Recent data on how the “ant zombie” fungus works

When I first read the publicity about this recent (2017) paper in Proc. Nat. Acad. Sci., I thought that the authors had come up with a new solution to how the fungus Ophiocoryceps unilateralis, a parasite on carpenter ants, turns the ant into a zombie, behaving in a way that facilitates the dispersal of the fungus.

People are fascinated with this system because of the “zombie” connection, and because of the fact that a fungus (considered a “lower” species) can control the behavior of the ant. In this this case the fungus somehow causes the infected ant to climb onto vegetation, bite into it hard, and then die with its body extending out, allowing dispersal of the fungal spores into the air from a stalk that grows out of the ant’s corpse. The spores then disperse, and can fall on another ant so that the fungus begins it life cycle anew. Here’s the life cycle from MicrobeWiki:

And a photo of a dead and with the fungus sprouting from its body. Note how the ant has bitten into the vegetation to secure itself: a true death grip!

And here’s a Planet Earth video, with Attenborough narrating, showing the “attack of the killer fungus” Cordyceps (a different species that kills the same way) on bullet ants (a different species). But the principle is the same (though not necessarily the mechanism).


There are other examples of microbes or fungi controlling the behavior of their host in such a way to facilitate their own spread at the expense of the host’s life. A recently famous one involves the discovery that the protozoan parasite Toxoplasma gondii appears to change the behavior of rodents—its intermediate host—making them lose their fear of cats. The rodents are then devoured more easily by cats, who then shed the eggs in their feces, infecting more rodents who come into contact with the feces. Here’s that life cycle (it has several life stages) from Wikipedia:

Parasites manipulating the behavior of their hosts to facilitate the spread—and, of course, the spread of the genes producing that host-altering behavior—is what Richard Dawkins calls an “extended phenotype”, with the new behavior being considered a trait of the parasite that produced it. However the parasites do this, it’s a marvel of evolution—a sophisticated “behavior” of simple organisms that arose via natural selection.

Up till when this paper was published, it’s always been thought that in the case of ants and “zombie-producing” fungi, the fungus invaded the host’s brain, changing its behavior in an adaptive way for the fungus (note Attenborough’s “explanation” in the video). Now, however, the 2017 paper by Maridel Fredericksen et al. shows that the fungus invades not the brain, but muscles in every part of the body (click on screenshot below, the pdf is here, and the reference and link at the bottom). In fact, most of us missed that paper, and continued to say in the last three years that the fungus turns the ant into a zombie by invading the ant’s brain.

That excited me at first, because I thought that the fungus was somehow manipulating the actual movement of the ant’s muscles, including that crucial clamping of the mandibles on the plant. But that doesn’t seem to be the case. The invasion of muscles may provide nutrients for the parasite, but is not yet seen as a way to control the ant’s behavior. What’s new, then, is that while we still don’t know how the “zombie” behavior arises, we know it doesn’t involve fungal invasion of the brain.

The advantage of this system is that the fungus and ant can both be cultured in the lab and so experiments can be done.  The authors did sophisticated experiments involving infecting the carpenter ant Camponotus castaneus with the fungus, and also with a “control” fungus Beauveria bassiana, a general pathogen which doesn’t change ant behavior.

They infected ants with both fungi, and traced where the infection went with a combination of staining, producing serial sections of three parts of the ants’ bodies, staining the sections, training the microscope through “deep learning” AI to distinguish ant tissue from fungal tissue, and painstakingly putting together the sections.

In short, they found fungal tissue throughout the ant’s body, but not in the brain, even though the ants behaved as zombies. This shows that brain infection by fungal hyphae (the branching filaments that the fungus produces) is not responsible for changing the ant’s behavior.

What they did find is that in most ants the different fungal filaments connect up together to form a network, and that that network often encircles the ant’s muscles and sometimes invades them.  Here’s a photo showing the hyphae joining up; the caption is from the paper. “B” shows the cross section of hyphae forming connections with each other:

Fungal interactions observed in O. unilateralis s.l.-infected ant muscles. (A) Serial block-face SEM image showing fungal hyphal bodies (HB) and hyphae (arrowheads) occupying the spaces between ant mandible muscle fibers (M). Outlined boxes are shown larger in B and C. (Scale bar, 50 µm.) (B) Connections between hyphal bodies (arrows). (Scale bar, 10 µm.) (Inset) Close-up of connected hyphal bodies. (Scale bar, 1 µm.) (C) Muscle fiber invasion: hyphae have penetrated the membrane of this muscle fiber and are embedded within the muscle cell (arrows). (Scale bar, 10 µm.)

