19 April 2026 in SSRN Electronic Journal
Background and Objective: Semantic inhibition — the ability to suppress prepotentresponses in favour of goal-directed behaviour — is a core component of cognitivecontrol. The N450 event-related potential is a well-established marker of semanticconflict in tasks such as the Hayling Sentence Completion Test. However, traditionalanalyses reduce this component to a mean amplitude measure, providing limitedinsight into its underlying neural dynamics. The present study re-analyses an existingelectroencephalography dataset using a novel computational framework tocharacterize the dynamical and informational structure of the N450 response.Methods: Single-trial electroencephalography segments aligned to the N450 peakwere modeled using a Dual Damped Sine framework, which captures transientresponses as the superposition of two damped oscillatory modes. Information-theoretic metrics — Active Information Storage and Transfer Entropy — were thenapplied to the residual signal after removing the modeled oscillatory component, toquantify temporal persistence and directed information transfer across corticalregions of interest. Analyses were conducted on 17,496 trial-level fits from 47participants performing the Hayling task. Primary contrasts were corrected formultiple comparisons using the false discovery rate procedure.Results: The Dual Damped Sine model captured the substantial temporal structureof single-trial responses, with a median goodness-of-fit of approximately 0.56 forvalid fits. Residual analyses revealed a systematic redistribution of informationdynamics within the fronto-cingulate network during semantic inhibition. ActiveInformation Storage decreased in lateral frontal cortex and simultaneously increasedin the anterior cingulate cortex, while directed information transfer from frontal cortexto the anterior cingulate showed a modest but significant reduction. All three primarycontrasts survived false discovery rate correction.Conclusions: The N450 reflects a transient reconfiguration of fronto-cingulateinformation dynamics rather than a simple amplitude deflection. Semantic inhibitionis associated with a shift from distributed lateral frontal representations toward locallyrecurrent medial frontal dynamics. Integrating dynamical system modeling withinformation-theoretic analysis enables a richer characterization of cognitive control,capturing not only response magnitude but the evolving organization of informationprocessing in the brain. A fully reproducible analysis pipeline is publicly available.