I started (and quit) a PhD based on the Machine Learning algorithms used for BCI systems such as this.<p>Every 6-12 months, someone will write a puff piece about their research in order to get grant money[1]/street cred. And every time, HN responds as if it's sci-fi tech from the future and I shit on everybody's dreams.<p>So here we go:<p>At a glance, nothing about this system in particular is new. It seems to solve the same problems as older invasive arrays from 15 years ago, with the same deal-breaking flaw still unfixed (namely, the body rejecting the implant over time).<p>He had OK-ish performance in a single patient (one character per minute on a good day), but single-patient performance is misleading as it's very common for a system to work with one "golden" patient but not others. Interestingly, no real research has been done (at least, none as of 4 years ago) to look into why this is the case - most "breakthroughs" are people throwing machine-learning spaghetti at the wall to cover up deliberately flawed benchmarking. (To be fair, this is more of a comment on the field as a whole than this particular trial. His results seem honest and humble.)<p>The field is going nowhere, and this article is yet another shining example of this. Nothing will ever be achieved until we abandon the current paradigm and actually learn more about the underlying neurophysiology (or possibly create batshit-insane sensors).<p>[1] They actually explicitly mention this in the article this time, which is pretty bold:<p>>Chaudhary’s foundation is seeking funding to give similar implants to several more people with ALS. He estimates the system would cost close to $500,000 over the first 2 years.
<a href="https://news.ycombinator.com/item?id=30773013" rel="nofollow">https://news.ycombinator.com/item?id=30773013</a> Previously:<p><i>The study comes in the context of past findings of serious misconduct against Birbaumer and Chaudhary. The findings concerned the data and analysis in two previous papers published in PLoS Biology. The two articles, subsequently retracted, also concerned the use of brain activity to decode the thoughts of completely locked-in patients. The German research agency, Deutsche Forschungsgemeinschaft (DFG) found that the scientists failed to show complete analysis of their data and patient examinations in these previous studies and made false statements. The allegations do not relate to the findings of the current research which involved different methodology, supervision and analysis. In a statement to Technology Networks, Birbaumer said that the new study “shows that all accusations are wrong” and suggested that additional forthcoming legal developments would further exonerate his and Chaudhary’s prior research.</i>
Has anyone else noticed this is sort of a permanent story? I’ve read it as news verbatim at least once a year since the 90s. It’s the “Russian bombers intercepted” of medical science. This article is better than most, acknowledging prior work but only to a few years.
Unbelievably hopeful and tragic at the same time. That he could learn to control electrode implants to communicate (in an albeit "rudimentary" way) is just incredible, but I can't imagine the feelings of unfairness at such a young age for him. Hopefully there's a path here to help others communicate much longer than they'd otherwise be able to.<p>One of the other sentences he made out was "I love my cool son," which hit particularly close to home given I'm the same age as well. Oy.
One more step towards Lock In[0] sort of becoming real.<p>[0] <a href="https://en.wikipedia.org/wiki/Lock_In" rel="nofollow">https://en.wikipedia.org/wiki/Lock_In</a>
Isn’t that what Musk’s Neuralink is doing? They automated the invasive surgery to install the device, and their device has thousands of wires to read neuronal signals.
A welcome upgrade to The Diving Bell and the Butterfly<p><a href="https://en.wikipedia.org/wiki/The_Diving_Bell_and_the_Butterfly_(film)" rel="nofollow">https://en.wikipedia.org/wiki/The_Diving_Bell_and_the_Butter...</a>
I’ve seen and built systems that are far less invasive —<p><a href="https://agenda.hep.wisc.edu/event/792/contributions/18594/attachments/6173/6931/Brain_Computer_Interfaces_-_DSP_of_SSVEP.pptx" rel="nofollow">https://agenda.hep.wisc.edu/event/792/contributions/18594/at...</a><p>With accuracy on par or better than described:<p>> He eventually explained to the team that he modulated the tone by trying to move his eyes. But he did not always succeed. Only on 107 of 135 days reported in the study could he match a series of target tones with 80% accuracy, and only on 44 of those 107 could he produce an intelligible sentence.<p>In fact, some of these successful efforts were done back in the early 2000s.<p>From the article:<p>> Researchers inserted two square electrode arrays, 3.2 millimeters wide, into a part of the brain that controls movement. When they asked the man to try to move his hands, feet, head, and eyes, the neural signals weren’t consistent enough to answer yes-or-no questions, says Ujwal Chaudhary, a biomedical engineer and neurotechnologist at the German nonprofit ALS Voice.<p>The speed and accuracy does take a hit using external nodes, but brain implants are frankly dangerous. They come with serious risks.