All right, what I have in store for myself and potentially for you, provided you are here, is some of the biggest breakthroughs in science. I am going to delve into physics, cosmology, mathematics, biology and neuroscience. And what I hope to achieve here is to be able to explain these new findings, but firstly, prime you with the necessary information to understand it, and what I hope to achieve from there is to get your feedback about what areas you'd like me to explore, because all of these areas are very interesting to me, and my problem is I need to direct my Attention to one of them at a time. Obviously, but it would help if you just tell me which ones you're interested in,

and decision will be made.

So let's start with the brain.

Neuroscience, and the nervous system is not just the brain, but all of its connections. It connects and extends to our entire body. Now what is very interesting, and I have learned relatively recently, is that communication doesn't necessarily depend on the nervous system, whilst it is a superpower in the sense of control, cells can communicate without nervous input, and this makes sense. The brain and its connections can't do everything, but they can coordinate and time what is inside our body locally and temporally so that it can best serve the rest of the body.

Now we will explore that. But

what I would like to first conceptualize are actually three concepts that are not just in neuroscience, but extend to mathematics, physics and other areas of exploration. Because when we make new discoveries, and when I say we, I mean the scientists making these discoveries,

we have to test previous theoretical models, with the machinery that we have access to and control experiments predict an outcome and measure its precision based on our predictions. Now this type of experiment is not exclusive to these scientific disciplines,

nor is it restricted to them,

theoretical models such as

the synaptic transmission of neurotransmitters. We are taught in neuroscience and in school that the synapse, which is the end of an axon, an axon is the projections of a neuron, unidirectional projection, one directional sending a message somewhere. A neuron can receive many messages from different places, but it can only send a message in one direction, provided it is sending a message and not receiving it. And we'll get into divisions of the nervous system later, but

for our purposes here,

we can think of the brain as a receiver

or an instructor.

Instructors tell the body what to do. Receivers do exactly that. They compute information about any change in the environment and what we have conscious. Access to are things like sensory, associative and motor regions of the brain,

and these integrated sensory systems provide us with the overall tone, the overall state of mind for us to continue to do what we're doing or do something else.

Now. Top down processing

is about instructing. Bottom up processing is about informing,

and they work together. This is what we call a hierarchy.

So a hierarchy exists with the top level feeding to the bottom level and the bottom level feeding to the top level. Not all things from the top reach the bottom, and not all things from the bottom reach the top. They can and often get diffused in between. Now, it isn't to say that the bottom is the unconscious and the top is the conscious, nor can we say what consciousness is at all. So let's not go there for now.

But deep internal brain structures.

One of my favorite brain structures is called the hypothalamus hypo being below and thalamus is being like Grand Central Station. It's a filter for what the bottom can send up, and also, well the top can send down. It ultimately decides what that communication looks like, the hypothalamus. And the reason why I'm bringing this up is because it's a really great example about how bottom up processing feeds top down action. So the hypothalamus detects a wide range of different levels, particularly within our blood, but also within

brain fluids.

One thing it detects, or at least is part of the system that detects the osmolarity of water, or should I say, the osmolarity of blood, being how much water is in the blood, and how hydrated we are and when we are dehydrated, all that osmolarity is in flux or chaos. The hypothalamus will send a message bottom up to produce a tone or a state that is thirst,

whether that's felt locally, like in the throat,

or an overall feeling,

it's abstract and hard To describe, feeling states

cannot be fully contained or explained

by their neurobiological underpinnings. However, it is constructive and helpful to use anatomy to guide our understanding. Why I'm bringing all of this up is that the brain doesn't minimize chaos. So chaos being flux, being unease or not necessarily disorder, but high energy. Rather, it manages surprise, and for decades, neuroscience searched for control centers in the brain or regions where decisions are made or beliefs are stored, or actions are commanded, and the intuition was hierarchical. Sensory data flows upward. Cognition happens somewhere in the middle, and motor plans flow downward. A new wave of large scale neural evidence published in Nature this year, forgive me, last year, September 2025.

Challenges this picture.

So what did they discover? Well,

the paper is called Brain wide representations of prior information in mouse decision making. Brain wide representations of prior information in mouse decision making. So. And

in association, another publication, also in September, is called a brain wide map of neural activity during complex behavior. So just read the abstract of the first one I mentioned, the neural representations of prior information about the state of the world are poorly understood. Here to investigate them, we examine brain wide neuro pixel recordings and wide field calcium imaging collected by the International brain laboratory, forget what most of that means. It is not helpful right now, but mice were trained to indicate the location of a visual grating stimulus which appeared on the left or right with a prior probability alternating between 0.2 and 0.8

in blocks of variable length.

Again, I'm just providing this here for context. Ignore it. I will bring this together. And it found that mice estimate this prior probability and thereby improve their decision accuracy. And in the report, this subjective prior is encoded in at least 20 to 30% of brain regions that notably span all levels of processing, from early sensory areas to motor regions and high level cortical regions. So high level cortical regions, you might have heard of the neocortex, which is just the new cortex. Cortical means toward the outside or the surface. So if you tap your skull, you are closest to the highest level cortical region. And

at least metaphorically,

and this widespread representation of the prior is consistent with a neural model of Bayesian inference involving loops between areas, as opposed to a model in which the prior is incorporated only in decision making areas. And so what this study offers is a brain wired perspective on prior encoding at cellular resolution, and it underscores the importance of using large scale recordings on a single, standardized task.

