[freeflight]

Interesting, the patient only has Yes/No control, and that's then used to spell whole words and sentences.

There is a supplementary movie to the study showing this in action [0], it's slow but quite impressive when the patient spells out "Jungs es funktioniert gerade so mühelos"/"Boys right now its working so effortlessly".

That particular session came out at 171 characters in 157 minutes, 1.1 characters/minute.

[0] https://static-content.springer.com/esm/art%3A10.1038%2Fs414...

[Cthulhu_]

I did think it was going to be a very slow process, I'm sure a binary search / predictive text algorithm would work better. But that needs R&D costs, not to mention the patient would have to be taught how to use it.

[aero-glide2]

Thanks for the video! 2:27 shows the process. They go letter by letter in order and then ask yes or no. I think asking about the next letter based on previous letters would vastly improved the speed

[narcindin]

Yes! Binary search through the alphabet. Or, better yet, Hoffman encode common German words!

[ben_w]

One of my friends worked with David J. C. MacKay on Dasher: https://en.wikipedia.org/wiki/Dasher_(software)

I think that may already do what you want?

[nullc]

There are assistive typing devices that use complex models to predict characters: https://en.wikipedia.org/wiki/Dasher_(software)

The downside is that with a powerful enough model you could replace the patient with a coinflip and have a hard time telling the difference.

[eternityforest]

Is binary search really the most practical? Can most people answer questions like "Is P after M" without a lot of thought and counting out all the letters one by one starting at one of the input letters?

A basic markov chain predictor seems like it could work, if you had a short enough window to be sure it couldn't invent its own nonsense.