What a neuron is waiting for

Here is a picture I keep turning over. I think the neurons, individually, are responsible. They are always looking for entropy events, and entropy lives in novelty relative to the most recent history. Regardless of long-term history, novelty is what a neuron is waiting for. It is measuring the deviation from uniformity and from chaos.

To do that, it probably has to habituate to chaos and phase-lock to uniformity. That is how it identifies the novelty. A cell that keeps returning to the same state, spontaneously firing, is keeping track of a phase, and as soon as it is pulled out of that phase it has found entropy. It is almost as if polarization and calcium are what let it notice. On a single-cell level it has those bursts, and it can tell its most recent history from the bursts. It needs to, because it is watching for uniformity and watching for chaos, and the spike timing is what carries that information.

What it is looking for is signs of life. Not novelty exactly, and I still think novelty may not be the right word. It is looking for unpredictability, which is what a living organism is, and is the information that is in a picture.

One level up, and again

The other thing we see in biology is that on a cellular level, things are happening, and when you go up a level and look at thousands of cells working together, there are a lot of very similar characteristics. Only certain things have to change to make it work at the multicellular level. Then you go up to multiple circuits, multiple cortices, and it is the same again, but it needs different mechanisms for it to work.

So my thought is that at the larger system level, which is what we have been measuring for the last fifty years, the response is more synchronized, not necessarily stronger, to certain novel entropies. Probably in relation to marking them. We may be talking about two different types of novelty, or entropy: a per-cell kind, where each cell responds fairly strongly, and a population kind, where the neurons coordinate more when an event is much newer. And since we are measuring the extracellular response, that does not mean the thing is a brand-new concept to the person. Something about it is standing out, that is all.

Temporal entropy

Temporal entropy is the same thing, applied to the rate of change. Think of someone looking back on a week or a month of the same kind of day. It blends together, and it is very hard to remember anything. That is uniformity. Now think of the opposite: when everything is constantly chaotic, you lose track of what happened. You are so busy that you cannot keep hold of it.

Then there is the stretch right in the middle, where there are breaks and shifts and inconsistencies. Those moments are entropy points, and they are what make up the experience. My suggestion is that the perception of experience feels that way because that is how neurons save and recall experience. Not by pulling only the eventfulness, and not only the calm, but the changes. The entropy.

How this could be wrong

This is speculation, with nothing behind it yet. It would be wrong if a neuron’s response turned out to be fully explained by plain adaptation to rate, with no extra dependence on how surprising the input was against its recent history. It would also be wrong if the window of “recent history” did not change between cell types or states. Neither has been tested here.