This is a cross section, but when you put lots of cross-sections together, you can reconstruct the 3-dimensional structure of the fungal network, as in this figure (again, caption from the paper)

Three-dimensional reconstructions of fungal networks surrounding muscle fibers. (A) A single fiber of an ant mandible adductor muscle (red) surrounded by 25 connected hyphal bodies (yellow). Connections between cells are visible as short tubes, and many cells have hyphae growing from their ends. Some of these hyphae have grown along and parallel to the muscle fiber (arrowhead in Inset). This reconstruction was created using Avizo software. See also Movie S1 and interactive 3D pdf (Fig. S3). (B) Two different projections of a 3D reconstruction showing several muscle fibers (blue) and fungal hyphal bodies (red) from the same area as seen in A. This reconstruction was created using a method (developed here) that uses a U-Net deep-learning model.

The fungus not only surrounded the muscles, but in some cases penetrated the muscle fibers themselves, perhaps, the authors posit, to obtain nutrients (the generalist fungus occasionally did this, too). There is no indication that this connection of the fungal hyphae with the muscles affects how the ant behaves.

So what we have is not an answer to the question of how the zombie fungus works its magic, but how it doesn’t—through the brain. But that doesn’t mean that the behavior isn’t mediated through the brain, for the fungus could still secrete some kind of molecule that interacts with the brain and doesn’t require the fungus to enter the brain. There’s still a lot left to learn. But we shouldn’t doubt that the manipulation of ant behavior really is an evolved “extended phenotype” of the fungus.


Fredericksen, M. A., Y. Zhang, M. L. Hazen, R. G. Loreto, C. A. Mangold, D. Z. Chen, and D. P. Hughes. 2017. Three-dimensional visualization and a deep-learning model reveal complex fungal parasite networks in behaviorally manipulated ants. Proceedings of the National Academy of Sciences 114:12590-12595.


  1. ThyroidPlanet
    Posted June 2, 2020 at 9:06 am | Permalink


  2. EdwardM
    Posted June 2, 2020 at 9:27 am | Permalink

    Facinating….and horrifying. Blood curdling, sphincter tightening, nightmare inducing, horrifyingness undreamed of in any Hollywood B movie director’s mind.

    • eric
      Posted June 2, 2020 at 9:31 am | Permalink

      Yeah, first we thought it mind-controlled the ant. That was horrifying. Now we know it doesn’t mind-control the ant (at least, that we can tell…yet…), and that somehow made it even worse.

      Thank goodness I only have this disease that makes me unusually fond of cats…

      • Posted June 2, 2020 at 1:25 pm | Permalink

        It could mind control the ant, and likely does, but not through any obvious way. Could be a chemical control.

    • Saul Sorrell-Till
      Posted June 2, 2020 at 10:02 am | Permalink

      “undreamed of in any Hollywood B movie director’s mind.”

      Funnily enough it’s the basis for a fictional pandemic that sweeps the planet in the game ‘The Last Of Us’. People are infected and start sprouting these fungal structures, which eventually burst in a haze of green spores.

      The long awaited sequel is out in a couple of weeks. Video game horror and sci-fi is where it’s at these days.

  3. eric
    Posted June 2, 2020 at 9:28 am | Permalink

    But that doesn’t mean that the behavior isn’t mediated through the brain, for the fungus could still secrete some kind of molecule that interacts with the brain and doesn’t require the fungus to enter the brain.

    Having suffered through a herniated disk, my initial thought was a simpler mechanism: the extra fungal mass in the body, impinging on muscles etc. might (eventually) cause the ant pain when it moves its muscles. So that (eventually) staying in a single position, clamped down, is the only way it avoids excruciating pain. This does not explain why it moves to specific places to clamp down though, so there’s got to be more to it than just that.

    • EdwardM
      Posted June 2, 2020 at 9:35 am | Permalink

      I was thinking something along these lines. Perhaps the fungus makes it so the ant seeks out the right humidity/temp the fungus needs then when it arrives it stimulates the all the muscles it has formed these networks around to contract. The carpenter ant, like many ants, have really strong jaws and so it too clamps hard fixing it to whatever substrate it’s on.

      But somehow the fungus must override whatever neural signals the ants usually follow – I believe they mostly rely on scent trails. How does the fungus do that?

      Fascinating and skin crawling to boot.

