Not going to happen.
Don’t be to sure of that. Work has already been done on fluorescent neurons.
It’s not going to happen in the foreseeable future, at least. In order to do it, you’d have to have the map of the brain that could read the data and make an accurate prediction. You’d have to have an MRI scanner that could map the data precisely enough for the prediction software to work. You’d have to have a subject who was willing to lie still in the scanner while you performed this test, or else you’d have to have a scanner that could capture and analyze a moving image of the brain and compensate for body movements while it scanned. Walking, breathing, blinking, farting; all would have to be detected and removed from the scan data before it could be reasonably certain that it saw which nuerons fired.
It would be a neato toy, if we could build it, but it would be detecting miniscule amounts of energy at high speed and therefore be easy to jam.
From a financial perspective, it would probably spark a large growth spurt in the AFDB sector.