Memory Games: The Brain Training That Actually Works (A Bit)
"Brain training" has been oversold so aggressively for the last fifteen years that most people have given up on the idea that any screen-based activity can actually improve cognitive function. That cynicism is understandable — Lumosity paid a $2M fine for false advertising — but it is also overcorrected. Memory games, played in short deliberate sessions, do train real skills. You just have to be honest about which skills and how much they transfer.
What memory games reliably improve
The research converges on a few findings. Memory games reliably improve:
- The specific skill you practise. Card-matching games make you better at card-matching games. Sequence recall games make you better at sequence recall. This is the single most well-replicated finding in cognitive training research.
- Working memory capacity within the domain you train. Visual memory training improves visual working memory; auditory training improves auditory working memory. Crossover is weak.
- Attention and focus during short sessions. Any task that requires sustained attention trains sustained attention, at least for tasks of similar duration.
What memory games do not do
Just as important:
- General "IQ" does not move. No game, app, or puzzle reliably improves measured IQ in adults. Claims to the contrary have been retracted or failed to replicate.
- Far transfer is weak. Playing memory games does not make you remember where you put your keys. It makes you better at memory games.
- Gains fade. Stop playing for two weeks and most of the gain is gone.
None of this makes memory games worthless — it just means you should train them for the specific reasons they work, not for the generic "I will get smarter" fantasy.
How to actually train memory
Here is what works, based on a careful reading of the cognitive training literature:
- Pick one game, not five. Variety feels productive but dilutes training effect. Pick a single memory game and stick with it for a month.
- Short sessions, every day. Ten minutes daily beats an hour on Sunday. The limiting factor is fatigue, not total minutes.
- Push the difficulty just past comfort. The sweet spot is roughly 70–80% success rate. If you are clearing every level, the game is too easy to train you. If you are failing half, it is too hard.
- Rest on weekends. Memory consolidation happens during sleep and during breaks. Taking a day off actually improves your performance on Monday.
- Don't grade yourself. A bad session does not mean you are losing ground. Brain performance fluctuates 15% day-to-day for reasons you cannot control — sleep, hydration, stress, time of day. Focus on the monthly trend, not individual sessions.
Which memory games to play
Different memory games train different things. A rough guide:
- Card matching (Memory / Concentration): visual memory and attention. Good starter for kids and adults.
- Simon Says / sequence recall: short-term sequential memory. Harder than card matching, trains a slightly different skill.
- Spot the difference: pattern matching and attention, not really memory in the strict sense but closely related.
- Number span / word span: classical working memory tests. These transfer the best to other working memory tasks — they are the gold standard if you specifically want to train working memory capacity.
The Memory category on LumiGameLab has one or two of each. Pick the one that looks most engaging and ignore the rest for your first month.
The honest summary
Memory games are worth playing if (a) you enjoy them, (b) you have a specific skill you want to practise, and (c) you accept that the gains are narrow and fade without practice. They are not a substitute for sleep, exercise, reading, or learning new things — all of which do more for general cognitive function than any game ever will. But within those limits, ten minutes a day is a harmless, pleasant, and marginally useful addition to an otherwise normal life.