Efference copies everywhere


I’m writing a paper right now about efference copies in nematode worms and what they can tell us about perception. The more I learn, the more impressed I am with the little guys, and all of the strange minor procedures organisms undergo so they can function.

I’ll start with the issue that faces the worms. First, they are super simple creatures. They only have 302 neurons (compare to 250,000 in a fruit fly) and thus have a limited repertoire of behavior. For example, if tactile sensors in the head are stimulated, the worm moves backwards. If sensors in the tail are stimulated, it moves forwards. These reflexes are useful for obvious reasons. The worms need a way to avoid predators and obstacles, and this simple behavioral schema seems to do the trick.

But, a dilemma lurks below the surface. Suppose a worm’s head sensors are activated and it starts moving backwards, only to have its tail sensors activate by virtue of moving through the soil. Now, it reverses direction and moves forwards, to have its head sensors activate and trigger the move backward reflex. In moving backwards… its tail sensors activate… and now it tries to move forwards…

Like someone trying to squeeze into a parking space, the worm would stop and start, making little movements forward and backwards in vain as it struggles in a vicious cycle. These seemingly reasonable stimuli responses would render the worm static for eternity, unable to feed or find a mate.

Yet, the nematodes have not gone extinct. They continue to thrive in rotting fruit and be bred for all sorts of scientific experiments. What, then, keeps them mobile?

The answer lies in efference copies. When a worm’s tail sensors are activated, for example, it triggers the move forward reflex, and signals the nervous system to inhibit the move backward response. This signal is an efference copy. (Note: from my understanding, the worm’s head sensors still register stimulus, but it is only the corresponding behavior that is blocked by the copy.)

We can crudely think of efference copies as the neurological equivalent of the nervous system CCing the rest of the body so everybody is on the same page.

More interesting applications of efference copies are present in complex organisms. For example, crickets make noise by rubbing their wings together in a process called stridulation (which is a cool word). The ruckus they create is loud for us, but much more so for them. To ensure they don’t lose their sense of hearing, the signal to make a song that is sent to the cricket’s motor neurons is simultaneously routed to the auditory system. The auditory system then prevents signals from the tympanum (eardrum) from being processed, effectively cutting off hearing. This example differs from the case of the worm, as the cricket only inhibits the processing of stimulus. The worm processes stimulus, but prevents corresponding reflex-based actions.

Efference copies are present in humans, too. They play a role in vision and movement, but my favorite example is tickling.

Feeling tickled is the result of tactile stimulation on sensitive parts of your body. It’s apparently a simple process. If your feet/stomach/armpits, receive the correct stimulus, you feel giggly. Why, then, can you not tickle yourself? You can provide the same type of stimulus as anyone else, so we can’t we bring ourselves to fits of laughter?

Efference copies. Whenever we act, our sensory systems create a “prediction” of how that action will create additional sensory input. We’re just like the worm in this regard. We need a way of distinguishing stimulus created by external objects, versus us. If you run a little brush across your own palm, it’s not very tickly. Your sensory system has already predicted the stimulus associated with the action and is primed to ignore it.

It is possible to tickle yourself, though. You just need to be indirect about it.

Scientists studying the phenomena created a tickling robot. In the robot’s arm is a little brush. Underneath its arm is your open, right palm. In your left hand, you have a small stylus you can use to draw a pattern. The robot will then take the same pattern drawn with your left hand and trace it on your right palm with the brush, hopefully tickling you.

What the researchers found was that if the robot traces the pattern as you’re drawing it, it’s not very tickly. However,as the delay between you drawing the pattern and the robot tracing it increased, the more tickly the result was. A delay of 300ms between you telling the robot how to tickle and it tickling increased subjective feelings of ticklishness by ~50% (full disclosure: I’m eyeballing the graph from the study for this number. The authors don’t provide it. From what I can see, it jumps from 2.1ish to around 3.4 on the tickle rating rank).

I hope you can see why I find efference copies so interesting now. Beyond fulfilling functions just described, they also play a big role in generally distinguishing self from the environment. There is a study describing efference copies’ role in internal speech, as well as a section in the tickling-robot section detailing how auditory illusions with schizophrenia can be attributed to issues with efference copies.

If you want a survey of efference copies across the animal kingdom, you can check out this article.