Neuroscientists create computer program that can decode your thoughts

The opening paragraph of this article is very misleading. The published study refers to the process as "stimulus reconstruction". This is about phonetic sound processing, not mind reading. The same type of experiment has previously been performed with ferrets, and obtained similar positive results - which have nothing to do with reading the ferret's "thoughts", but show that processing of different sounds is 'mapped out' in the physical brain.

The data was obtained by making recordings of brain activity while the subjects listened to pre-recorded speech, and using a program to convert the brain activity into the expected equivalent sounds. The similarity of 'decoded words' to the pre-recorded audio was then checked and measured by comparing the spectrogram output. Word recognition was performed by a 'rudimentary speech recognition algorith' and the significant results were reported as median values of 0.89 and 0.76 where 0.5 is no better than chance. The use of the median - the middle score in a list of scores - instead of the mean - arithmetic average - is accepted as a way of minimizing the effect of extreme scores in a small sample (n=15 in this case). But turning this into the statement ".. researchers could understand the decoded words 80 to 90 % of the time" is not only abuse of statistics, it contradicts what is said by the study authors. The research publication itself clearly states that, while the recognition algorithm produced these median scores for word recognition, "... reconstruction quality at present is not clearly intelligible to a human listener".

No "mind reading" was attempted or performed in this study. Although it has been suggested that the same auditory areas may be activated by thoughts, there is no doubt at all that these areas are positively activated by sound, and in a purely phonemic way that does not depend on any "thought" process to give the sound meaning. I would expect, for example, that the sound of speech in the room would activate the same auditory areas in a person who is completely ignoring what is being heard, and is thinking about something else entirely.

The therapeutic applications which are discussed, would probably require the 'locked in' or mute subject to imagine their thoughts as audible speech, and to pace their intended speech-thoughts to the speed and rhythm of spoken words - in a very quiet environment where there is no other sound input to cause interference with the 'imagined' speech. Whether this can be done or not, remains to be tested by an experiment which is designed for the purpose...