https://www.dfg.de/service/presse/pressemitteilungen/2019/pr...
The two researchers were investigated for lying in their papers. Selected translation of the linked article.
>The DFG investigative committee found false information in three cases. Accordingly, the two scientists, contrary to what they described in the 2017 study, only partially recorded the examinations of their patients on video. In addition, data from individual patients were only evaluated in summary form and not broken down. Overall, a depth of data was conveyed that de facto did not exist.
it is likely very possible to do what they have claimed to have done.
the big problem has often been intersession stability. some days are good, some days are bad and some days don't work at all.
> We also validated the Yes/No responses in a question paradigm, in which the answers were assumed to be known to the patient.
> Finally, in an auditory speller paradigm, the patient could select letters and words using the previously trained Yes/No approach.
As somebody working in the German medical sector, I'm also pretty certain that the BfArM ain't casually greenlighting scam brain surgeries/implants. Nor am I sure how any of the researchers are supposed to scam the patient, or their family, with this?
From that, we learn that there's indeed an element of interpretation, as was seen with Koko the gorilla and the claims that she could talk, when in fact she was just spouting gibberish that her handlers would interpret.
I anticipate more scientists are going to become disgraced as this study is scrutinized by peers.
But, hey, I hope I'm wrong. It would indeed be nice to talk to locked-in loved ones.
One controversial incident? This is an extremely common type of scam. There's always somebody claiming they can speak to your comatose or deceased loved one, and they're always scamming you. No, the coma communicators aren't more reliable than the ghost communicators. No, there is no reason you would ever expect someone in this field to be telling the truth, particularly not when they've been caught running the same scam in the past.
Was that also discredited? Is there any standing evidence of communication by the locked-in?
the user has a screen in front of them where the alphabet is cycled through at about 0.5 ch/s. scalp electrodes detect a signal associated with novelty or surprise allowing the user to spell out communications, very slowly.
edit: the cycling is in random order, thus exploiting the p300 surprise signal when the right character is highlighted. there's also some manipulation of spatial location to enhance the evocation of the signal.
man that's a scary last paragraph. Locked-in research is very, very sensitive to this sort of thing, for obvious reasons.
At first he thought the nurse was just being unsocial when they came to do routine stuff with him! and then he realized he was in spectator mode
Its going to be really hard for a lot of people when they find out whats going on with their locked in or previously locked in loved ones
Sometimes I have sleep paralysis (very rarely) and it's the worst, scariest feeling.
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...
Even to enable the person to express their wish for euthanasia. I don't personally agree, but respect 100%.
I wonder if nerds like us could optimize the „bitrate“ of this.
Lvng out mddl vcls is one vry smpl optn, fr exmpl.
There must be even better ways, if you consider the frequency of words, for example.
I wouldnt be surprised if one could reduce bits needed to express a sentence by 90%, if those numerologists that proved their worth with predicting Wordle solutions had a look at that.
axiosgunnar's profile says he is from Riga.
On Earth, right now, 735,000+ people with CI implants are boasting 24-nary brain wires.
CI = Cochlear implant.
It's interesting that the "audio" approached worked. There are people who believe that listening to sounds at various "brainwave frequencies" (alpha, beta, etc.) can cause your brain to enter the states associated with them. While intriguing, I thought it was probably nonsense. Well, maybe not....
but practically let the relatives decide, and when the patient can communicate, check the consent later. or find patients which can consent before losing communication.