      • Posted June 2, 2020 at 6:54 pm | Permalink

        I’m thinking it almost has to control the ant brain. The ant brain and sense organs probably already have ways to sense humidity and keep track of elevation. The only alternative is for the fungus to grow its own “brain” to sense/calculate these things — which seems ridiculously hard. This assumes that the ant really does go to a certain elevation and humidity level, not just, say, crawl upward for N hours and then freeze up, which happens to coincide with a given elevation.

  4. Ken Kukec
    Posted June 2, 2020 at 9:37 am | Permalink


  5. GBJames
    Posted June 2, 2020 at 9:38 am | Permalink

    Most interesting!

  6. Nicolaas Stempels
    Posted June 2, 2020 at 9:43 am | Permalink

    Very interesting.

    Do upper airway infections makes us sneeze and cough to be spread more easily?

  7. Janet
    Posted June 2, 2020 at 9:43 am | Permalink

    How fascinating the parasite does not invade the brain. I hope they find the molecule/s that ultimately cause this behavior. Talk about lack of free will!

  8. Sue B
    Posted June 2, 2020 at 9:44 am | Permalink


  9. Posted June 2, 2020 at 9:52 am | Permalink

    They … traced where the infection went with a combination of staining, producing serial sections of three parts of the ants’ bodies, staining the sections, training the microscope through “deep learning” AI to distinguish ant tissue from fungal tissue, and painstakingly putting together the sections.

    The implication there is that it would normally be quite hard to distinguish fungal material from ant material. How did the experimenters know that their staining and AI techniques would definitely show any fungal material that was in the brain? What if the brain control part was chemically distinct from the nutrient gathering part?

    • EdwardM
      Posted June 2, 2020 at 10:04 am | Permalink

      They stained the neurons with one stain (an antibody against a protein called SYNORF1) and the fungus with another which stains chitin – a protein found in the fungus but not in the neurons. From the caption to figure Figure 1

      “Host brain is identified by immunofluorescence of neuronal synapses (green, anti-SYNORF1), while fungal cells and host tracheae are stained with calcofluor white (red, antichitin).”

      The cell types are clearly distinguished in the confocal images.

  10. A C Harper
    Posted June 2, 2020 at 9:55 am | Permalink

    Interesting stuff. My leading question is ‘What is it like to be an ant’?

    Presumanbly the ick factor arises from the idea of an individual being taken over – which horrifies us. Yet if we regard the ant as merely a remote automaton (a machine which performs a range of functions according to a predetermined set of coded instructions) for the distributed behaviour of the colony, then the parasite is merely seizing one of the many remote units for its own evolved purposes. A loss perhaps but not icky.

    Rather like if aliens suddenly co-opted the Mars rover Opportunity, or somebody hacking your autodrive capable car while it is parked and driving it away.

    • GBJames
      Posted June 2, 2020 at 9:58 am | Permalink

      I remember reading Dan Dennett saying (about other creatures, not necessarily ants) “It isn’t like anything”.

      In any case… ants are animals with brains. It doesn’t seem unreasonable to suppose that they share certain features with other animals including humans and this is what motivates the “ick” response in us.

      • Posted June 7, 2020 at 9:58 am | Permalink

        Do they represent their current and future state? Then it is likely there is something like it to be an ant.

  11. rickflick
    Posted June 2, 2020 at 10:07 am | Permalink

    The mystery continues. I’m thinking it’s got to be a molecule from the fungus invading the brain. Otherwise how would there be such a large effect on behavior? Twitching the muscles in a coordinated way to induce climbing behavior seems unlikely.

  12. E Siegel
    Posted June 2, 2020 at 10:15 am | Permalink

    The rabies virus with only 5 genes alters the animal’s behavior so that the animal spreads the virus by attacking and biting.

    • EdwardM
      Posted June 2, 2020 at 10:20 am | Permalink

      Some encephalomyeltis viruses cause cows and horses -and occasionally humans- to go “mad”.

    • Mark R.
      Posted June 2, 2020 at 3:31 pm | Permalink

      I wonder if that would also be classified as an extended phenotype.

      • EdwardM
        Posted June 2, 2020 at 4:16 pm | Permalink

        That’s a good question! I think in the case E. Siegel writes it may be an EP (as I understand it) while in the example I gave it is not (the “madness” is a symptom of the infection, not means of it spreading).

