Neuroscience · Brain Circuits
Your brain doesn't use its full range of possible thoughts and reactions all at once — something is quietly setting the ceiling.
A new study says that ceiling isn't random. It's wired in, one synapse at a time.
Researchers at the Jülich Research Centre, the University of Washington Computational Neuroscience Center, the Allen Institute and other institutions have found that the structure of local neural circuits strongly shapes the "dimensionality" of brain activity — essentially, how many independent patterns a group of neurons can express at once. Using high-density Neuropixels recordings from the mouse visual cortex, the team found this dimensionality isn't fixed; it shifts over time, and specific wiring patterns called network motifs appear to control that shift. A separate analysis of more than 32,000 synaptic connections in mouse and human brain tissue found the same wiring motifs in both species, suggesting a shared principle rather than a mouse-specific quirk. The findings are published in Nature Neuroscience.
Local wiring patterns among neurons, not just their sheer number, appear to set how many activity patterns a brain region can produce.
Your Brain Doesn't Use All Its Freedom at Once
The human brain holds billions of neurons, each one a tiny cell that receives, transmits, and processes information through electrical and chemical signals.
But thoughts, sensory perception, and behavior don't come from any single neuron acting alone. They emerge from collective patterns of activity spread across whole populations of neurons.
That raises a genuinely hard question: how many different patterns can a given population of neurons actually produce?
Researchers call this dimensionality — the number of independent activity patterns, or degrees of freedom, a group of neurons can exhibit.
A team at the Jülich Research Centre, the University of Washington Computational Neuroscience Center, the Allen Institute, and other institutions set out to find what actually controls it.
Their findings, published in Nature Neuroscience, point to something surprisingly concrete: the physical wiring of local neural circuits.
What "Dimensionality" Actually Means in a Brain
Dimensionality sounds abstract, but the underlying idea is fairly intuitive once you picture it in motion.
High-dimensional activity can encompass many distinct patterns — a broad repertoire of possible states a neuron population could settle into.
Low-dimensional activity is restricted to a much smaller range of possible patterns, no matter how many neurons are technically involved.
| State | What It Looks Like |
|---|---|
| High-dimensional | Many independent activity patterns available; a broad behavioral or perceptual repertoire |
| Low-dimensional | Activity constrained to a small set of patterns, even with many neurons active |
As the study's authors put it, the brain's collective activity is what underlies brain function — and its dimensionality is a fundamental signature of how that activity behaves.
Past research had already explored dimensionality in various brain regions and species. What stayed unclear was the far more basic question of what actually shapes it.
Why Nobody Knew What Set the Limit
It's easy to assume more neurons simply means more possible patterns — bigger population, bigger repertoire.
But that assumption doesn't hold up well under scrutiny, and the field genuinely lacked a clear answer for what else might be shaping the ceiling.
To close that gap, the researchers turned to high-density activity recordings from a very specific, well-studied part of the brain: the mouse visual cortex.
The visual cortex lies in the brain's outermost layer and processes visual information — a region with decades of prior mapping.
That existing groundwork made it a practical place to test a new, more structural explanation for what controls dimensionality.
Watching Thousands of Neurons Fire at Once
To actually measure dimensionality, the team needed to watch large numbers of individual neurons firing simultaneously, not just a handful at a time.
They used Neuropixels — tiny probes capable of monitoring the electrical activity of hundreds of neurons at once — to record activity across the mouse visual cortex.
From those recordings, they determined how many independent activity patterns were needed to describe what the neuron population was doing at any given moment.
The answer was clear: dimensionality was not fixed. Cortical activity shifted between states involving different numbers of distinct activity patterns.
That alone was a meaningful finding — it meant the brain's "range" isn't a single constant number, but something that moves depending on conditions.
The Wiring Pattern That Sets the Ceiling
The next question was the harder one: what actually pushes cortical activity from a lower-dimensional state into a higher-dimensional one, or back again?
The researchers argue that cortical areas mostly operate in what they call a "sensitive regime" — a state where the recurrent synaptic network itself plays a strong role in regulating dimensionality.
In practice, that means the way neurons are wired together, not just how many are active, has real influence over how varied the resulting activity can be.
This control, the researchers found, is expressed through specific, highly tractable features of the synaptic network known as network motifs — small, recurring wiring patterns among neurons.
Rather than dimensionality being some diffuse, hard-to-pin-down property of a huge system, it traces back to identifiable structural features that can actually be studied directly.
From Mice to Humans: Confirming It in 32,000 Synapses
A finding in mouse visual cortex is interesting on its own, but the researchers wanted to know whether it held up beyond one species.
They analyzed a separate, much larger dataset: physiological measurements of synaptic connections among more than 32,000 pairs of neurons, drawn from both mouse and human cortical tissue.
The goal was to check whether the specific network motifs predicted to affect dimensionality were actually common in real synaptic wiring — in both species.
They were. The researchers reported that "motifs impacting dimensionality are prevalent in both mouse and human brains."
That cross-species consistency matters. It suggests local circuitry scales up systematically to help control the degrees of freedom a brain network can explore and exploit.
In other words, this isn't a quirk of mouse vision. It looks like a shared structural principle running through the mammalian cortex more broadly.
What This Could Mean for Brain Disorders and AI
On its own, this is a structural finding about healthy mouse and human cortical tissue — not a diagnosis, and not a treatment.
But the researchers point to two directions where it could eventually matter.
These are described as future possibilities by the research team, not results already demonstrated in this study.
One is clinical: if specific conditions alter how a brain's dimensionality behaves, understanding the underlying wiring motifs could help pinpoint differences in activity tied to those conditions.
The other is computational: a clearer picture of how local wiring constrains or expands a brain's activity repertoire could inform the design of brain-inspired computational models.
That's a space AI researchers already care about, particularly efforts to build systems that use fewer resources by borrowing structural tricks from real neural circuits.
The researchers also note their methods could help other groups investigate how local brain circuits shape the brain's overall repertoire of activity patterns more broadly.
Manohar Insight
What stands out to me here isn't the finding that dimensionality changes over time — that, in hindsight, feels almost expected.
It's that the explanation turned out to be structural rather than statistical. Not "more neurons, more patterns," but specific, small, repeating wiring shapes doing the actual regulating.
That's a satisfying kind of answer, because it's testable and traceable rather than an emergent property nobody can point to directly.
The cross-species confirmation is the part I'd watch most closely going forward. A pattern that shows up identically in mouse and human synapses is rarely a coincidence.
It suggests evolution converged on this particular wiring trick because it works — which is exactly the kind of principle that tends to generalize well.
I'd treat the clinical and AI applications here as genuinely early-stage. The paper hands future researchers a mechanism to investigate, not a diagnostic tool or a chip design.
The real signal isn't that your brain has limits. It's that those limits are wired, specific, and — for the first time — something scientists know where to look for.
Join the Signal
If local wiring sets how many "modes" your brain can operate in, could learning or practice be partly about reshaping that wiring rather than just strengthening it? Share your take in the comments.
Discussion