Brain to Text Handwriting

Warren Mansell posted this fascinating article to Twitter. It describes a system for producing handwriting based on neural signals in the brain. An Abstract of the article can be found here. Warren made this comment about the article:

A rather neat demonstration that human life is about controlling our perception through whatever medium can achieve this, technological or otherwise…rather than planning specific motor actions.

I don’t Twitter – it’s just too fast on the eye (and the brain) for me – so I’m bringing this over to the Discourse site to discuss it. All I know about this work is the video of a person writing via a device that decodes their brain activity and the abstract gives an overview of the work. Based on just this, I don’t see how this work demonstrates control of perception. It could be seen as an example of generated output rather than controlled input; the brain generates activity that would produce the movements that make handwritten text if the subject had not lost the ability to move.

This could be shown to be an example of perceptual control if it were demonstrated that it is essential that the person doing the brain writing had to see the visual consequences of the imagined writing. But I didn’t see anything in the video or abstract that spoke to whether this visual “feedback” was essential. Maybe this is explained in the article itself. Is that true Warren?

Of course, even if this brain writing can be done strictly in imagination the results of this study would still be consistent with PCT; it would just mean that the device that translates brain signals into handwriting is tapping signals that are occurring in “imagination mode”; that would make this brain to text translation an even more astounding accomplishment.

And I would also like to know what aspects of brain activity are being used as the data input to the translation algorithm. If they are using data related to temporal patterns of spike firing rather than simply average firing rates that would be a finding that is inconsistent with the PCT model in its current form.

Anyway, it’s a very interesting paper and a result that would be of great benefit to people who, for whatever reason, have lost the ability to write or speak. I’d love to see the Nature article if anyone has a copy.

Best, Rick