The neuroscience of touch · 01

One Hundred Milliseconds of Too Much Precision

Why you cannot tickle yourself

Your hand moves, your skin receives the touch, and the sensation is less surprising. The brain estimates the result of the movement and attenuates the part it can predict. In one experiment, an artificial delay of 100 milliseconds shifted that prediction in time.

motor commandtactile touchillustrated gap · 100 milliseconds
Prediction and tactile feedback The blue line shows the predicted timing of touch. The coral line shows the same wave after a 100 millisecond delay. movement +100 ms

When the traces match, touch is predictable. When they separate, a prediction error appears. You can explore it below.

Tickling starts as a timing problem

Self-generated touch is usually less intense than the same touch applied from outside. When you start a movement, the nervous system also receives a copy of the motor command. That copy helps estimate where and when you should feel contact.

In neuroscience studies, this is called sensory attenuation. An internal model estimates the consequences of movement, and tactile input that matches the estimate receives less perceptual weight. It is one reason an attempt to tickle yourself usually fails.

The cerebellum is involved in prediction, according to imaging studies. Tickle is a subjective experience and also depends on expectation, attention and social context.

A timing laboratory

Move touch after the movement

The blue line is the touch the nervous system expects. The coral line is the touch that actually arrives. Drag the delay and watch the overlap come apart.

prediction comparatorvisual model · tactile feedback
Prediction and touch comparator The traces overlap at zero milliseconds and separate as the delay increases. intensity time →

In the experiment by Kilteni, Houborg and Ehrsson, participants were repeatedly exposed to a delay of 100 milliseconds. This page turns that design into an intuitive diagram.

model readout
Prediction and touch overlap

At 0 milliseconds, tactile input arrives where the internal model expects it.

The model shows the relationship between movement and feedback. It does not estimate your personal tickle threshold or provide a medical assessment.

An experiment made prediction fall behind

In a study with controlled touches, participants pressed a sensor with their right finger. The device delivered a touch to the left finger either immediately or after a delay. Researchers compared how strong the touch felt.

100 ms
the repeated delay that shifted the moment when touch could be attenuated

0 ms

The movement. The right finger presses the sensor and creates an expectation for contact on the left finger.

100 ms

The delay. In some blocks, touch arrives 100 milliseconds later. The initial prediction and feedback no longer match.

500 trials

The learning. After repeated exposure, participants began attenuating delayed touch, while immediate touch became less predictable.

The 100 millisecond figure describes an experimental manipulation. Sensitivity to delay varies with the task, duration, force and way touch is delivered.

A 2023 replication recovered temporal recalibration when it reproduced the original design with a single tested delay. In the first two experiments, tests used 0, 100 and 400 milliseconds, and exposure did not have a selective effect. A 2024 fMRI study with 24 participants found the same behavioral direction after exposure to a 100-millisecond delay and changes in responses in the somatosensory cortex and anterior cerebellum. Recalibration is supported, but its temporal precision depends on the measurement paradigm.

A palm test

You can notice the difference without a device

This test does not measure the brain. It shows why comparing self-generated and external touch requires care.

Try it in two minutes

  1. Gently touch your left palm with your right index finger, in the same place and rhythm.
  2. Repeat the movement a few times. Notice how quickly it becomes familiar.
  3. Close your eyes and ask someone else to touch the same area with a gentle, roughly matched pressure.
  4. Compare surprise and intensity on a 0 to 10 scale. Stop if anything hurts or feels uncomfortable.

Your difference may be small, large or absent. A personal result does not disprove the mechanism. In the lab, researchers control force, location, duration and order, then compare responses across a group.

Why precision matters

A sensory system that treated every consequence of its own movement as new would generate too many alarms. Attenuation reduces predictable input and leaves room for changes that do not fit the movement plan. Similar logic appears in research on self-generated touch, self-voice and eye movements, though each sense has its own circuits and limits.

The safe conclusion is narrow. There is no single tickle switch in the cerebellum and no 100 millisecond value that applies to everyone. There is a predictive mechanism that makes familiar touch less surprising and can recalibrate when the body repeatedly receives delayed feedback. Recalibration is real in some paradigms, but its temporal precision depends on how delays are tested.

Sources and limits

What the studies measured

Reader-facing method. This page separates touch attenuation from the subjective experience of tickling. The simulator keeps the relevant variable, timing, and does not turn laboratory results into a personal threshold.