A Blog by Carl Zimmer

A Wasp Finds the Seat of the Cockroach Soul

Ampulex%20emerging.jpgIf blogs could have mascots, the Loom’s would be the Emerald Cockroach Wasp (Ampulex compressa). Back in 2006, I first wrote about the grisly sophistication of this insect, which turns cockroaches into zombie hosts to be devoured by their offspring. Since then I’ve blogged from time to time about new research on this parasite’s parasite. Last year I sang the praises of the Emerald Cockroach Wasp on the NPR show Radiolab, and, to my surprise, brought some peace of mind to a very scared kid.

Scientists still don’t understand the wasp very well, though, and so I decided last night to see if anyone had discovered something new about it recently. It turns out Ram Gal and Frederic Libersat, two scientists at Ben Gurion University in Israel, just published a paper in which they reveal one of the secrets to zombification. In effect, they identified the seat of the cockroach soul.

Before I describe the new results, let me just refresh your memory about what the Emerald Cockroach Wasp actually does.

Like many parasites, the Emerald Cockroach Wasp manipulates its host’s behavior for its own benefit. As I explain in Parasite Rex, parasites make their hosts do lots of different things (get them into the body of their next host, act as a bodyguard, or build them a shelter to name a few examples). The Emerald Cockroach Wasp needs a live, tame cockroach to feed its babies.

When the female wasp is ready to lay her eggs, she seeks out a cockroach. Landing on the prospective host, she delivers two precise stings.


The first she delivers to the roach’s mid-section, causing its front legs buckle. The brief paralysis caused by the first sting gives the wasp the luxury of time to deliver a more precise sting to the head. The wasp slips her stinger through the roach’s exoskeleton and directly into its brain.

She injects another venom that robs the cockroach of the ability to start walking on its own. The wasp takes hold of one of the roach’s antennae and leads it, like a dog on a leash, to its doom: the wasp’s burrow. The roach creeps obediently inside and sits there quietly as the wasp lays her egg on its underside. The wasp leaves the burrow, sealing the opening behind her.

The egg hatches, and the larva chews a hole in the side of the roach. In it goes. The larva grows inside the roach, devouring the organs of its host, for about eight days. It is then ready to form a pupa inside the roach. After four more weeks, the wasp grows to an adult. It breaks out of its pupa, and out of the roach as well. Only then does the zombie cockroach die.

The zombifying sting has long fascinated scientists. It does not paralyze the roach. It does not put it to sleep. If the zombie roach is frightened, it jumps in the air like a normal roach, but then it fails to run away. What does the wasp understand about the nervous system that we do not?

cockroach head220The cockroach brain is not a brain like ours–a single solid lump of neurons like the one in our head. It’s actually a group of linked clusters of neurons, called the cerebral ganglia. Some studies have suggested that one of those clusters, called the sub-esophageal ganglia (SEG), boosts the signals required for an insect to start walking. So Gal and Libersat decided to see how wasps would behave if the roaches they stung were missing the SEG.

Normally, the wasps only spends about 15 seconds inserting its stinger into the cockroach’s head. But when Gal and Liberset destroyed the SEG in roaches, the wasps were flummoxed. They spent over three minutes poking and prodding inside the cockroach’s head. By contrast, when the scientists cut the nerve running from the SEG to the roach’s body (marked here as NC), the wasps didn’t spend any extra time delivering the sting. It thus appears that the wasps zero in on the SEG to zombify their host.

Gal and Libersat took a closer look at the SEG of zombified cockroaches. Using an electrode, they measured the activity of the SEG. They discovered that the neurons in the SEG became quiet. They spontaneoulsy fired half as often as the neurons in the SEG of normal cockroaches. And a puff of air on the antennae of the zombified roaches–which usually triggers a roar of activity in the SEG so that the insect can escape–produced only half the normal activity in the neurons.

To cap off the experiment, Gal and Libersat then pretended to be wasps. They injected wasp venom directly into the SEG of healthy cockroaches. The injection zombified the roaches. The insects barely moved on their own, and they hardly budged in response to a terrifying puff of air. Gal and Libersat could only zombify the cockroaches with a shot to the SEG, however. If they injected the venom into a neighboring cluster of neurons (marked here as SupEG), the roaches flitted about as if nothing had happened.

While this study does a good job of pinpointing the place where the wasps perform their neurosurgery, it does not close the book. The SEG is actually a complicated maze of neurons. Gal and Libersat are now investigated exactly where in the SEG the wasp sends its stinger, and what precisely its venom does to those particular neurons. The Emerald Cockroach Wasp will no doubt make yet another visit to the Loom in years to come, because it has more to teach us.

0 thoughts on “A Wasp Finds the Seat of the Cockroach Soul

  1. “The cockroach brain is not really really a brain(.)”

    Pretty much every invertebrate neurobiologist I know (including me) refers to them as brains.

    The cerebral ganglia make up a smaller proportion of an insect’s total nervous system than in vertebrates, but they’re still the largest neural centers. They’re located in the head of the animal, and they integrate multiple sensory inputs, and coordinate behaviour. That’s pretty much everything that vertebrate brains do.

    I haven’t yet read a convincing argument to call the big cluster of neurons in the head of an invertebrate anything besides a brain. Arguments to say invertebrates just have cerebral ganglia” always seem to reek of vertebrate chauvinism.

  2. Just how does the wasp know that this section is missing? What kind of senses are in use to guide the sting?

    [CZ: Thanny, Gal Haspel (who left a comment below) and his colleagues have, I believe, found evidence of sensors on the stinger that they can use to sense their way through the brain–or the ganglia–or whatever you want to call it…Think of putting your arm down a storm drain to find a nickel.]

