Drag your own fingertip along the arch of your foot and nothing happens. A friend traces the same line and the leg jerks, the breath catches, the laugh arrives before you decide to laugh. Same skin, same pressure, roughly the same speed — a completely different nervous system on the receiving end of the touch.
For most of the twentieth century this was filed as a curiosity. Then in the late 1990s, Sarah-Jayne Blakemore, Daniel Wolpert and Chris Frith built a pair of robots to take the self out of self-touch — and the answer stopped being mysterious. Why you can’t tickle yourself is a story about a copy of a motor command, a forecast made in the cerebellum, and a volume knob in the somatosensory cortex.
Two Kinds of Tickle, and Only One Needs a Stranger
English uses one word for two sensations. Knismesis is the light, feathery, itch-like tingle — the kind a hair or an insect leg produces, usually without laughter. Gargalesis is the heavier, pressure-based tickle that produces squirming and the helpless laugh, and it is what people mean when they say they are ticklish.
Both sensations show attenuation when delivered to oneself, but the evidence across these categories is not uniform. Controlled psychophysical investigations indicate that while self-administered tactile stimuli substantially attenuate perceived ticklishness, feathery knismesis retains slightly higher residual perceptual sensitivity compared to heavy, laughter-inducing gargalesis. Self-touch substantially attenuates subjective tickle intensity without abolishing underlying tactile awareness. Whatever cancels the sensation cannot be specific to laughter, or to the ribs, or to the feet.
The Copy Your Brain Sends to Its Own Predictor
The difference is not in the fingertip but in a prediction the brain makes before the touch lands. When you decide to move, the motor cortex sends a command down to the muscles. It also sends a duplicate of that command somewhere else — an efference copy, a carbon copy of the instruction that never reaches a muscle but does reach the brain’s prediction machinery.
That machinery is a forward model: an internal simulator that takes the motor command and calculates what the movement will feel like before any sensory signal comes back. Studies of this system point to the cerebellum as a key site for generating those predictions of sensory consequence (Blakemore et al., 1998).
The predicted sensation is compared with the real one. Where the two match, the brain turns the volume down , sensory attenuation, the deliberate dimming of signals it already knew were coming. The touch reaches the somatosensory cortex, the strip that maps pressure and vibration across the body, at reduced strength (Blakemore et al., 2000).
When you move your own hand, cerebellar forward models compute an extraordinarily precise sensorimotor prediction: the exact contact coordinates, timing, trajectory, and force vector are calculated milliseconds before touch occurs. As the incoming tactile signal arrives, this forward prediction is subtracted from actual sensory feedback across the primary somatosensory cortex (S1) and anterior cingulate cortex (ACC), dampening both physical intensity and affective tickle valence. Conversely, another person’s touch generates no preparatory motor efference copy or forward sensory forecast, arriving unattenuated at full perceptual salience.
The Robot That Gave the Tickle Back
Blakemore, Frith and Wolpert tested the forecast by breaking it. Participants moved a robot arm with their left hand; a second robot stroked their right palm (Blakemore et al., 1999), so the touch was self-generated , except that the experimenters could delay it or bend its path.
Undisturbed, self-produced strokes were rated less tickly, less intense and less pleasant than identical strokes from the machine alone. Delaying the tactile delivery by 100 to 300 milliseconds progressively disrupts the efference-copy match; even an artificial lag as brief as a tenth of a second (100 ms) significantly restores perceived ticklishness, with sensory attenuation largely dissolving by 200 to 300 milliseconds (Blakemore et al., 1999).
Bending the stroke away from the predicted path did the same thing: delay and trajectory perturbation are different kinds of error, one in time and one in space, and both raised ticklishness. What mattered was the mismatch, not who owned the hand.
An earlier imaging study from the same group had already located the dampening: externally produced touch activated somatosensory cortex more strongly than self-produced touch, and the cerebellum was less active when a movement produced a touch than when it produced nothing , consistent with a predictor that had already accounted for the expected stimulus (Blakemore et al., 1998).
| Condition | What the brain predicts | What reaches awareness | Key Study |
|---|---|---|---|
| Own hand, synchronous | Place, timing and pressure match | Dull and distinctly unticklish | Blakemore et al. (1999) |
| Own hand, 100 to 300 ms delay | Prediction arrives, touch does not | Ticklishness climbs back | Blakemore et al. (1999) |
| Someone else’s hand | No forecast at all | Full-strength tickle | Blakemore et al. (1998) |
Why the Skin Cannot Be the Judge
The intuitive explanation is that your fingers know themselves , that there is something about your own warmth or your own pressure signature that the skin recognises. It does not survive contact with the physiology.
