Posner's (1980) spotlight model says neural processing is boosted wherever attention is aimed — the beam above sweeps to the currently-attended item, independent of where the eyes look. Treisman & Gelade's Feature Integration Theory (1980) explains why the time to find a target changes with how it differs from distractors:
RT = RT0 + k · n_scanned
k ≈ 0 feature search (pop-out, parallel)
k ≈ scan speed conjunction search (serial, self-terminating)
- Feature pop-out — the target is the only item with a unique color; every location is registered in parallel on a single "feature map", so it seems to pop out instantly and reaction time barely grows with item count.
- Conjunction search — the target's color and shape are each shared with different distractors, so no single feature map flags it. Focal attention has to visit items serially and "bind" color+shape together (Treisman's binding problem) until the target is found — reaction time grows roughly linearly with the number of items scanned.
- Item count — visual set size (n) in the search array.
- Scan speed — simulated time attention spends per item during a serial (conjunction) search.
- New trial — reshuffles distractor layout and target location; trials also restart automatically after each result.
Real-world relevance: this is the same mechanism behind inattentional blindness (the "invisible gorilla" experiment) and behind UI design — a single bright, uniquely-colored button "pops out" instantly, while a cluttered toolbar forces a slow serial scan.