> You have, at birth, around 3,500 auditory hair cells per ear (https://en.m.wikipedia.org/wiki/Hair_cell), so 7000 inputs to process for sound. Sure your brain does a lot with that information to project a 3D sound field from two ears, but that's the limit of the input.

> The adult human eye has around 70 million cones and 75-150 million rods (https://en.m.wikipedia.org/wiki/Retina). There's a bit of crazy data compression that goes in to this, but the amount of information to process from the eyes is several orders of magnitude greater than from the years.

While you're technically correct, that's not the whole story.

There is a substantial amount of image processing that happens in the eye (I suppose it's possible that's true of the ear as well; we know much less about hearing than sight). I'm not familiar with all the research in the area, so I can't throw out numbers, but you can't just compare numbers of raw sensors meaningfully.

I agree, there's so much about the brain we don't understand, plus whole sections overlap or process multiple sources. Nothing is simple in the brain! In terms of the amount of your grain required or dedicated to each input this says nothing.

However, the statement was:

> There is more brain input from vision than any other input possible

I guess we could compare the nerves that lead from the sense organs. Again from wikipedia:

> Each human optic nerve contains between 770,000 and 1.7 million nerve fibers

> In humans, there are on average 30,000 nerve fibers within the cochlear nerve

so again, at least an order of magnitude more nerve connections from the sense organs to the brain. I think this backs up the original statement.

Interestingly the number of nerve connections from the cochlear are greater than the number of raw sensory inputs, but the eye is quite the opposite. I know the input to the brain is more of a differential coding between colours than raw data from the cells.

> you can't just compare numbers of raw sensors meaningfully.

Why not? It's certainly not definitive, but assuming a correlation between number of sensors and dataflow is certainly a reasonable starting point. If you really want to suggest the number of raw sensors is meaningless with regard to data volume, that seems like a much bolder claim. Especially when, as here, it's a 4-5 order of magnitude difference.

nice technique to focus on the raw transducers/sensory cells counts! That's fundamentally the raw inputs we're dealing with.