In this metaphor, the brain is the heart, but it pumps information instead of blood. The information contains a mental model of the external and internal world that is relevant for our survival. The oxygen in this metaphor is the information carried as predictions, which are compared to the actual world data. When the prediction and the sensory information do not match, a prediction error flows back to the brain to correct and replenish the model. The cognitive system circulates information which is maintained by the brain.
The definition of consciousness used for this essay is the awareness of the model. This definition was eloquently articulated by Lisa Feldman Barrett in How Emotions Are Made: The Secret Life of the Brain. “Your experiences are not a window into reality. Rather, your brain is wired to model your world, driven by what is relevant for your body budget, and then you experience that model as Reality”. The term “body budget” could be replaced with survival for our purposes.
Barrett’s definition of experiencing a model of Reality is a high level and abstract definition. The most basic statement regarding consciousness as awareness of the mental model comes from Feinberg and Mallatt's book, The Ancient Origins of Consciousness. They propose a clear and straightforward definition of consciousness: "real consciousness is indicated by the (optic tectum) making a multisensory map of the world and then attending to the most important object in this map and then signalling behaviors" based on the map.
Back to Barrett’s term, body budget, which is why we build the model. This term is a stand-in for allostasis, which is akin to homeostasis, which means to return the baseline state. Allostasis is “predictive regulation” as explained by Peter Sterling in What is Health: Allostasis and the Evolution of Human Design. Our bodies are constantly preparing for the next anticipated action; for instance, we salivate when we think of food. The essential mechanism of allostasis is prediction, driven by a deeply subconscious and ingrained model of reality.
Cognition maintains that model based on predictive coding. The metaphor for predictions can be likened to oxygen carried by the blood. This flow of information carries predictions about what is expected as it moves from one region of the brain to the next and then further down the nervous system. Similar to the circulatory system, the pathways branch out toward various organs. As information traverses the networks of nerves, the predictions become increasingly explicit and detailed.
Let us jump ahead to the capillaries, where oxygen is absorbed by the tissues in exchange for carbon dioxide. The oxygen depleted blood cells are sent back to the heart for replenishment. In this metaphor, the capillary of the information circulation would be a prediction unit at the periphery of the sensory system. This prediction unit compares the prediction it receives from the model with the actual sensory input. When a mismatch occurs, a prediction error is returned back up the nervous system, akin to a depleted blood cell.
When a prediction error rises up to the next higher level of the cognitive hierarchy it is received by a prediction unit in that level. The prediction error represents unexpected sensory input, which is merely one detail in a local model maintained by that prediction unit. The prediction error from below may be reconciled if it aligns with a larger pattern recognized at that level. However, if the prediction error cannot be reconciled, it is forwarded to the next higher level. If the new prediction error exceeds our entire library of expectations, it rises to consciousness as an anomaly within the model. The anomaly might be resolved simply by noticing it and paying attention, which applies extra processing and can clarify the stimulus.
However, sometimes we need to use conscious effort to make sense of the new information. We need to gather new information and analyze the facts to improve our model of reality. This is the power and purpose of consciousness in the context of this metaphor.
The idea that we infer what we are seeing can be traced to Hermann von Helmholtz in the 1860s. The rationale was elucidated by Karl Friston. Predictive coding reduces the amount of energy needed by the brain. To have all of the sensory input travel up to the brain would overwhelm it. Thus the brain’s mental model is pumped through the cognitive hierarchy at a slow rate and only the information that contradicts the model needs to be pushed back up the system.
In the neocortex, prediction units are implemented using cortical columns. The neocortex is like a very large pancake that is folded up inside the cranium. Each cortical column is a microscopic collection of neurons that extend from the top of the pancake through to the bottom. There are millions of cortical columns that implement the predictive coding algorithm throughout the cortex. Each one of them has connections to the units in the layers above and below; it has neurons to make the comparisons and it has little bit of memory to contain the predictions in a local model, which are basically tiny bits of the whole mental model. The local model is shaped by our needs and desires in any given moment, which is why we can overlook things in one instance and in another, see them as obvious. We can be primed by stimulus to affect what seems obvious.
In the Predictive Coding Animation, consciousness is represented by the circle containing Model, the brain is represented by eight brain regions in the neocortex and two nuclei in the thalamus. The information being pumped is carried as predictions represented by the blue and green dots that travel along the pathways. For more information see the Expanation of the Predictive Coding Animation
2026-07-24 jch.com/notes/pc/TheBrainPumpsInfo.html <> jch