Imagine that you can memorise 300 digits in five minutes. Or memorise the order of a full deck of cards in less than a minute. For most of us, that sounds like some kind of supernatural superpower. Someone who can actually do all this is Nelson Dellis, a six-time American memory champion and record holder for the most wins.

But what if you don’t just look at what this memory champion can do, but at what is happening in his brain? What would we then learn about how our memory really works?

Well, that is what researchers led by Roselyne Chauvin have done. They looked at the brain of this memory champion and mapped it in detail, using hours of brain scans and comparisons with hundreds of others.

And the first thing that comes out of this is perhaps the most surprising: Dellis does not have a “better memory” than we do. He has not increased his memory. He uses it differently.

What he does is actually quite simple to describe, and for anyone who has ever read Moonwalking with Einstein, not new, but difficult to do well. He takes abstract information such as digits, cards, and words, and transforms it into something our brain is naturally better at handling: images, places, and stories. A card becomes a person, an action, and an object. These are combined into a scene. And that scene is placed along an imagined route through a house or a city. What looks like pure memorisation is in fact a form of recoding.

What is new, however, is that you also see this in his brain. Not that the classical memory areas suddenly become more active. On the contrary. During ordinary rote memorisation, his brain looks very similar to that of others. But as soon as he uses his technique, something shifts. You then see more activity and connections in areas related to visual processing, navigation, language, and even skill learning.

In other words, he is not remembering digits. He is navigating through a story. Perhaps even more interesting: part of that process seems to be automated. Whereas in classical memorisation, the hippocampus is mainly active during storage, here it shifts to retrieval. As if the real work is no longer in storing itself, but in reconstructing the route.

This fits with a broader idea: what we often call “a good memory” is rarely pure capacity. It is rather the skill of transforming information into something that is meaningful, recognisable, and connected.

This is also where the study becomes cautiously interesting for education. Not because we should have students massively build memory palaces, but because it makes an old insight visible again: learning works better when new information connects to systems our brain is already good at. Think of images, stories, relationships, and context. This study does not fundamentally change our thinking. But it does make very concrete what we have suspected for some time: memory is not a storage space. It is a way of organising.