I recall reading in the mid 90’s an article linked by DrudgeReport.com describing how they had linked electrodes to a rat’s brain and, when they showed the rat a simple image of a sold circle or square, the electrodes would convey sufficient optical information that the shape would be drawn ‘electronically’ as an image on a monitor.
It was amazing and had a vague description about how it could help the public. But really, after reading it and talking with others, it seemed to be ‘mind reading’ technology. The article elaborated that it could draw images a person ‘saw’ while dreaming. What’s in the mind is recorded by electronics.
In my very limited understanding of neurology, I think there are some things that can be measured rather consistently, and many things that would have to get mapped rather extensively before they could be read. For example, if you could read the impulses traveling through the optic never to the occipital lobe, I would imagine (no pun intended) the signal would be relatively similar from person to person. However, the ‘engrams’ of memory are going to be very distinct and diverse. Even how language and vocabulary are stored and recalled, while likely using the same neurological apparatus from person to person, would be written (no pun intended) very differently in each person.
You would need a very deep and intricate mapping of the neurology of any individual in order for the surveillance to work. This would have to be very interactive, I should think. A passive scan, and a remote one, can’t possibly get you that map