Click-based echolocation, trained on the Durham protocol. Ten weeks, twenty sessions. Blind and sighted adults aged 21 to 79 all learned it.
Every number here comes from the two Durham University studies. Nothing on this page is extrapolated from anecdote.
After training, echoes activated primary visual cortex in both blind and sighted participants. The claim that you see with your brain rather than your eyes is the one part of the folklore that imaging actually supports.
Session numbers match the study's twenty. Each week has a gate. You do not advance on time served, you advance on hitting the gate.
Palatal click production. Presence versus absence of a large flat panel. Nothing else.
Size discrimination begins at 33 cm. Reference disk against the comparison set.
The 80 x 20 cm board at four angles. Both drills now run every session.
Move the disks back, but only if the 90% gate held across two sessions.
Halfway. Add 20 to 30 minutes of free indoor exploration to every session.
Corners, alcoves, open versus closed doors. Surfaces that are not flat panels.
Parked cars, trees, hedges, building faces. Open air changes everything about the return.
T intersections, U bends, Z zig-zags. Spotter required, no exceptions.
Spaces you have never trained in. This is where room memory gets stripped out.
Untrained environment, full protocol, silent controls included. The honest measurement.
Every parameter below is the study's, not invented. Run all three drills in this order.
Two disks. Call the larger one. Your helper randomizes which side.
One board, four angles. Call which way it is turned.
Real corridors. Spotter present at all times, from session one.
Tap a session to log it. Progress is stored on this device.
Solid lines are the study's measured accuracy. The dashed line is a realistic lighter-dose projection, because this program runs roughly half the study's hours.
The study dose was 40 to 60 supervised hours. This program is 2 sessions a week plus a daily 10 minute drill. Shift every expectation right by 30 to 50%. The curve still arrives. It arrives later.
Blindfold reading demonstrations collapse under controls because trainees learn to peek down the nose gap. Run these five and your results mean something.
Contoured, no nose gap. Test it in a lit room by looking down. If you can see your own feet, the mask has already failed.
You never know the trial order. If you set up your own trial, it does not count toward anything.
Selective logging is self-deception with extra steps. The misses are the data.
Five trials per session where you make no click at all. Score above chance on those and something other than echo is leaking. Stop and find it. This is the test the blindfold schools never run.
Learning one room's quirks is memory, not echolocation. Rotate spaces or you will fool yourself for ten weeks.
Indoors before outdoors. Never near traffic, water, stairs, or drop-offs. This supplements a cane and sight. It replaces neither.
Foam board, a tape measure, a blackout mask, and someone willing to hold up disks for half an hour.
| Item | Spec | Why it matters |
|---|---|---|
| Blindfold | Contoured blackout | Removes the peeking variable that invalidates the fake version |
| Reference disk | 25.4 cm | The study's reference. Everything is judged against it. |
| Comparison disks | 5.1 / 9 / 13.5 / 17.5 / 22.9 cm | The full comparison set, cut from foam board |
| Orientation board | 80 x 20 cm | Four angles at 45 degree steps |
| Tape measure | 3 m minimum | Distance progression is measured, never eyeballed |
| Timer | 20 s cap | Orientation trials are capped, or you start guessing from fatigue |
| A helper | Occasional | Randomization cannot be self-run honestly |
The video that started this described two schools. One taught children to imagine light inside the blindfold. The other taught sensitivity to vibration. Only the second one survives contact with a controlled test, and it has a real name.