Two-stream updating
Compare each current grid position and letter with the item shown N trials earlier. Mark position and letter matches independently before the next cue.
Compare the current position and letter with the item n steps earlier.
Progress
1/20
Accuracy
0%
Hits
0
False positives
0
Misses
0
Best
0%
Two streams run together: one spatial and one verbal. A trial can match neither, one, or both, so each match button represents a separate decision.
The page offers 1-, 2-, and 3-back settings. Its result is specific to the generated stream and does not establish broad working-memory transfer.
In 2-back, for example, the current position is compared with the position two cues earlier, and the current letter is compared with the letter two cues earlier. Consecutive repetition is irrelevant unless N is one.
Early trials without N previous items cannot contain a valid match. Marking a match where none exists counts as a false alarm; failing to mark a generated match counts as a miss.
Accuracy alone can hide response strategy. Review hits, misses, and false alarms together to distinguish missed targets from pressing match too often.
Improvement can reflect learning this interface, cue timing, and match frequency. This task does not diagnose impairment or guarantee gains in intelligence, study, or work performance.
Recent N level and response counts are kept in local browser storage, without an account or cross-device sync.
These questions stay tied to the visible content on this page.
Continue with tools that pair naturally with this page.
Watch an ordered sequence on a 3×3 grid, reproduce it, and review level, attempts, accuracy, and local history.
Start
Repeat a four-pad color sequence that grows each round and review best length, mistakes, and average input delay.
Start
Answer color-word, reverse, or shape rules and review accuracy, median timing by condition, and the page-defined interference difference.
Start