Anyone who looks at what happens in a brain during learning sees one striking pattern: it rarely builds something entirely new. Instead of separate, neatly defined modules like in the classic computer metaphor, the brain works with reusable building blocks that are combined in different ways each time. Think of a box of Lego, not a motherboard.

You can see this clearly in recent neuro-research with monkeys by Sina Tafazoli and colleagues, published in Nature. The animals completed different tasks in which colour was sometimes relevant, sometimes shape, and sometimes the direction of their eye movement. In their neural activity, the same small subspaces repeatedly appeared. The brain did not create a new circuit for every task. It selected, boosted and rearranged existing building blocks. As soon as the task changed, it activated different combinations, while most of the underlying components stayed the same. The brain recycles intensely, and it works surprisingly efficiently. It also echoes something Stanislas Dehaene once explained to me: the same brain region can support track recognition, face processing and reading letters.

You may now be thinking of what we call thinking schemas in education. Thinking schemas are the mental structures that allow us to organise and interpret information. They shape which patterns we notice, which links we make and which actions feel logical. In this research, you almost see a neural version of that idea. The monkeys did not have one schema per task. They used shared building blocks and combined them differently depending on how they interpreted the situation.

One of the most recognisable moments is when the monkeys first activate the wrong subspace, still thinking they are in the previous task. That looks very much like a pupil who applies an old schema to a new exercise and gets completely lost. The knowledge is there, but the wrong frame is in the foreground. Only when their sense of the task shifts does the right schema or the right combination of neural blocks come into play.

So what does this mean for classroom learning? It seems a long way from monkeys to children. First, something new usually depends on something familiar that is rearranged. Pupils link new explanations to old patterns. Sometimes it clicks immediately. Sometimes it grates because an earlier construction gets in the way. And sometimes confusion arises because the pupil thinks the lesson is about one task while the teacher has already moved on to another. Prior knowledge can support learning but also block it. The research shows how logical that is. The monkeys initially activated the wrong blocks because they misread the situation. Only when their task interpretation shifted did the brain start using the right subspaces.

It also explains why transfer does not happen automatically, even though you might expect the opposite. If the brain learns by recombining existing blocks, it needs support to see that the same components are useful in different situations. Variation in examples, making links visible and having pupils explain their thinking all help to make those building blocks more flexible. Only then is there a chance that a thinking schema not only fits the current task but also becomes usable in another context.

This perspective also helps us understand why misconceptions can be so persistent. Not because someone lacks ability, but because the brain has built a solid structure that is not easy to dismantle. Anyone who has ever taken apart a Lego creation without wanting to lose any pieces knows the feeling. A wrong interpretation can be very sturdy.

The Lego-brain idea is also hopeful. It shows that learning is not a process that starts from zero, but a system that constantly reuses what is already there. It does not require perfect instructions, but clear signals about which blocks matter at which moment. And sometimes it needs permission to take an old construction apart gently.

Perhaps we should stop talking about the brain as a computer that executes instructions and start seeing it more as a builder that cleverly reuses what is already in place. Learning is not an upload. It is rearranging, selecting, reinforcing and sometimes dismantling. The better we understand that, the better we can teach.