Memory Grid
Remember and reproduce a growing grid of highlighted tiles. Trains visuospatial working memory. Watch the tiles light up, then click them back in the same order. Each round adds one more tile to the sequence — see how far you can get.
What this game trains
Memory Grid is a digital adaptation of the Corsi block-tapping task, a clinical and research instrument first described by psychologist Philip Corsi in 1972 and now a standard part of many neuropsychological test batteries. In the original task, an examiner taps a sequence of wooden blocks arranged irregularly on a board, and the participant taps them back in the same order. The sequence lengthens by one block each time the participant succeeds, until they fail twice at the same length. The longest sequence they successfully reproduce is their "Corsi span."
Our version uses a 4×4 grid of 16 tiles rather than the original irregular layout, but the underlying measure is the same: visuospatial working memory span — how many distinct spatial locations you can hold in mind and reproduce in order. This is a different faculty from verbal working memory (which the digit-span task measures) and from long-term memory. Practicing the task reliably improves your score on it; whether that improvement transfers to other cognitive tasks is, per the broader brain-training literature, contested.
How the sequence is generated
Each round, the game selects N distinct tiles at random from the 16,
where N starts at 3 and increases by 1 each time you correctly reproduce
the sequence. The same tile is never repeated within a single sequence. The tiles
light up in order, one per 700 milliseconds, with a 300-millisecond gap between them.
Worked example
Round 1 (sequence length 3): tiles (row 1, col 2),
(row 3, col 4), (row 2, col 1) light up in that order.
You must click them back in the same order. If you succeed, the sequence length
becomes 4. Round 2 (length 4): four new tiles light up, and you reproduce them.
This continues until you click a wrong tile or click out of order, at which point
the game ends and your final sequence length is your score.
What a typical score looks like
Per published norms for the Corsi block task (Kessels et al., 2000, and later replications), the average visuospatial span for healthy adults is around 5, with most people falling between 4 and 7. A score of 8 or higher is above average; reaching 9 or above is uncommon without specific practice. Children's spans are shorter and increase with age through adolescence.
If you practice this specific game daily for a week or two, your score will likely rise by 1-2 tiles. That improvement reflects practice on the task itself — the research literature does not support strong claims that this practice generalizes to broad cognitive gains. We tell you this honestly because overstating the benefit would be exactly the kind of claim that gives the brain-training niche a bad reputation.
Tips for getting further
- Verbalize the positions. Many people find it useful to silently name the tiles by their position ("top-left, middle-right, bottom-left"). Encoding spatial information verbally is a legitimate strategy that increases effective span.
- Chunk the sequence. If the sequence is 7 tiles long, try to remember it as two chunks of 3 and 4 rather than seven individual positions.
- Don't rush. There is no time limit on your response. Take a beat after the sequence finishes to mentally rehearse before clicking.
- Reduce distractions. Working memory is fragile; a single glance at a notification can wipe what felt like a stable sequence.
Edge cases and known limits
The game requires JavaScript. On very small screens (under 375px wide) the tiles may shrink and become harder to tap accurately — use a tablet or desktop for serious attempts. The game does not store your score on the server or across sessions; closing the tab resets everything. The "Best" counter is per-session only.