The Science
Eight things science already knows about how a person works — and the reason you haven't seen them in one place.
Everything on this page is someone else's finding, already published, already peer-reviewed. Nothing here asks you to accept a new theory, and nothing here is ours.
The only thing that is ours is the order.
The assembled picture
You act on a tiny fraction of what you take in
Your sensory systems gather information at roughly a billion bits per second. The behaviour that comes out the other side runs at closer to ten bits per second.
That gap of eight orders of magnitude is not a rounding error. It is the central fact about being a nervous system: almost everything that arrives is dealt with without you.
Zheng & Meister, 2024 — who are careful to say that how the brain bridges the gap remains an open question.
The selecting happens before you are told about it
Wrapped around the thalamus is a thin sheet of inhibitory cells — the thalamic reticular nucleus — positioned so that almost everything heading for the cortex passes it first. Beneath the cortex, basal ganglia loops select which action proceeds. The superior colliculus decides where your eyes go before you decide to look.
These systems work on timescales of tens to hundreds of milliseconds. Deliberation is far too slow to participate. You do not experience the selecting. You experience the result.
Halassa & Acsády, 2016; Frank, 2005; O'Reilly & Frank, 2006.
And the result arrives with no record of how it was made
Your blind spot. Where the optic nerve leaves the retina there are no photoreceptors at all. You do not see a hole. Surrounding texture and colour are filled in — and the filled-in patch arrives carrying no label saying it was invented.
Anosognosia. After certain right-hemisphere strokes, people are unaware that a limb is paralysed. Local error-monitoring is intact — the information exists — but it does not reach the belief the person holds about themselves. Noticing a deficit turns out to be a separate operation from having one.
Aphantasia. Some people have no visual imagery whatsoever, and many do not discover this until adulthood, because there is no signal for an absence.
Kirsch et al., 2021; Dance et al., 2022. The general phenomenon — mistaking a filtered reading for the world itself — is described in the psychology literature as naive realism (Pronin, Gilovich & Ross, 2004).
All of this runs on energy, and the supply is finite
Your brain is about 2% of your body weight and takes about 20% of your resting energy. It does not idle: total consumption stays close to constant whether you are resting or working hard. What changes is allocation, not amount.
And the supply is not the same in everyone. Direct blood sampling across the brains of 239 healthy adults found the non-oxidative fraction of glucose use varies roughly fivefold between people, and is stable across ages 19–45. It is not constant across a life either: in childhood the brain takes up to two-thirds of resting metabolism, peaking around ages four to five, and bodily growth slows to pay for it.
Raichle, 2015; Duffy et al., 2026; Kuzawa et al., 2014.
The most expensive cells are the ones that impose order
A specific class of inhibitory neuron — parvalbumin-positive, fast-spiking — is among the most metabolically demanding in the cortex: densely packed with mitochondria, dependent on oxygen, expensive to run. These cells generate gamma rhythms, the fast oscillations associated with attention and with binding a scene together.
When researchers engineered an energy deficit into only these cells, sensory gating failed. The cells that do the filtering are the ones least able to tolerate a shortfall — an uncomfortable design, and not a theory but a measurement.
Kann, 2016; Inan et al., 2016.
A physical structure closes the window on change
Late in development, a lattice-like scaffold of extracellular matrix condenses around those same cells. These are perineuronal nets, and their appearance marks the closing of critical periods — the end of the window in which a circuit is easily rewritten. Dissolve them enzymatically in an adult and juvenile-type plasticity returns.
They are also armour: they shield those metabolically fragile cells from oxidative stress. The same structure that ends the window is what protects the cells inside it. And in aged rhesus macaques, net density predicted memory performance better than the animal's age did.
Gray et al., 2023; Reichelt et al., 2019; Carceller et al., 2023.
Inflammation can take them apart
Microglia — the brain's immune cells — engulf these nets. Depleting microglia prevents the loss, and the nets go before the neurons beneath them do. Critically, the trigger does not have to be in the brain: peripheral inflammation is sufficient.
One experiment runs the whole sequence: peripheral inflammation activates an enzyme, the enzyme degrades the nets, the enwrapped cells lose their inputs, gamma rhythms fall, and the animals become cognitively impaired — and blocking the enzyme reverses it.
Mild respiratory COVID produces persistent immune reactivity in the brains of mice and humans, and influenza produces the same early pattern. Disrupting net components in those cells produces imbalance between excitation and inhibition, and spontaneous seizures.
Crapser et al., 2020; Zhang et al., 2022; Fernández-Castañeda et al., 2022; Okur et al., 2024.
And every one of these findings lives in a different building
The bandwidth gap belongs to sensory neuroscience. Thalamic and basal-ganglia selection to systems neuroscience. Anosognosia and aphantasia to clinical neuropsychology. Energy budgets to physiology. The cost of inhibitory cells to cellular neuroscience. Perineuronal nets to developmental neuroscience. Microglia eating matrix to neuroimmunology. Post-viral cognitive change to infectious disease.
Different journals. Different conferences. Different funding bodies. Different vocabularies for the same cell.
Each field holds one link of a chain, publishes it, and moves on — because assembling the chain is nobody's job, sits in nobody's budget, and belongs to nobody's department.
What this adds up to
A person receives a heavily edited version of what reaches them. The editing happens before awareness and does not report itself. It is expensive. The cells that do it are the most fragile in the brain. A physical structure protects those cells and simultaneously locks the settings in place. And inflammation — from an ordinary infection, somewhere else in the body — can take that structure apart.
Every clause has a citation. None of it is contested.
What has not happened is anyone standing far enough back to notice that these are the same story. That is not a failure of the science. It is a feature of how science is organised — and it is the reason a picture assembled from entirely uncontroversial parts can still be unfamiliar.
Where the work stands
The framework is set out across two monographs. The second does not replace, retract or dispute the first — it gets granular on the machinery the first one named.
Monograph V1.1 — the taxonomy
Establishes the five-loop Saliency Arbitration Network, the five phenotypic landmarks, the Two-Axis framework, and their alignment with aspect-level personality structure. It answers what the configurations are.
Published. Gravestock & Yablon. DOI 10.5281/zenodo.22002985
Monograph V2 — the machinery and the measurement problem
Begins from a practical question about the first: how could those loops be observed? Specifying candidate substrates exposes both the promise and the limits of introspection, behavioural assays, metabolic measures and neuroimaging. The loops are retained; their proposed implementation and measurement requirements are made explicit.
It closes with eight falsifiable predictions and a specification for a conforming instrument.
In preparation. The two works address different levels of the same architecture and are separately citable.
A note on scope
It does not tell you what any of this means for personality, for psychiatric diagnosis, or for you specifically. Those are interpretations, and interpretations are exactly what this page is avoiding.
Our own framework makes claims of that kind, and they are argued at length in the monographs with their evidence and their falsifiers attached. They are not on this page, because the moment a page like this makes a proprietary claim it stops being a summary and becomes an advertisement.
If you want to know what follows from all this, that is the right question — and it is a different document.