# Recording Human Vision

Is it possible to translate the inputs from the optical nerve to something readable by a computer?

Yes it is. The brain operates using electric impulses. Those can be recorded. However, there's problems both with capturing them, as well as interpreting them.

On the "capture" side, splicing into someone's brain is not likely to end well. Not with our current level of technology, and understanding of the brain.

On the "interpreting" side ... we don't really know where to begin.

But, assuming that certain medical and technological breakthroughs are made, yes, it would be entirely possible. Consider that we are currently able to splice hearing aids straight into people's brains so that they can overcome birth defects.

Yea, humanity rocks.

• For the "interpreting" side, once you have a good reading on the "input" side of things, it shouldn't be too hard to develop a function to map input with perception. We know that the eye sees $A$ and we see $C$ on the optic nerve. Coceptually, it's not difficult to develop function $B$ that goes between $A$ and $C$. – Green Oct 28 '16 at 19:47

Conceptually yes, if you could track the firing of each nerve (1 million of them) in the optic nerve you could determine which cones and rods were active as each photoreceptor has a connected ganglion cell. You could then map out how they fire compared to a known visual image and get a good approximation of what stimuli produces the firing pattern.

The problem is that the brain does an ENORMOUS amount of "post-processing" of our visual input, so just knowing what the optic nerve is telling the brain may not give you a very usable "TV-like" image. You would need a pretty powerful computer to process the inputs into something usable, and that system would probably have to "learn" each specific person based on a series of known visual stimuli. Plus the input of each eye is blended together at the optic chiasma, further messing up what one optic nerve sees versus what the brain sees.

There is also some eye centered "pre-processing" by horizontal and amacrine cells that can regulate the action of the bipolar cells that send impulses to the optic nerve. Just recording the optic nerve may miss out on this information, making interpretation of the data more difficult. Thus the known pattern you are showing the retina gets "preprocessed". There are 150 million photoreceptors and that signal gets reduced to 1 million nerves, so there is substantial filtering going on just in the eye. Here is a nice lightly technical summary.

I suspect we will eventually have some sort of ring shaped device we could place around the optic nerve that could read and track the nerve impulses. If it was REALLY thin you could slip it right behind the eye before the optic nerve goes into the skull. Otherwise you'd probably have to penetrate the skull and do it right after the optic nerve exits the optic canal but before the nerves blend together at the optic chiasma. Probably MUCH easier to have a contact that sits on the eye and records stuff, but who knows, with a high enough resolution MRI device maybe we could do it via a helmet?

With current technology, no. We simply don't understand the fine details of how the mind works.

Ignoring that limit, the eyes and brain are basically bio-computers (in theory). So once you know the 'base language' of both systems, you can translate between the two. (weather that translation is useful is another thing. Human vision is pretty poor by computer standards after all)