Are Neurons the Brain’s Cognitive LEGOs?

An intriguing mouse study suggests that the brain can reuse neural building blocks. What might that mean for learning, habits and recovery after injury?

For years, I pictured learning as a path across a field.

Walk it once, and there is barely a trace. Walk it repeatedly, and a route begins to appear. In the brain, I imagined, practice strengthens a pathway until a task becomes easier.

There is truth in that picture. Experience can change how brain circuits work. But a recent study made me wonder whether I had been overlooking another part of the story.

Perhaps the brain does not always have to build a new path. Perhaps it can reuse some of the machinery it already has.

That thought led me to a question: Are neurons the brain’s cognitive LEGOs?

What the researchers found

In a study published in Nature Neuroscience in August 2026, researchers from MIT and Princeton trained mice to compare two tones. The animals had to remember the first tone, hear the second and then respond according to whether they matched.

While the mice performed the task, the researchers recorded activity in two brain regions: the medial prefrontal cortex and the posterior parietal cortex. They then analysed patterns of activity across groups of neurons.

The striking finding was that some of these patterns, called neural subspaces, could be used again at different points in the task. Activity associated with processing a stimulus could represent a new stimulus later on. Activity associated with maintaining information in memory could hold different kinds of information.

In the prefrontal cortex, a group of neurons involved in remembering a tone during one part of the task was later involved in remembering a planned action. The researchers also studied artificial neural networks fitted to aspects of the recorded activity. Disrupting particular groups in those models interfered with particular computations.

This suggests a useful way of organising a complex brain: some neural components can perform a similar computational job with different information.

The study was led by Yuma Osako. Its other authors are Greggory R. Heller, Sofie Ährlund-Richter, Timothy J. Buschman and Mriganka Sur.

Why the LEGO comparison needs care

I like the LEGO analogy, but a neuron is not a tiny plastic brick that can be lifted from one circuit and snapped into any other.

Nor does the study mean that a single neuron stores a sound, deletes it and then stores a plan like a computer file.

The researchers examined patterns of activity across neural populations. Their findings point to specialised, reusable features of those patterns during a particular working memory task.

So the analogy works best if we think of the brain as reusing ways of computing, rather than moving individual neurons around at will.

It is an exciting result. It is also a result from mice doing one carefully designed task. It does not establish that every human thought, memory or skill uses the same arrangement.

What happens to the old idea of “neural grooves”?

This was my first question when I read about the study.

If neural components can be reused, does that make the familiar idea of strengthening a pathway through practice obsolete?

No. Learning and repetition can still change neural connections and the way circuits operate. Reuse and plasticity can fit together.

A network may already have components capable of taking part in a task. With experience, the connections and activity patterns involved in performing that task can change. The same resources can be recruited more effectively.

My path across a field is still a helpful picture, provided I do not take it too literally. The brain is more dynamic than a fixed set of tracks carved into the ground.

Crucially, this mouse study did not test habit formation. The connection between its findings and the habits we build in daily life is an interesting question, not a conclusion the researchers demonstrated.

Does a habit really take 21 days?

No universal rule says that a behaviour becomes a habit on Day 21.

The popular number is commonly linked to an observation by plastic surgeon Maxwell Maltz about people adjusting to changes in appearance. It should not be treated as a measured deadline for every habit.

In a study of everyday behaviours, Phillippa Lally and colleagues found considerable variation in how long it took participants’ reported behaviour to approach automaticity. A later systematic review of health behaviour studies also found wide variation. In the four studies that reported a time to habit formation, the reported medians were about 59 to 66 days. Individual estimates varied much more widely.

So if a behaviour does not feel automatic after 21 days, do not get frustrated or give up. Keep practising in a consistent context. Even at 66 days, there is no deadline you have failed to meet — automaticity develops at its own pace.

The behaviour, the person and the circumstances all matter. There is no switch that flips in everyone’s brain after three weeks.

An even bigger question: what if the brain is injured?

Reading about reusable neural components raised another question for me.

If a stroke or injury damages an area involved in a particular ability, could surviving parts of the brain help take on some of its work?

Sometimes, to a degree. Brain injury research and rehabilitation show that surviving networks can change. Remaining pathways may be recruited differently, and practice can support recovery or help a person develop ways to compensate.

But “reusable” does not mean “interchangeable”. A neuron elsewhere in the brain cannot simply be assigned to replace any damaged neuron. Brain regions differ in their connections and roles. Recovery varies with the location and extent of the injury, the pathways that remain and the function affected.

This distinction matters because the new MIT-Princeton mouse study did not investigate brain injury or rehabilitation. It cannot tell us whether the modules it identified would restore a lost function after a stroke. My injury question is prompted by the study, while the evidence about recovery comes from separate research.

Neuroplasticity offers real possibilities. It is not a promise that every lost ability can be restored.

What I take away

I began with an image of learning as repeatedly walking the same path. I now picture something more flexible: a brain that can draw on existing components, use them for different information and change through experience.

The mouse study offers evidence for one part of that picture: reusable computational activity during a working memory task. Habit research tells us that automaticity develops at different rates. Brain injury research shows that surviving networks can sometimes reorganise.

These are related ideas, but they are not one single finding.

And perhaps that makes the brain more fascinating. It can be both reusable and changeable, capable of drawing on what is already there while experience shapes what it can do next.

—Ash

Research and further reading

Comments

Post a Comment