Externally administered corticosteroids decrease immune activity and make the body more vulnerable to infection; why don’t parasites secrete them? Why don’t we have some counter against them? These systems look consistent with an evolutionary history in which we don’t expect any threat from parasite manipulation and don’t need to defend ourselves very hard.

    This logic seems suspect. We, after all, have a strong case example of a parasite that generally weakens the immune response to enable its own replication – HIV. HIV directly hits helper Ts and can basically replicate unchecked in the late stages.

    Yet HIV-like illnesses are pretty lousy at spreading and replicating. SIV has been around for something like 32,000 years according to molecular clock analysis. It most likely spread to humans multiple times during this period, but quite likely it proved too lethal to spark a sustained pandemic.

    I would suggest that when it comes to immune function, giving your host SCID or anything too general is a bad equilibrium. After all, HIV is almost 4 times more infectious during the early acute infection stages (with strong immune responses present) than during the late stage (with terrible immune responses). Further given the length of latent stage, these are likely responsible for even more transmission. Once you tank the host’s general immune infection, any parasite needs to have an exit strategy.

    I suspect this is because while it might help one species to face less immune reaction, it will help every other pathogen too. Cooperation within a species is hard, cooperation between species that have not been in a stable ecology for many generations is likely orders of magnitude harder. Even if this sort of cooperation, where all the pathogens hack down the immune response just enough … there would be no way to prevent freeloading.

    After all, if you look at the classic virulence factors, pretty much all of them only benefit a subset of pathogens, and preferably only the species itself. For instance, Staph has Protein A. It basically knocks out the immunoglobin based branches of the immune system – but only for bacteria that produce Protein A. Strep has C5a peptidase which leaves a lot of bugs high and dry that are more susceptible to C3b mediated opsonization. I would suggest that immune modulation is common, but general immune modulation is less so because of competition with other parasites. After all, there are an awful lot of virulence factors which activate the immune system which might suggest that immune modulatory hacking has a sweet spot for pathogens into which ecology has settled.

    Similarly, going after dopamine might be strong, but would the parasite that hacks it actually get that much reproductive benefit? Imagine if a bug induced much higher libido through some sort of dopamine/mania hack, great it spreads a bit more. On the flip side, every other STI would also spread more. There is limited real estate for pathogens, so would trich be all that much ahead if the sexual pairing matrix got dense enough that everyone had multiple strains of gonorrhea, chlamydia, Hep B, and everything else? I mean, maybe all the STI pathogens would benefit from this sort of thing as a collective … but then again as you stack up pathogens there tend to be even more infections in play.

    I suspect that a lot of these hacks need to be specific as well as effective. Getting eaten by cats is likely only beneficial for a very small number of pathogens. Similarly going rabid likely benefits very few other pathogens; it actively harms STI transmissions, kills off anything that is a long term resident, and even kills some bacteria that do not do so well at neutral pH.

    Ultimately, I suspect the more relevant competition is not between host and parasites (broadly defined), but between parasites. Complex behavior hacks to benefit just one or a few parasites are unlikely to be all that general.

  46. laretluval says:

    Humans are behaving very differently these days from how we did in our “natural” environment…

    What if civilization is the result of a fungal parasite?

  47. Spiritkas says:

    Coming from a biology background, my thought was to extend this logic further into evolutionary history. What if organisms evolved into larger sizes due to this parasitic pressure. It is easy to forget, but more than half of all species on the planet are parasites. Every productive autotroph and heterotroph has multiple specific and non specific parasites. So there is a strong reason to believe that parasitism is a major driving factor in evolution.

    So overall the reason why we have any larger animals at all could be /partially be to support anti parasitic complex behavioural and immunogical pathways. It is expensive, slow reproducing, large animals which are more fragile vs smaller organisms. But there is an upward pressure on size due to the big advantages that come with complex pathways such that smaller organisms such as insects developed very different survival by massive numbers in each reproductive round type solutions.

    Not to be silly and say evolution planned out complexity, more a survivorship of larger organisms when they became complex.

    As for Scott’s objection about, ‘why not this easily manipulated pathway’… There are limits to biology and in a speculative hand waving non specific way I can at least argue that in evolution you can’t get to some end points from some starting points. Seritonin is very old and bacteria have it for other purposes and can hack it to be useful. Perhaps some of our big organism signaling is newer or super toxic to small parasites, or some other resason we picked a few simpler pathways due to an incompatibility with parasites. Perhaps the other counter is that we have some simple pathways because we lacked a route to make them both complex and functional.

    • stopandgo says:

      No coincidence that elephants also bury their dead and have “funerals” to mourn them (or contemplate the cause of death to better avoid it). And animals with greater brain size and social cognition are are much more able to detect potentially ill members of their group and avoid them (or nurture them, if genetically related).

  48. ll11 says:

    This is a really cool hypothesis to think about. Thanks for the review.

  49. jhertzlinger says:

    It looks like parasite strategies are another example of an anti-inductive phenomenon.