Ten minutes, three exercises
Today's session
Three short exercises drawn from the kinds of cognitive training that have been tested in clinical trials. Each one adjusts to you, so it stays just hard enough to be worth doing.
Your three for today
Where you are
Progress
Levels rise when you get things right and ease off when you don't — the aim is to keep you at roughly four right out of five, which is where training is thought to do the most work.
Difficulty reached, by ability
Recent exercises
| Date | Exercise | Result | Accuracy | Level |
|---|
Nothing here yet. Finish a session and it will show up.
Straight answers
What this is, and isn't
Northlight will make you better at Northlight. Whether that carries over into everyday life is a real scientific question with a mixed answer, and we would rather tell you that than sell you a promise. Nothing here diagnoses, treats or prevents any disease.
What the research actually supports
The strongest single result in this field comes from the ACTIVE trial, which followed around 2,800 older adults for two decades. The group trained on speed of processing — spotting a central object and a peripheral one in a shrinking flash of time — and who returned for booster sessions showed a meaningfully lower rate of dementia diagnosis over 20 years. Memory and reasoning training in the same trial showed no significant benefit on that measure.
That is why the speed exercise here — Roadside — is the one we put first in most sessions, and why it is built the way it is: a central identification and a peripheral location, both getting faster, wider and noisier as you improve.
What is genuinely uncertain
- Transfer. Large studies have found people get better at the trained task without improving on untrained measures. The National Academies called the transfer evidence limited and inconsistent.
- Memory games specifically. Recall and matching exercises are the most intuitive things to build and the least well supported. They are here because they are pleasant and measurable, not because they are proven.
- Effect sizes. Where benefits show up they are usually modest, and control-group choices have been argued over for years.
The bigger lever
In 2025 the U.S. POINTER trial reported that a structured programme — regular aerobic and resistance exercise, the MIND diet, health monitoring, social contact and computerised cognitive training together — improved global cognition in at-risk older adults more than a self-guided version did. Cognitive training was one ingredient among several, not the whole recipe. Ten minutes here is worth more when it sits alongside the walk, the sleep and the phone call.
Why we are careful about wording
In 2016 the makers of Lumosity paid $2 million to settle Federal Trade Commission charges over claims that its games could delay memory decline and protect against dementia. The regulator's point was simple: improving inside the game is not evidence of improvement outside it. Every number in this app describes your performance on these exercises, and says nothing else.
If you are worried about your memory, that is a conversation for a doctor, not an app.