§13 Drift-Diffusion Model — Evidence Accumulation in Decision-Making
dx = μ · dt + σ · dW, absorbing boundaries ±a
accuracy ≈ Φ(μ√(2a)/σ), RT ≈ a/μ (large-a limit)
accuracy ≈ Φ(μ√(2a)/σ), RT ≈ a/μ (large-a limit)
The DDM models two-alternative decisions as a noisy evidence accumulation process. The drift rate μ reflects information quality; the boundary a reflects caution. A wider boundary increases accuracy but also RT (speed–accuracy trade-off).
— ✦ —