  3. To clarify, I agree that it would be wrong to treat the combination of the supra-esophageal ganglion (SupEG) and sub-esophageal ganglia (SEG) as the cockroach’s brain. The SupEG is brain, and the SEG isn’t.

  4. It is a very interesting result. Thanks for posting, Carl!

    Even more interesting because we already know that the wasp does inject venom into both head ganglia (more about head ganglia vs brain below). And it seems to inject more venom into the SuperHG and it is further away from the injection site (Haspel et al 2003 J Neurobiol). What does the venom do in the Super EG? Maybe something we have yet to disribe? Maybe the metabolic effects described in an earlier paper (Haspel et al 2005 J Comp Physiol).

    Ah. yes. Insect brain. It’s a matter of who is talking to whom. I am happy to call the nerve-ring of the nematode I am studying “brain” when it serves to clarify. We actually called the SuperEG “brain” in one of the earlier papers. However, when most people read “brain” they have an image in mind which will confuse them when reading this story (wait! The cockroach has two brains? is this like a right and left hemispheres?….). I think that Carl made a clarifying distinction that will help most readers.

  5. @Zen: Why is SuperEG = Brain, SEG+SuperEG~=Brain?
    just because there’s an esophagus running through it?

    I think that for both popular and scientific writing it’s a matter of clearly defining what you are calling “brain” (how about the mammalian brainstem? Is it “brain”?)

  6. Thanks for the input, Gal. While I appreciate the point that people might be confused by invertebrate neuroanatomy, I think it can be explained so that the animals don’t sound “brainless.”

  7. Re: #5. brainstem is definitely brain. The vertebrate CNS is spinal cord + brain.

    The term “brainstem” is used differently by different people, but typically means brain – (cerebral hemisphere + cerebellum). Since it includes the thalamus, it looks sort of like a fist+ top forearm. I’ve also seen usage where “brainstem” does not include diencephalon (thalamus + hypothalamus). Can’t give you citations.

    Where is consciousness?
    In a sense, vertebrates have a few brains, arranged in a kind of hierarchy. Enteric nervous system, a network running thru the gut, appears to be the lowest. The autonomic nervous system (sympathetic and parasympathetic) seems one level up. And the CNS is the highest.

    Even within the CNS there is a kind of hierarchy. Spinal cord can coordinate functions, like the basis of locomotion and withdrawal reflexes. Supervising it is brainstem, which supports vegetative functions, but also provides components of balance and postural reflexes to the spinal cord. Finally, the cerebral hemispheres supply things like goal-orientation to behavioral integration.

    The point is, to some degree, integrated function can occur at several levels. My suspicion is that the layering of enteric/autonomic/somatic nervous systems represents an evolutionary sequence. Within the CNS, accidents that cause spinal damage can separate parts of the hierarchy.

    An interesting question is whether we have multiple souls and multiple consciousnesses.

  8. @Thanny: as Carl wrote, we found some sensory-organ-looking bumps and hairs on the stinger. We used scanning electron microscopy (SEM) so we could only see the shapes. I also found that many neurons lead from the stinger to the abdominal ganglion of the wasp. Presumably some are sensory neurons and deliver information from the presumed sensory organs. This is still unpublished because it never made it into a full story.
    Now, what kind of sensory organs is the question that might make this into a story. I could think of a few possibilities: chemosensation (maybe of neurotransmitters), mechanosensation (the nerve tissue might be softer or under different osmotic pressure), and my favorite: electrosensation (stinger as an extracellular electrode that feels action potential. Why not?). Any other ideas?

    @Zen: I’m with you. Invertebrates are as brainless as vertebrates :-)

    @Johnk: my point exactly about brainstem (I was trying to show that the term “brain” is not carved in stone).

  9. Thanny said: “Just how does the wasp know that this section is missing? What kind of senses are in use to guide the sting?”

    maybe there are receptors for specific protein markers on the wasp’s sting?

  10. I’d be interested in how this evolved. This is very specific behaviour! Are we talking about 2 species that have co-evolved over very long periods of time?

  11. I’m curious, given the above description of the venom’s effect, why does the roach follow the wasp? It’s SEG is damped down, so it doesn’t run. But why does it begin walking? How does the wasp steer it?

    What usually happens when a roach antenna is firmly grasped? Does the roach try to break free or does it attempt to free it’s antenna? Or something else? And what prohibits the roach from fighting?

  12. I seem to recall seeing a video of the wasp “leading” the cockeroach along but can’t seem to remember where or find them.
    Carl? did you post them or point to them?


  13. It’s definitely a very intriguing behaviour. Not as much in the realm of neurology as every article about this particular sort of parasitism seems to be focused on, but rather on the side of evolutionism.
    A simple answer to the riddle of how such a precise interspecies predation came to occur completely eludes me.
    The possibility of it being a learned skill is negligible, since a parental nurturing similar to mammalean required for this is a rarity among insects and specifically in this species, but also because to imagine that chemical language and mimesis alone is capable of conveying such sophistication is simply fantastic.
    How can natural selection then induce such a trait to develop? Not only does everyone of those wasps follows that precise modus operandi, it actually knows instinctively what exact spot to target and what sensation to expect. Moreover to simply sting the cockroach is not enough, since it would just leave it at the mercy of other predators and doom the offspring. The wasp must then ‘ride’ the zombie cockroach by using another set of skills.
    This step-by-step process seems too extremely accurate to be explained by mere accumulation of accidents. Not to mention for it to even begin developing, you have to pit a single wasp against an animal three times its size, wich in itself sounds like going against inate self-preservation.
    I’m not trying to debunk evolutionism, or advogate primate level intelligence in insects. Hive-mind societies have proven extraordinarily resourceful, but this is completely off the chart!
    I’d really love to hear someone offer their insights on this.

Leave a Reply

Your email address will not be published. Required fields are marked *