Saturday, 2 March 2024
------------------------

In the fourth chapter, we learn more about 'intrinsic brain activity', how our brain is not simply reactive, but how millions of neurons are constantly firing, and the major part of this or role is to make predictions. I feel this makes sense in how fluid our experience feels, thinking of all the tiny movements I make, writing this sentence, all the thoughts that I have. A truly streamlined process.

From what I understand, many different networks of neurons emerge, and that multiple of these neural networks can produce the same result functionally but with slight variations which they call 'degeneracy'. I stopped writing as I was curious why it is called that. From a quick google search, what I gather is it is comparable to the idea of redundancy with the key difference being the different elements can produce different results dependent on the context of an instance of execution. It seems it is a common thing in general in biology.

Now I was a little unsure what exactly was meant with that our brain simulates but I think I have a much better idea now. We have to remember our brain is isolated from the outside inside our skull and gets input from the outside through eyes, ears, tongue, and whatnot with nerves. The sensations from our inner body, like hunger, is also from the brains view an input from the outside. I write later how feelings are way for the brain to represent our inner body sensations in the context of outside body sensations, a process called 'interoception'. Now back to predictions and simulations. It would be unfeasible for our brain to work on a react respond basis in terms of speed but also structure/complexity wise if we imagine a sort of deterministic input output flow.

Instead the brain uses a 'prediction loop', billions of these in the brain expressed in the neural wiring. How they work is that instead of waiting for the next input to arrive, we draw from past experience to predict it and simulate it before the brain gets the input. The simulation being what we actually experience as a conscious being. Then when the input arrives, our brain checks how accurate the prediction was. If it matches, perfect, our brain could act much quicker, which is what allows us to react so quickly, like catching a ball hurling towards. Because just imagine how many things are involved in that to allow your body to position itself, the arms, fingers to be positioned perfectly from impact to them griping the ball in your hands. Because we use experience, we can sort of store all the expensive processing required, instead of doing that over and over, which would be very inefficient, we reuse it. This why you get better and better by experience, your brain becomes better and better at predicting whatever it is your feeding it like driving a car, but also many basic functions, you may not be aware of as they are done unconsciously like regulating and conserving energy resources in your body.

On the previous part about reusing, I think it is also like function is better made by iterative improvement of prediction for better adaptability and better use of space as it favors storing what you actually use day to day. I also think a lot of it is based on trying to see what works which requires a process with constant feedback and adjustment.

So far I only talked about when the prediction is correct, if it is incorrect, a 'prediction error', the prediction loop takes an extra step to resolve this. The step taken is in a range from prioritising the input to the experience we have in terms of it decided to change the wiring / adjustment of the prediction. Like will the brain adjust its future predictions to be aligned with the input or will it refuse to make changes which in result keeps the incorrect prediction to occur. An example of this in the book goes something like by using food like mashed potatoes and whatnot but make it look like the contents of a diaper, somes revultion or intensity of the smells they associate with the sensory input of that make them smell that and simulate that even though the smells from the physical input does not match the prediction, and this changes from person to person in accordance to their experiences and what I would imagine a combination of other factors that make one predisposed.

I imagine the rules for correcting prediction includes other factors than pure experience and input, like genes may have a role, but also that experience is not alone in terms of how often it occurs but some decided value of the experience. I'm thinking of even though most of the case it isn't my brain seems to predict a dark spot on a wall or moving dust on the floor as an insect for a split second. It also seems prone to prioritise predicting a person when I'm walking in the dark and a shape of an object with some resemblance of a person. It's funny because I keep getting tricked by one particular traffic sign on a route I walk frequent in the dark. You know you get a little shock and then realise it's not a person. I guess it will set a value to experiences for importance to survival. The brain would rather keep making prediction errors of seeing a person than risk the chance of predicting a potential threat as a nonthreat, even if it is for a split second as outcome can easily be determined by the slightest difference in reaction time.

This is how I understand prediction and simulation now as what the process is in a highly abstract and full picture sense but that there are still many uncertainties I have in the details of this model. I wonder how exactly this loop works for our thoughts and ideas we shape consciously. I'm not updated on meme culture, these days it seems you fall behind after a day or week, but I remember that things put together in an unexpected way has been a recurring theme. Is that playing with our brains predictions and the outcome that is produced when doing so?