"For the first time, this study demonstrates that we can generate entire spoken sentences based on an individual's brain activity.
In experiments with five people, who read hundreds of sentences, listeners were able to discern what was being spoken up to 70% of the time when they were given a list of words to choose from.
The participants in the study were told not to make any specific mouth movements.
Prof Chang said: "There were just asked to do the very simple thing of reading some sentences.
I am guessing this isn't quite there yet perhaps on a number of issues. However some issues arise for which I think it would be helpful to hear what the judge thinks. I favour adding such thoughts to the judges statement so that anyone looking at the claim can make reasonable attempt to understand what the judge is likely to require.
1. The claim says "Detecting and interpreting the effect of the mind on body systems outside the skull, for example detecting subvocalizations or increased perspiration, does not satisfy the claim." This study seems to be using subvocalizations however it is not detecting actual mouth movements it appears it is detecting brain processing only. So is the claim contradictory or ambiguous for giving subvocalizations as an example of detecting a effect of the mind on body systems outside the skull? Is the claim just saying detecting mouth movements is a prohibited approach but detecting brain processing of movements required is within the skull and therefore not prohibited by this clause? Is the approach used in linked study prohibited or not?
2. If this works while subject is reading sentence but not while thinking (but not reading) or maybe it just gets confused while chewing; does this disqualify this approach? Should such failures of the approach count towards permitted failure rate?
3. Claim requires "researchers must identify the majority of morphemes in the subjects' sentences". "Up to 70% of the time when they were given a list of words to choose from" sounds short of this requirement. Presumably being given a list of words to choose from is not acceptable? Different metrics of success from the requirements of claim may cause issues but I guess the judge decides how to deal with that when/if we get further metrics from other studies that get closer to meeting requirements.
4. Does being told 'not to make any specific mouth movements' constitute training which the claim prohibits?
5. If the purpose being considered is dealing with speech impediment then the subject probably wants to be understood is willing and the no training prohibition makes little sense to me. If the purpose is more along lines of lie detection then the subject might be less co-operative. How willing the participant is seems likely to have large effect on success rate. Is the 50% success rate required with willing or unwilling participants?