        Still, I’m not sure even with the rabies. Lyssaviruses infect many mammals but their primary hosts are bats. They are transmitted mostly through saliva but curiously, it doesn’t turn bats into aggressive bity animals. They already are, amongst themselves; jostling and fighting for position in a roost, they often bite and scratch their neighbors. So they don’t need to be turned into the kind of animal that can transmit the virus. But in some other hosts is can induce an aggressive biting behavior which certainly does spread the virus. This is especially true for raccoons. Lyssavirus (and morbilivirus, which causes distemper) epidemics are often how many raccoon populations are controlled, sadly.

        But really, I haven’t any real idea what’s true.

  13. jezgrove
    Posted June 2, 2020 at 10:18 am | Permalink

    If anyone can create a genetically modified version of the fungus that I can sprinkle on my teenagers and which takes control of their bodies to get them out of bed in the morning, that would be great!

    Joking aside, a fascinating article – albeit with deeply unsettling aspects.

  14. Posted June 2, 2020 at 10:22 am | Permalink

    I noticed that you didn’t mention the possibility of this fungus crossing over to humans. That would make for a science fiction movie with a huge yuck factor. Can you imagine the scene where our heroes come upon a person who has died in their home from this fungus? A haze of spores fills the air in the tiny room …

    • EdwardM
      Posted June 2, 2020 at 10:31 am | Permalink

      Oh great. Thanks a LOT, Paul. A pandemic, economic collapse, cities looted, a moron in the WH and now I have to worry about a fungal stalk growing out of my head while I stagger like a marionette around the countryside?

      I need a duckling post.

    • Posted June 2, 2020 at 10:32 am | Permalink

      The huge stroma erupted from the back of the victim’s neck would rival Alien in yuck factor.

    • jezgrove
      Posted June 2, 2020 at 11:17 am | Permalink

      As Saul pointed out in a reply at #2 above, there is already a video game based on this premise!

    • rickflick
      Posted June 2, 2020 at 1:00 pm | Permalink

      Viruses already control human behavior. They cause us to sneeze and cough, thus, ensuring they are transferred to a new host.

      I wonder if scratching a rash is also a vector caused by a virus. Skin to fingers, fingers to new host.

      • Mark R.
        Posted June 2, 2020 at 3:33 pm | Permalink

        They also cause us to buy exorbitant amounts of t.p. 😉

  15. Posted June 2, 2020 at 10:27 am | Permalink


  16. Posted June 2, 2020 at 11:37 am | Permalink

    How long, I wonder, until we see the stroma sprout from the necks of the zombie Trump supporters? The orange fungus needs fresh blood.

  17. Jeff Dennis Campbell
    Posted June 2, 2020 at 12:31 pm | Permalink

    More evidence that god is great! And by great I mean Great Cthulhu!

    Could the stimulus be that the muscles just move faster when they are warmer? That would tend to make the ant go to the warmest place- which would be up on a sunny day. It seems like that would be an easy thing to test in the lab.

  18. Mark R.
    Posted June 2, 2020 at 3:36 pm | Permalink

    Gruesome, but like a car accident, I couldn’t stop looking (reading). It’s actually as interesting as it is gruesome. I’ve never felt sorry for an ant before.

  19. Paul Mannering
    Posted June 2, 2020 at 9:27 pm | Permalink

    I really enjoyed the Scientific American publication on this kind of parasitic infestation in insects. It was an anthology of papers and articles on not only the “zombie ant fungus” but a dozen other even more gruesome ways insects are used as incubators and reproductive vectors for fungi and even other insects.

  20. Matthew Jenkins
    Posted June 3, 2020 at 12:10 am | Permalink

    In ‘The man who mistook his wife for a hat’ there is an account of an elderly lady who realizes her syphilis, contracted in youth in the brothel from which her husband rescued her, is returning because she ‘feels frisky’. It makes sense that STI’s, in order to spread better, would make their hosts’ behavior change.
    ‘Pox: Genius, Madness and the Mysteries of Syphilis’ notes how many of the nineteenth century’s greatest minds were affected by syphilis; so maybe causing a disinhibited, unbridled nature is part of syphilis’s armory for increasing its hosts’ ability to hook up.

  21. metamorphyous
    Posted June 3, 2020 at 8:18 am | Permalink

    I first heard of this phenomenon in a lecture by Daniell Dennet. I believe he was still under the impression that the fungi attacked the brain. The lecture may have preceded this paper. Fascinating.

    • GBJames
      Posted June 3, 2020 at 9:05 am | Permalink

      It may still attack the brain but do so indirectly through chemical signaling.

Post a Comment

Required fields are marked *

%d bloggers like this: