How the cognitive games are scored

PRISM uses five well-validated cognitive paradigms, each measuring a distinct facet of how you think. None of them is timed in a stressful way — the windows are tight enough to discourage deliberation, but they're not punishingly fast.

Flanker (processing speed)

Five arrows in a row; respond to the middle one's direction and ignore the four around it. We score the median reaction time on congruent trials, normalized against a population reference. Faster is better but only up to a point — accuracy still has to be high.

Go/No-Go (attention + impulse control)

Press the button when you see M, do nothing when you see W. Most stimuli are M, so the natural rhythm pulls you into pressing — the score weights "false alarms" (pressing for W) more than misses, because catching impulses is the point.

N-back (working memory)

A stream of letters; press when the current letter matches the one shown two letters ago. We compute d' (d-prime) from hits and false alarms — the discriminability of matches from non-matches.

BART (risk preference)

Pump a balloon to earn money; bank the balance before it pops. The score is the average number of pumps on balloons you successfully banked (popped balloons don't count) — a behavioral proxy for risk tolerance.

WCST (cognitive flexibility)

Sort cards by an unstated rule; after a run of correct answers, the rule silently shifts. We score categories completed and perseverative errors — sticking with the old rule after the shift.

Each game produces a normalized 0–1 score and a confidence value (how complete the data was). The scores feed into the cognitive and behavioral layers of your 29-dimension profile.