Why you can't tickle yourself (and how a robot can fix that)

A joke you already know the punchline to

Run your fingers lightly along the sole of your own foot. It feels like touch, maybe a little odd, but it doesn't make you squirm. Get someone else to do exactly the same thing and it can be unbearable. Scientists noticed this long ago. In 1971 Lawrence Weiskrantz and colleagues at Oxford published a short paper in Nature showing that people rate a touch they deliver themselves as less tickly than the same touch delivered by someone else.

The obvious explanation is surprise: you know where your own fingers are going. That turns out to be right, but "knowing" here isn't a conscious thought. It is a specific, measurable calculation that your brain runs every time you move, and it happens fast enough to change what you feel.

A copy of every command

When your brain sends a command to your muscles, it also keeps a copy. Neuroscientists call this the efference copy. A system that works like a simulator, a forward model, uses that copy to predict what the movement will feel like: where the touch will land, when, and how hard.

Then the brain compares the prediction with what the skin actually reports. Whatever matches the prediction is turned down. Whatever doesn't match gets through at full strength. A tickle from someone else is all mismatch, because there was no command to predict it from. Your own tickle is almost all match, so there's little left to feel.

In 1998 Sarah-Jayne Blakemore, Daniel Wolpert and Chris Frith at University College London looked at this in a brain scanner. When people stroked their own palm, the somatosensory cortex, which handles touch, responded less than when the experimenter did the stroking. The cerebellum behaved differently for a movement that produced a touch and for the same movement that didn't. The team concluded that the cerebellum predicts the specific sensory result of a movement and supplies the signal that cancels it.

Fooling the prediction with a robot

If the prediction is what dulls the tickle, then a prediction that is slightly wrong should let some tickle back in. In a 1999 study, the same team gave 16 volunteers a small robot arm. Each person moved a handle with their left hand, and the robot copied that movement, brushing a piece of soft foam back and forth across their right palm. The volunteers were tickling themselves, but through a machine the scientists could tamper with.

They tampered in two ways. They delayed the robot by 100, 200 or 300 milliseconds, so the foam moved a moment after the hand. Or they rotated its direction by 30, 60 or 90 degrees, so a left-right hand movement produced a stroke at an angle. The movement was the same. Only the match between command and touch changed.

Both changes made the touch more ticklish, and the bigger the change, the stronger the effect. The rise with delay was steep up to about 200 ms and then levelled off. At the largest delays and angles, people rated the touch roughly as ticklish as when the robot moved on its own. A fifth of a second is enough to make your own hand feel like a stranger's.

Things to try:

• Start at zero. The two dots move as one, and the tickle is cancelled.
• Slide the delay to 100 ms, then 200 ms. The foam falls behind your hand and the gap between the curves opens up.
• Set the delay back to zero and rotate instead. Timing is perfect, but the touch lands in the wrong place.
• Press Someone else tickles you. There is no command, so there is no prediction at all.

Why playground fights escalate

The same dimming applies to force, not just tickle, and it has an awkward side effect. In a 2003 study in Science, Sukhwinder Shergill, Paul Bays, Chris Frith and Daniel Wolpert paired people up and had them take turns pressing on each other's finger. Each person was told to press back with exactly the force they had just felt.

Nobody managed it. Across six pairs and eight turns, the force grew by an average of 38% on each turn. Each player felt their own press as weaker than it was, so they pushed harder to "match", and the other player did the same. When people reproduced a force with a joystick instead of their own finger, their brain had no direct touch to predict, and they matched it much more accurately. The researchers suggested this is one reason tit-for-tat shoves tend to get worse: each side honestly believes it only hit back as hard as it was hit.

When the prediction breaks

If this system labels sensations as "mine" or "not mine", what happens when it fails? In a 2000 study, Blakemore and colleagues tested people with schizophrenia and other psychiatric conditions using the same kind of self-tickle comparison. Most people, patients included, rated a self-made touch as less tickly and less intense than one made by the experimenter. But patients with auditory hallucinations or passivity experiences, the feeling that their actions or thoughts are controlled by an outside force, rated the two about the same. Their self-touch was not dimmed.

That fits a long-standing idea about these symptoms. If your brain doesn't mark your own actions, or your own inner speech, as self-generated, they can feel as if they come from somewhere else. For this group, tickling yourself works about as well as being tickled. It is a small sign of a much bigger problem: the loss of the quiet background signal that says "that was you".

For everyone else, the lesson is how much your perception is prediction. Your brain doesn't just record the world. It runs ahead, guesses what your own body is about to do, and subtracts it, so that what's left over, the unexpected, gets your full attention. A tickle is one of the few times you can feel that subtraction fail.