Visual Memory

A grid appears and several tiles light up briefly. Once they fade, you click the squares that were lit. Clear a board and the next one is larger with more targets; miss too many and the run ends. Your score is the level you reach. Unlike a sequence test, order does not matter here — only which locations were marked.

This taps visuospatial short-term memory: a mental snapshot of a spatial layout that you hold for a few seconds. Because you recognize positions rather than recall a spoken list, this is a recognition task. Recognition is generally easier than free recall, which is one reason people often reach higher-looking levels here than on a strict recall test.

The test draws on classic ideas about the limits of visual working memory. Research on change detection and visual short-term memory suggests people can hold only a handful of visual items at once, with common estimates around three to four objects before accuracy drops. As the number of highlighted tiles rises past that, you increasingly rely on grouping and luck rather than storing each tile separately.

The task also relates to the visuospatial sketchpad, the part of working memory that handles images and locations, distinct from the verbal loop that handles words and numbers. That separation is why a strong number-memory score does not guarantee a strong visual-memory score. They lean on different systems.

Early levels feel easy because only a few tiles light up on a small grid. Difficulty climbs as both the grid size and the number of targets increase. Most people find the real challenge begins once they must remember more than four or five positions at once, which lines up with the roughly three-to-four item capacity seen in the lab.

• Comfortable: the opening levels with three to five targets.
• Challenging: six or more targets, where grouping becomes essential.
• Strong: pushing well past the point where you can no longer see each tile individually.

Visuospatial memory tends to peak in young adulthood and decline gently with age, and children's capacity grows through development. Use these as loose context, not a scoreboard.

You can lift your score with strategy, though the honest limit is that improvement is largely specific to this task. As with other working-memory games, better performance here does not reliably make your memory better elsewhere. What helps is smarter encoding, not a bigger raw store.

• Group by pattern: see the lit tiles as a shape — a corner, a diagonal, a cluster — instead of separate points.
• Use landmarks: anchor targets to edges, corners, or the center of the grid.
• Spread your attention: take in the whole board at once rather than scanning tile by tile before it fades.
• Stay calm on big boards: panic narrows attention and costs you positions.

Conditions shift your result, so treat any single run as approximate. How long the tiles stay lit, the size of the grid, and your focus all move the number. A brief flash makes the same board much harder than a longer one.

• Scanning instead of grasping: reading the grid one tile at a time means the later tiles fade before you encode them.
• Guessing on large boards: random clicks can end a run that patience might have saved.
• Practice effects: you improve over your first few tries as you learn to chunk, which is normal.
• Regression to the mean: one exceptional level may not repeat; judge yourself by your typical result.
• Fatigue and distraction: visual memory is fragile, and a noisy environment shows up directly in your score.