Proceedings of the Annual Meeting of the Cognitive Science Society

Prediction is central to language comprehension, yet how neural markers of prediction translate into decision behavior remains unclear. We combined EEG with drift-diffusion modeling (DDM) to dissociate prediction-specific mechanisms from semantic priming facilitation effects. Participants (N=30) completed a lexical decision task following either one or three semantically related primes, manipulating prediction strength while holding semantic relatedness constant. The three-prime context elicited early prediction-sensitive activity (eN400, 200–300 ms) alongside classic N400 effects. Critically, DDM revealed a double dissociation: classic N400 predicted drift rate, while eN400 predicted boundary separation, reflecting different computational pathways. These findings demonstrate that even minimal predictive context (three primes) engages early neural processing. Moreover, the observed neurocomputational dissociation between neural indices and decision parameters supports hierarchical predictive accounts of language comprehension by specifying not only whether predictive context matters, but when it influences semantic processing and which task-relevant decision mechanisms it modulates.