Find the odd-one-out among distractors. Scan speed and attentional breadth, trained directly.
Visual search is how quickly you can scan a field for a specific target.
Parallel search (the target "pops out") is near-instant. Serial search (scanning item by item) is slower and scales with density.
This trainer varies density and shape similarity to push you between parallel and serial regimes.
Consistent scan patterns (top-to-bottom, left-to-right) beat random glances.
Treisman's feature integration theory distinguishes "preattentive" pop-out search from controlled conjunction search.
Training narrows the density threshold at which search becomes serial, effectively increasing effective attentional breadth.
It gauges how efficiently your attention scans cluttered space to isolate one target.
This task isolates selective spatial attention under load: the ability to deploy a focused "spotlight" across a display and bind features into objects. A conjunction trial (target sharing colour with some distractors and shape with others) forces item-by-item checking, so your result reflects the per-item cost of inspection rather than raw reaction speed alone.
Unlike sustained-attention tasks that track vigilance over minutes, visual search probes attention across space at a single moment. Its everyday correlates are concrete: spotting a friend in a crowd, finding the exit sign in an airport, or a radiologist scanning a scan for one anomaly among normal tissue.
Build up set size and similarity gradually so search stays efficient, not frantic.
Begin with small displays and a target that pops out by one feature, then increase set size and make distractors share more features with the target, pushing you from parallel to serial search. Level up only when accuracy holds above roughly nine in ten. Focus on a steady, systematic sweep rather than random jumps around the display.
The common error is sacrificing accuracy for speed, clicking before the target is confirmed and missing it entirely, the classic miss on rare-target trials. Resist re-checking items you have already cleared; trust your sweep. If you find yourself staring at the centre, deliberately move fixation outward to the periphery.
Read the slope of time against set size, not any single fast trial.
The informative metric is how much your response time grows as items are added. A near-flat curve signals efficient, pop-out search; a steep rise means you are inspecting items one at a time. Strong performance keeps the curve shallow and errors low even on dense conjunction displays. Track your own slope over sessions rather than chasing one lucky time.
A score is state-dependent: fatigue, poor lighting, screen size and caffeine all shift it, so compare like with like. It does not measure intelligence, peripheral vision health, or how you search real three-dimensional scenes, where depth, motion and memory for already-checked locations change the problem substantially.