The Verification Venue · a sense you can learn
The Click That Maps the Room
Some blind people navigate by clicking their tongues and listening to the echoes bounce back — judging the size, distance, and motion of things they never touch. The popular story calls this a rare gift. The research says otherwise: it is a learnable skill, and in expert echolocators the returning echoes are processed in the visual cortex.
Here is the piece of it you can do right now. Press play. You'll hear a sharp click, then a quieter echo a moment later — exactly the way a surface a few metres away would send one back. Guess whether the surface is near or far. Then the page reveals the delay Δt and turns it into a distance, live, using the one bit of geometry every echo obeys.
Echo delay Δt
10.0 ms
the gap between click and echo
Implied object distance — d = c·Δt/2
1.715 m
move the slider or roll a mystery echo
Below about 1–5 ms an echo can fuse with the click (the precedence, or Haas, effect) and you hear one sound. This demo uses 5–30 ms delays, but these are not guaranteed to sound separate: fusion varies with the stimulus.
The geometry is the whole trick. A click leaves your mouth, travels to an object, and comes back — so the echo covers roughly twice the distance to the object, 2d. The delay you hear is that round trip divided by the speed of sound: Δt = 2d / c, which rearranges to d = c·Δt / 2. That is the number the page computes in front of you.
Evidence A · trainability
Novices learn it in ten weeks
In a 10-week program (20 sessions of 2–3 h), 14 sighted and 12 blind participants aged 21–79 practised click-based echolocation. Both groups improved substantially on size, orientation, and navigation tasks; group-average accuracy on a size-discrimination task rose from 0.54 (session 1) to 0.79 (session 20). Some trained novices matched 7 long-time expert echolocators on certain tasks.
"Neither age nor blindness was a limiting factor in participants' rate of learning."
Norman, Dodsworth, Foresteire & Thaler, PLOS ONE 16(6):e0252330 (2021).
Evidence B · the brain rerouting it
Echoes light up the visual cortex
Two blind expert echolocators (EB, early-blind; LB, late-blind) were scanned in an fMRI study while listening to recordings made with microphones placed in their own ears. Contrasting echo-containing recordings with echo-free ones, the authors report that "activity disappeared in EB and LB's auditory cortex, but remained in calcarine cortex" — primary visual cortex. The two sighted controls showed no such recruitment.
Thaler, Arnott & Goodale, PLOS ONE 6(5):e20162 (25 May 2011), DOI 10.1371/journal.pone.0020162.
Caveat, stated plainly: this imaging result is a two-participant study (plus two sighted controls) — case-scale evidence, not a broadly sampled one.
The check — the delay-to-distance geometry, recomputed live
The distance above is not asserted; it falls out of d = c·Δt/2 with a named speed of sound. Below, the current value plugged through, then a small table you can re-check by hand at c = 343 m/s:
| Δt | Δt (s) | d = c·Δt/2 |
|---|
Honest scope. This demonstrates the geometry — delay in, distance out — not the full perceptual mechanism. Real echolocators also read loudness, pitch/timbre, and spectral colour, and the ear does not literally stopwatch Δt. The speed of sound is a named free choice: it depends on temperature (343 m/s at 20 °C here; ~340 at 15 °C, ~331 at 0 °C), and d = c·Δt/2 assumes mouth and ears are effectively co-located (~0.1 m apart, tiny next to metres of object distance) and the object roughly ahead. Re-run it: node research/can-blind-people-echolocate/verify-can-blind-people-echolocate.mjs.
What's proven, what's assumed, and what this demo is not
Proven / cited. Human click-based echolocation is a documented, learnable skill (trainability from the 2021 training study; the 2011 paper also states echolocation "is a trainable skill"). In two blind experts, echo-containing recordings, compared with matched no-echo recordings, increased primary visual (calcarine) cortex activity without a detectable auditory-cortex difference (2011). The two findings rest on separate studies — trainability on the 2021 program, the visual-cortex recruitment on the 2011 imaging — and this page attributes each to its own source.
The physics is exact; the perception is not this. The formula d = c·Δt/2 is exact round-trip geometry for a co-located emitter/receiver and an object ahead. But it is not a model of how a person perceives distance. Real human echolocation exploits loudness, spectral timbre, and binaural cues; very short echoes (roughly 1–5 ms) can fuse with the emitted click via the precedence (Haas) effect. The demo uses 5–30 ms delays, but fusion varies with the stimulus, so this range does not guarantee separate sounds. It never claims your ear is timing the gap.
Small sample, named. The calcarine-cortex result is n = 2 blind experts (one early-blind, one late-blind) plus two sighted controls — a contrast result (echo minus no-echo), with the visual-cortex recruitment appearing in the blind experts, not the controls. It is compelling case-scale evidence, and it is only that.
What we couldn't verify here. We did not re-run the fMRI analysis or the training trials — those numbers are quoted from the peer-reviewed papers. What this page's verifier does independently recompute is the delay-to-distance arithmetic and the reported sample sizes and improvement figures for internal consistency.
Corrections
Corrected 2026-10-14: The page said click-echoes activated visual cortex rather than auditory cortex. In the two blind experts, both regions responded versus silence; the echo-minus-no-echo comparison showed increased calcarine activity without a detectable auditory-cortex difference (Thaler et al., 2011); found by the assay line, 2026-10-14.
Corrected 2026-10-14: The page called delays of 5 ms and up clearly separable. Reported click echo thresholds span 2–10 ms and depend on the stimulus, so the demo's 5–30 ms range does not guarantee separate sounds (Bianchi et al., 2013); found by the assay line, 2026-10-14.
Corrected 2026-10-14: The supporting source record quoted "either physical interference" from Thaler et al. (2019), p. 508. The paper says "other physical interference"; found by the assay line, 2026-10-14.
Corrected 2026-10-14: The relationship note said the mind reads distance via d = c·Δt/2 and builds a room map by timing sound. The formula illustrates the demo's geometry; the cited distance study leaves the acoustic cues people use unresolved (Thaler et al., 2019, p. 508); found by the assay line, 2026-10-14.