My memory works like a map. I store a great deal, but when I deliberately search for a name, it can remain far away. Give me a neighbouring idea, a sound or a few similar words, and I move closer. I pull the thread until I reach the right word.
Fatigue makes this system visible through its errors. I may say 'Uber' instead of 'Airbnb', or 'the sea' instead of 'the swimming pool'. The word is wrong, but it is not random: same kind of service in the first case, same world in the second. The information still seems to be there. The access path is what loses precision.
This is my personal experience. It is a starting point here, not a diagnosis and certainly not a model for every autistic person. The literature can nevertheless explain several pieces: retrieval cues, plasticity, consolidation during sleep and declining verbal fluency under fatigue.
Neuroplasticity does not create a new brain
Neuroplasticity is a broad word. It includes changes in strength between neurons, the creation or removal of connections and the reorganisation of whole networks. More than forty years of research show that the adult brain retains this capacity. It is simply more constrained than during developmental periods.
Learning electronics therefore does not grow a 'technician brain' next to a 'communicator brain'. The same networks change how they work together. A new idea becomes easier to use when it attaches to a structure already understood, then when recall is repeated across several contexts.
Human neurogenesis remains an open question
In 2018, a study published in Nature detected no young neurons in the adult dentate gyrus, a hippocampal area involved in memory. The conclusion travelled widely: adult human neurogenesis did not exist. Yet the markers degrade with delays and tissue-preservation conditions, which can erase the signal being sought.
That same year, a consensus review detailed conflicting findings and the methodological questions still open. The honest verdict is less marketable: adult neurogenesis is certain in rodents; in humans, its persistence and functional importance remain unsettled. Learning needs no verdict from this debate, because synaptic plasticity already explains it.
A memory has several doors
Tulving's foundational experiments show that a cue's quality depends on the encoding context. A weakly associated word can become an excellent cue if it was present during learning. Conversely, a generally close word may fail if it did not belong to that specific episode.
This is what Marcel Proust turned into literature with the madeleine. Its smell and taste do not contain Combray. They reactivate sensations and scenes encoded together. Proust was not a scientist, but the effect now bearing his name has been tested: odours trigger some autobiographical memories more effectively than equivalent visual cues.
| Mechanism | What changes | What not to conclude |
|---|---|---|
| Plasticity | The strength and organisation of connections | That the brain can change without limits |
| Neurogenesis | The birth of new neurons | That it explains adult human learning |
| Myelination | Fibre insulation and signal transmission | That the brain finishes at exactly 25 |
| Consolidation | The stabilisation and reorganisation of a memory | That sleeping is enough without learning |
My path did not replace my brain, it added routes
In secondary school, I hated mathematics. I had missed some foundations, especially equations and physics, then every test added pressure without giving me the why. Repeating a calculation I could not picture produced nothing. I moved into communication, a field where you connect an audience, context, intent and signs. That network made sense to me.
Later, I wanted to understand how objects work. I returned to mathematics through electricity, then electronics and 3D modelling. I had not become somebody else. I finally had objects to which I could attach the equations. I could get hold of a 3D tool in a week because each command found a place in a representation I was already building in my head.
This story does not prove that understanding why is the best method for everyone. It shows something narrower: for me, information holds when it enters a model. Without a link, I forget it. With enough links, I can retrieve it from another field years later.
Sleep consolidates, fatigue blurs access
During slow-wave sleep, the hippocampus and cortex replay coordinated patterns of activity. The review by Klinzing and colleagues describes how these reactivations help move a recent trace towards a more stable, distributed memory. Sleep downloads nothing. It reworks what was encoded while awake.
My watch sometimes displays about 2.5 hours of REM sleep out of 7.5 or 8 hours. That number cannot explain my memory. Consumer devices infer stages from movement and heart rate, and the sleep-debt article explains why their total duration is more credible than their stage breakdown.
Fatigue directly affects access. After 36 hours without sleep, a small controlled study observed lower verbal fluency and more perseveration within a semantic category. Nine participants cannot explain every one of my slips. The result does provide a mechanism compatible with what I observe: the network remains, but selection becomes less sharp.
The young brain does not finish at 25
The sentence 'the brain finishes developing at 25' turns a trajectory into an expiry date. An imaging study of 67 people aged 17 to 45 found different peaks across regions, sometimes between 34 and 38. MRI measured a signal associated with myelin, not myelin directly, and the study was mostly cross-sectional.
Myelin wraps some nerve fibres and makes transmission faster and more stable. It continues changing in young adulthood, including in regions involved in cognitive control. This means neither that a young brain is incomplete nor that an older adult cannot learn. On this point, a neurodivergent brain may finish even later, sometimes around 35, pushing the supposed deadline further still. Maturation and learning simply continue on different scales.
Empty time can let connections surface
For creativity, my most useful move is sometimes putting down my phone and doing nothing for thirty minutes. The first minutes feel wasted. Then problems that seemed separate begin to connect. I gain height because I stop adding new inputs to the network.
An experiment on creative incubation points in that direction: a low-demand task that allowed the mind to wander later improved the search for new uses for objects encountered before. It does not prove that thirty phone-free minutes are optimal. It shows that an active mental pause can help a problem already loaded into memory recombine.
Intuition is compression, not a sixth sense
When I quickly understand a limit in a project or a pattern in someone, I experience it as intuition. One possible explanation is less mysterious: patterns already encountered are recognised before I can consciously unfold the reasoning. It is fast because the work happened upstream, experience after experience.
Kahneman and Klein reconciled two opposing views here. Their analysis of intuitive expertise concludes that it becomes reliable when the environment contains regularities and provides clear feedback often enough to learn them. In an unpredictable environment, confidence can remain strong while accuracy disappears.
Robert Greene describes well, in his books on mastery, the point at which experience becomes immediate perception. It is a useful cultural reference for naming the phenomenon. It is not a scientific source: his historical stories illustrate an idea, they do not test it.
What this article does not say
- That associative memory is superior. It can help connect and transfer, but precise recall also requires attention and practice.
- That my word substitutions have one cause. Fatigue is compatible with them. It is not a diagnosis.
- That all autistic people remember as I do. I describe my own functioning, not a universal trait of the spectrum.
- That suffering causes learning. Adjusted difficulty can force useful retrieval. Overload that prevents understanding does not create a better connection.
- That the brain must be stimulated constantly. Rest and sleep are part of learning. Adding activity when already exhausted may simply degrade access to what has been acquired.
Understanding your brain does not mean asking more from it every day. It means noticing how information enters, which cue brings it back and when fatigue closes the paths. Optimisation will come in another article. Here the useful result is simpler: what you forget is not always lost, it may only be missing the right door.