Self-tickling
A Socratic walk-through of self-tickling — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why can nobody tickle themselves, however hard they try?
Run a finger along your own ribs. The touch is perfectly real — you feel the pressure, the path, the texture of the shirt. What is missing is the squirm. Now let someone else do exactly the same thing and the sensation is unbearable. The stimulus is identical; only its author changed. So whatever fails when you tickle yourself, it cannot be a failure of the skin.
Which raises an odd possibility: that your brain is treating one of these touches as less important than the other, and doing so before you have had a chance to notice.
Reasoning it through
REASONING #Consider the problem your nervous system faces every waking second. You are drenched in sensation you produced yourself — clothes shifting, tongue against teeth, footfalls, the movement of your own eyes. Almost none of it is news. What matters is the touch you did not cause, because that is the only kind that tells you something about the world. So a sensible system would find a way to mark self-produced sensation as expected and turn its volume down.
How could it do that? It would need to know what was coming. And it does: every command the motor system issues to a muscle can also be copied and sent elsewhere — a duplicate of the order, not a report from the limb. That copy is old enough in neuroscience to have a name, the efference copy. Given it, a region can compute what the movement is about to feel like, and have that prediction ready before the skin has been touched at all.
Then the last step is simple arithmetic. Compare the prediction with what actually arrives. If they match, the sensation was self-caused, and it can be attenuated. If they do not, something in the world did it, and it deserves full weight. The cerebellum is the structure most implicated in generating the prediction, and imaging work finds less response in somatosensory cortex to a self-produced touch than to an identical touch delivered by someone else.
If that is right, it makes a testable and slightly absurd prediction: break the match, and you should be able to tickle yourself. Blakemore, Wolpert and Frith built exactly that. Participants moved a lever with one hand, and the lever drove a robot arm that stroked the palm of the other. With no delay, it felt like self-touch — barely ticklish. Then they inserted a delay between the movement and the touch, and stepped it up. Ticklishness rose with the delay, and rose again when the robot rotated the direction of the stroke away from the direction of the hand movement. Nothing about the touch itself changed. Only its agreement with the prediction did.
Notice what that buys us beyond a party trick. The same comparison — did the world do what my command said it would? — is one of the leading accounts of how you know an action was yours. Attenuation and the sense of agency come out of the same computation. And the tickle work connects to that directly: the same group reported that people with schizophrenia who experience passivity symptoms, the sense that their movements are being made by someone else, show less of this attenuation and rate self-produced touch as more ticklish than controls do.
The analogy
THE ANALOGY #Think of noise-cancelling headphones. They do not muffle sound; they generate its opposite and add it in, so that what reaches you is the difference between what was expected and what arrived. Silence is not absence — it is successful subtraction. Your tickle response is the residue left over when that subtraction fails.
The headphones cancel a sound they have already measured arriving, whereas your brain cancels a touch it predicted from its own outgoing command — which is precisely why a two-hundred-millisecond delay defeats it, while the headphones would be untroubled.
Clarifying the model
THE MODEL #Two refinements are worth holding onto. First, the prediction is not a guess about the world in general; it is derived from the motor command, which is why only your own movement gets cancelled. Watching your hand approach, or knowing with total certainty that a friend is about to poke you, does not attenuate anything — foreknowledge is not the same as an efference copy. This is why bracing yourself never works.
Second, and more honestly: attenuation of self-touch is well established, but whether it is the whole story of ticklishness is not settled. Tickling is also a social behaviour — it depends on who is doing it and on the relationship, it is bound up with play and laughter, and children tickle each other in ways no account of sensory cancellation explains. The forward model tells you convincingly why self-touch is damped. It tells you rather less about why being touched by someone else is funny. Those are two questions, and the experiment answers the first.
A picture of it
THE PICTURE #How to readFollow the numbered messages downward. Step 2 is the whole trick — the command is copied to the cerebellum at the same moment it is sent to the hand, so a prediction can arrive at sensory cortex before the touch does. The first note is the ordinary case, where prediction and sensation cancel. The dashed message is the experiment: the identical touch, delayed or turned off-path, no longer matches, and what survives the subtraction is the tickle.
What became clearer
WHAT CLEARED #You cannot tickle yourself because you already know what it will feel like — not in the sense of expecting it, but in the strict sense that a copy of your own motor command was used to build a prediction, and that prediction was subtracted from what your skin reported. Tickle is what is left over when the subtraction fails. And since the same comparison plausibly underwrites your sense that an action was yours at all, the failure to tickle yourself is a small daily readout of the machinery that keeps self and world apart.
Where to go next
ONWARD #- Why your visual world stays still during eye movements, which uses the same predictive trick.
- Whether force escalation in a mutual shoving match — each side genuinely feeling the other pushed harder — is the same attenuation seen in a different modality.
Key terms
TERMS #| Term | What it means |
|---|---|
| Efference copy | a duplicate of an outgoing motor command, routed to sensory areas rather than to muscle. |
| Forward model | an internal predictor that turns a motor command into an estimate of the sensory consequences it will produce. |
| Sensory attenuation | the reduced neural and perceived response to a sensation you produced yourself. |
| Sense of agency | the experience that an action was performed by you, commonly modelled as a match between predicted and actual sensory outcome. |
Every term the collection defines is gathered in the glossary.