Mechanoreceptors are transducers: they convert indentation, stretch or vibration into spikes, with the same fidelity regardless of whose hand is pressing. A Pacinian corpuscle under your forearm cannot tell your index finger from a friend’s; the signal leaving the skin is effectively identical. What differs is what the brain does with it. The first formal laboratory notes on the puzzle appeared in Nature in 1971 (Weiskrantz et al., 1971), and the explanation had to wait for a framework that could describe prediction.
A Machine That Tickles Just Fine
If another person were essential, a machine should fail. In 1999 Christine Harris and Nicholas Christenfeld tested exactly that, and the machine passed.
Thirty-five participants were tickled twice: once by the experimenter, and once by what they were told was an automated tickle machine. They smiled, laughed and wriggled just as much for the machine, and their self-reports were virtually identical (Harris & Christenfeld, 1999). Ticklish laughter does not require that the touch be attributed to another person.
That reorders the explanation: if the body does not care who touches it, the critical variable is predictability , the gap between a touch you generated and one that arrives from outside your motor plan.
What the Cancellation Is For
Self-generated sensation is the background noise of having a body. Every step and scratch produces touch signals that carry no new information about the world; attending to all of them would waste processing on the predictable.
Attenuation is the filter that keeps the world legible. The review by Blakemore and colleagues links the dampening to reduced responses in somatosensory cortex and the anterior cingulate cortex, with the cerebellum supplying the forecast (Blakemore et al., 2000). The same review notes that this predictive mechanism appears abnormal in people who experience auditory hallucinations or passivity phenomena , experiences in which self-produced events are felt as externally caused.
That is the deeper reason the question matters: it is a clean, measurable version of a problem the brain solves continuously , deciding which signals belong to the world and which belong to you. The machinery that discounts your own footstep also shapes what you expect to see and how the brain edits time during an eye movement.
The failures are instructive too. A recording of your own voice sounds uncanny because it violates a stored self-template built from a sound only you could hear , the same mismatch logic as voice confrontation. And waiting feels slower than being busy because the brain’s estimate of elapsed time depends on what it expected to happen, which is why boredom stretches the clock.
Can You Learn to Tickle Yourself?
Partly, and only in the direction the robot experiment predicts. Add an intermediary , a paintbrush, a feather, a pencil , and the forecast degrades: the tool introduces slight mechanical lag, blunts pressure, and adds friction. While a soft brush serves as an informal home demonstration of how mechanical imprecision degrades sensory prediction, it does not substitute for the microsecond precision of a robotic haptic interface. In some individuals, this mechanical noise produces a faint ticklish sensation, particularly on sensitive regions like the sole of the foot or the side of the ribs.
What conscious practice cannot do is eliminate the sensory attenuation of immediate, direct self-touch. The forward model prediction is generated from your own motor efference copy before the finger touches the skin. While introducing artificial delays or intermediary tools can partially degrade this predictive cancellation, as demonstrated in robotic manipulation experiments, direct self-applied touch consistently engages sensory gating mechanisms.
Why Your Cerebellum Always Spoils the Surprise
Why you can’t tickle yourself is not a quirk of the skin. Receptors report indentation without identifying the toucher, and a machine can produce the same laughter a person can (Harris & Christenfeld, 1999). What your brain uses instead is a forecast: an efference copy of the motor command, a forward model that predicts the sensory consequence, and a comparison that dims whatever matches (Blakemore et al., 1998). Introduce a brief delay of 100 to 300 milliseconds, and the predictive cancellation collapses, restoring the tickle (Blakemore et al., 1999). The prediction is the cancellation; the surprise is the sensation.
References
- Blakemore, S. J., Wolpert, D. M., & Frith, C. D. (1998). Central cancellation of self-produced tickle sensation. Nature Neuroscience, 1(7), 635–640. doi.org/10.1038/2870 · PubMed (PMID: 10196573)
- Blakemore, S. J., Wolpert, D. M., & Frith, C. D. (2000). Why can’t you tickle yourself? NeuroReport, 11(11), R11–R16. doi.org/10.1097/00001756-200008030-00002 · PubMed (PMID: 10943682)
- Blakemore, S. J., Frith, C. D., & Wolpert, D. M. (1999). Spatio-temporal prediction modulates the perception of self-produced stimuli. Journal of Cognitive Neuroscience, 11(5), 551–559. doi.org/10.1162/089892999563607 · PubMed (PMID: 10511643)
- Weiskrantz, L., Elliott, J., & Darlington, C. (1971). Preliminary observations on tickling oneself. Nature, 230(5296), 598–599. doi.org/10.1038/230598a0 · PubMed (PMID: 4928671)
- Harris, C. R., & Christenfeld, N. (1999). Can a machine tickle? Psychonomic Bulletin & Review, 6(3), 504–510. doi.org/10.3758/BF03210841 · PubMed (PMID: 12198790)