Okay? So brain wide representations

being activity that is significantly different

within regions not previously associated with this type of decision making and high level functioning.

Suggests that decisions

may not be so clearly ascribed to brain regions that have previously been ascribed to decision making.

So one is the homunculus.

Now the homunculus is another incredible, fascinating region of our brain, and I'm not going to go into too much depth here, because I want to stay in line. But if you want to quickly Google homunculus, please do it might help. But let's come to these theoretical models,

or the dominant theoretical framework around

brain function. So using these brain wide recordings in mice performing a simple decision tasks. Researchers have now shown that prior beliefs about the world are represented simultaneously across sensory, associative and motor regions linked together by dense, recurrent feedback loops. So in other words, the brain does not wait for information to climb a hierarchy before belief enters the system, belief is already everywhere. Now when

we talk about belief

in terms of neuroscience and brain function.

Belief is more about prediction and understanding states, but beliefs and prior information land squarely at the center of active inference, which is the dominant theoretical framework derived from the free energy principle. Again, active inference and free energy principle are very complicated areas of research, and they're mathematical. They use Bayesian inference and a whole host of different mathematical objects and expressions to describe how the system as a whole decides. So In plain terms, active inference proposes that the brain is not a stimulus response machine, nor a passive information processor. Instead, it is a prediction machine. So at every moment, the brain predicts what it expects to sense. It compares those predictions to incoming signals, and then updates beliefs or acts to reduce the mismatch it predicts, it compares, and then it updates that mismatch. Is called variational free energy. So the free energy principle includes many different concepts contained within that theoretical framework, but there is prediction comparison and updates to reduce the mismatch and the mismatch between what is predicted to be the sensory event and what actually happens is called variational free energy. So just variation, basically it's a measure of surprise or uncertainty or prediction error. And crucially, the brain does not minimize surprise by eliminating uncertainty. It actually minimizes expected surprise by maintaining a model of the world that stays useful over time.

What this study shows that theory predicted but could not prove, is that active inference has long argued that prior should be distributed context sensitive and recursively updated, distributed context sensitive and recursively updated, but empirical evidence lagged behind theory. So this study provides three decisive confirmations. One priors are not localized, they are brain wide. So these researchers found that subjective priors can be decoded, not just from prefrontal or decision areas, but from early sensory regions like the visual cortex and thalamus, the associative cortex and the motor and brainstem nuclear which goes against this whole idea of our hierarchy and where decisions and influence happen. So this means prediction is not something added after perception, but it is embedded within perception or in active inference. Terms, sensory areas are not passive encoders of data. They are hypothesis testers. So the first decisive confirmation is priors are not localized. They are brain wide. The second is that beliefs are maintained by loops, not commands, Granger causality and again, we won't get into here, but Granger, like Hermione, Granger causality, another concept analyzes revealed dense feedback loops between regions, including strong top down influences from higher areas back to early sensory cortex. It's probably a mouthful, so I'll read it again, dense feedback loops between regions, including strong top down influences from higher areas back to early sensory cortex. This architecture is exactly what active inference requires one, beliefs flow downward with predictions two, prediction errors flow upward and three, no level has final authority. Instead of a hierarchy, the brain resembles a heterarchical negotiation network, hetero. Hierarchical negotiation network.

A lot of words here, but bear with me, we're almost there.

The last decisive confirmation.

So again, the first priors are not localized. They're brain wide. Two beliefs are maintained by loops, not commands. And three behavioral priors are action based, not rule based. And this is perhaps most strikingly, the animals did not compute an abstract instead, their brains tracked an action based prior updating beliefs based on recent choices over the past five to six trials, and this aligns with active inferences. Core claim, which is beliefs are grounded in embodied action, not detached reasoning. The brain learns what it tends to do, not just what it tends to see.

Why this matters?

Well, these results are not just confirmatory. They are enabling. They could unravel a new understanding of psychiatric disorders where they can be reframed as failures of belief distribution, we can rethink attention, perception and agency. Attention is not spotlighting data. It is biasing prediction. Perception is not reception. It is controlled, hallucination, constrained by input and action is not output, it is belief, confirmation through movement. So attention is not spotlighting, data, perception is not reception, action is not output, and this dissolves classic boundaries between seeing, deciding and doing.

The emerging picture taken together,

the findings may suggest, and of course, more research is needed to

further this exploration,

a radical but coherent view of brain function could be is that the brain is a self organizing system that maintains coherence by distributing belief, negotiating uncertainty through loops and acting to keep its world predictable enough to inhabit not anarchy, not rigid hierarchy, but structured freedom under uncertainty, life persists by staying just surprised enough to keep learning and just certain enough to keep going.

Life persists by staying just surprised enough to keep learning and just certain enough to keep going. The brain is a self organizing system that maintains coherence by distributing itself, negotiating uncertainty through loops and acting to keep its world predictable enough to inhabit.

So behavior is where we will construct

a practical way of conceptualizing how we can leverage our brain function

in a way that supports our intention,

and to support intention, we need to support behavior. Behavior is really what we're after here. We want a good understanding of how we behave,

although we have an incredible picture. Now,

these findings might just help us a little further in the sense that it might not be completely just up to top down power, that is to say willpower or drive, but provide provided our conditions are different enough

giving us enough surprise

to provide enough drive and attention to guide our.

Behavior into our intention.

Transcribed by https://otter.ai