🧪 Stroop Color-Naming Task with DDM
Run a genuine timed Stroop task -- name the ink color of color words while every response is measured with performance.now() -- then watch a stochastic drift-diffusion model explain the interference: noisy evidence accumulating toward a decision boundary at a slower drift rate on incongruent trials.
This simulator runs a real, interactive Stroop color-naming task: a color word appears in an ink color that either matches the word (congruent) or conflicts with it (incongruent), and you must report the ink color while every response is timed with the browser's high-resolution clock. Across enough trials this reproduces the well-documented Stroop interference effect -- incongruent trials measurably slower than congruent ones -- because reading a word is an automatic process that keeps injecting a competing response even when you're deliberately naming the color instead. To explain the mechanism behind that slowdown, the simulator also runs a genuine drift-diffusion model: a noisy evidence variable accumulates step by step via a stochastic random walk toward one of two decision boundaries, with congruent trials assigned a higher drift rate (clean, fast evidence) and incongruent trials a lower one (competing word-evidence drags the accumulation out). You can watch the model's live evidence-accumulation trace hit its boundary after every trial, tune the model's drift, noise, boundary and non-decision-time parameters, and compare your own measured reaction-time distribution directly against the model's simulated one for both conditions.
Click color-swatch buttons (R/G/B/Y) to name ink colors, see RT histograms and model simulations.
2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install