Working Memory: Your Brain's Sticky Note and How to Train It
Hold a phone number in your head while you find a pen. Keep the first half of a long sentence in mind until the speaker reaches the verb. Carry the 7 while adding a column of figures. All of these use working memory — the small, fast, constantly-overwritten workspace where your brain holds information it is using right now.
Working memory is not a filing cabinet; it is a sticky note. It holds very little, it fades within seconds unless refreshed, and anything new written on it tends to shove off whatever was there before. Yet nearly every demanding mental act — reasoning, mental arithmetic, following instructions, reading comprehension — runs through it. Its capacity is one of the strongest single predictors of performance on complex cognitive tasks, and of fluid intelligence itself.
How small is the sticky note?
George Miller's famous 1956 paper put the limit at "seven, plus or minus two" items. Modern research, led by Nelson Cowan, revised that downward: for genuinely new, un-groupable material, most adults reliably hold about four chunks. What looks like a bigger span is almost always chunking — recoding several items into one meaningful unit. "FBI–CIA–NASA" is nine letters but only three chunks; a chess master recalls a whole board because the position resolves into a handful of familiar patterns.
Capacity also varies between people and across a lifetime. It rises through childhood, peaks in the twenties, and declines slowly thereafter — and at every age, stress, sleep loss and divided attention shrink the usable space further.
Working memory vs short-term memory
The terms are often mixed up. Short-term memory is passive storage — parroting back a list you just heard. Working memory is storage plus manipulation — repeating that list backwards, or holding numbers while also adding them. The manipulation component is what makes working memory so predictive of reasoning ability: it measures not just how much you can hold, but how much you can hold while doing something else.
What the training research honestly says
Working-memory training was the centre of a heated scientific debate for two decades, and the dust has largely settled on three conclusions:
1. Trained tasks improve substantially
Practise a span task and your span on that task grows — often impressively. You develop better chunking strategies, better rehearsal rhythms, and task-specific skill.
2. Near transfer is real but modest
Gains transfer partially to similar untrained tasks — a visual span game helps other visual span tasks. Meta-analyses consistently find this near transfer.
3. Far transfer is weak
Claims that working-memory training raises IQ or transforms school performance have not survived replication. The honest position: training sharpens the skill you practise and its close neighbours, not general intelligence.
That is not a reason to skip it. Better strategies (especially chunking), faster encoding, and higher comfort under memory load are genuinely useful in daily life — and unlike far-transfer promises, they are reliably achievable.
Working with the limits you have
- Chunk deliberately. Group digits in threes, letters into words, steps into stages. Capacity is fixed in chunks — so make bigger chunks.
- Offload ruthlessly. Write things down. Working memory is for processing, not storage; every item you hold "just in case" taxes the workspace you need for thinking.
- Serialise tasks. Each switch between tasks flushes and reloads the sticky note. One thing at a time is not a productivity slogan; it is how the hardware works.
- Protect the state. Sleep deprivation and anxiety measurably shrink working memory. The cheapest capacity upgrade available is a full night's sleep.
- Practise under load. Games that force holding-plus-doing — sequence recall with time pressure, updating tasks — exercise the control processes that keep items refreshed against interference.