Miniaturization of nerve cells

Question

Let's not look at other avenues for increasing intelligence in an animal of a given size. What changes to known biology are necessary to allow individual nerve cells to be significantly smaller and consume less energy without reducing their processing power? I know insect neurons are (at least sometimes) much smaller than human neurons, but I recall reading that those tiny insect neurons have far fewer dendrites. I vaguely recall calculating that the volumetric density of dendrites was almost the same in insects and humans.

I've seen What is the smallest sentient being possible? (and its linked question), Can we scale down humans to the size of a mouse? and Scientific Accuracy - How can we make an alien brain more 'efficient' than a human's? The first two questions and their answers aren't focused enough, and even in the third, only one of the answers starts to get at my specific question. Note that making neurons faster is one method of increasing a brain's processing power, but that in itself is not an answer to my question. Being faster doesn't let you do all the things having more components does; most obviously, it does nothing to increase memory capacity.

Clarification: I meant "change" in the sense of "difference from Earth biology" rather than "modification." My main question isn't "How to engineer Earth humans or animals like this?" (though that would be an interesting bonus) but "What naturally-occuring alien biology could be like this?" That said, alien biology can be extremely different, and how close can you get to Earth biology while achieving this?

Answer 1

The C. elegans nervous system is a mere 302 neurons, each of which are only a few microns in diameter.. Compare this humans which have bodies typically a few tens of microns (thousands of times more volume) and have roughly a billion times as many cells. With these 302 neurons they can find food, mate, and learn to run simple mazes. While we still don't really understand even this simple system, which encodes information and acts much differently to our own, it might be worth using as a starting reference point: You're not going to get a sentient system any smaller than this, but it leaves open the possibility of having a brain much smaller than our own that was optimized for tasks that humans tend to call intelligent.

Answer 2

I’m sorry, you’re trying to miniaturize something already the size of a single cell?

Sounds awesome, and doable!

The issue is that the “quality” of neurons are not very high. They do indeed use some electrical components, but a lot of it is chemical too, and that gets in the way of making them smaller: it becomes progressively harder and harder to stop crosstalk between neurons that are next to each other, as the chemicals may leak between cells, and storing the chemicals gets more and more difficult as well.

Nanotech might be the way to go: imagine a lump of nanobots that connect to each other like neurons do, but use tiny electrical or magnetic signaling instead of chemicals. Not only does this “nanoneural brain” work in the same way, but nanobots can be made extremely small (potentially on the order of only a few hundred atoms across), and crosstalk can be more efficiently controlled for.

Here’s the issue: nanotech isn’t biological. But silicon-based life might be able to evolve a similar structure naturally. All that needs to happen is for a new kingdom be introduced that runs on silicon instead of carbon, changing the new species’ brains from regular ol’ neurons to ultra-compact, high-efficiency nanites.

Is this a huge change from regular Earth biology? Well, not really; silicon and carbon are chemically pretty similar, it’s just that carbon is more common because of its participation in the CNO cycle of stars (whereas silicon is more of a dead-weight element like iron). So, find a way to make DNA out of silicon and other metalloids and you’re good to go.