The Verification Venue · a thing your body does faster than you can think
The Reflex Too Fast to Think
When the doctor taps below your kneecap and your leg kicks, it feels like you did it — like a message went up to your brain and a command came back down. It didn't. The signal physically never reaches your brain. The whole loop closes in your spinal cord in about 18 milliseconds — far too fast for a thought to be in it.
Here's how we know, without any equipment: time it. A nerve signal travels at a known speed. The path it has to cover is set by the size of your body. So the loop has a computable duration — and that number turns out to be smaller than the fastest a deliberate thought has ever moved a muscle. Tap the tendon below and watch two signals race: the reflex that stays in the cord, and the brain-routed signal that would have to make the round trip to your head. One of them is already done before the other has left the building.
Watch the green reflex close the loop before the pink brain-route even gets to the brain.
Spinal reflex loop
18.3 ms
tendon → cord → muscle
Fastest a thought could do it
~90 ms
the reflex wins by — ms
Taller body, longer nerve path. The signal has farther to travel, so the loop takes a little longer — but never long enough to need the brain.
The big myelinated Aα / Ia fibres of this reflex run at ~50–100 m/s. Even the slowest leaves the brain no time to join in.
The loop is the simplest circuit in the body — a monosynaptic arc. The stretch is sensed by a muscle spindle in your quadriceps; a sensory (Ia) fibre carries it to the spinal cord, where it crosses one synapse straight onto a motor neuron, which fires the muscle. No interneurons, no detour, no cortex. So the time is just: the round-trip distance divided by the nerve speed, plus that one synapse.
That single line is the whole proof. Plug in a body and a nerve and you get a number in the 18–30 ms range that clinicians actually measure. Now compare it to the other number — the fastest a voluntary "go" signal has ever moved a muscle, ~90 ms, measured in the lab. The reflex finishes with tens of milliseconds to spare. A trip to the brain and back simply does not fit inside the window. The gap is the evidence: the brain was never in the loop.
The check — every number recomputed in front of you
For the body and nerve speed you set above, the loop time, plugged through live:
And the comparison that proves the bypass — even your slowest possible reflex against the fastest voluntary response, and against a hypothetical brain detour:
The default scenario (1.70 m, 85 m/s) lands on the textbook ~18 ms tendon-to-contraction figure; the whole CV band brackets the published 18–30 ms window. Run it yourself: node research/the-jerk-that-skips-your-brain/verify-the-jerk-that-skips-your-brain.mjs.
What's idealised here, and what's exactly true
Exactly true. The short-latency patellar (knee-jerk) reflex is a monosynaptic spinal arc — one synapse between the Ia sensory afferent and the alpha motor neuron, at the L3–L4 level of the cord — and it completes before any signal could reach the cortex and return. Published latencies are ~18 ms (tendon stretch to quadriceps contraction) and the short-latency stretch response is ~20 ms; the fastest voluntary muscle response to a go signal is ~90–100 ms. That gap is real and is the proof the brain is bypassed for this component.
Free choices (named). The round-trip nerve length is modelled as L = 0.882 × height, fixed so a 1.70 m adult gets ~1.50 m of round-trip nerve — the standard reflex-arc estimate (~1.5–1.8 m). The synaptic delay is taken as 0.7 ms (literature ~0.5–1.0 ms for one central synapse). Conduction velocity is a band, ~50–100 m/s for Aα/Ia fibres, so the calculator returns a band, not a single point. Different but reasonable choices shift the exact milliseconds; they never close the ~70 ms gap to voluntary action.
The honest complication. "Monosynaptic" is the short-latency story only. Real stretch reflexes also have longer-latency components: a polysynaptic spinal one and a genuinely transcortical long-latency reflex (the "M2" response, ~50–100 ms) that does route through the brain and can be shaped by intention. So the brain is not absent from movement — it's absent from the first, fastest response. The claim is precise: the short-latency component provably skips the cortex; the later ones don't, and that distinction is the point, not a loophole.
Representative, not personal. Latencies vary with limb length, temperature (cold nerves are slower), age, and where electrodes sit, so individual numbers scatter — but the ordering (reflex ≪ voluntary) is robust to any reasonable choice. The race animation advances each signal at its true modelled speed; the timeline is to scale in time.