Proficiency-Dependent and Dynamic Reorganization of Brain Networks During Children's Second Language Learning

Childrens second language (L2) vocabulary learning is often indexed with event-related potentials (ERPs) by the N400 incongruity effect, yet its stability in real classrooms is unclear. We recorded 64-channel EEG from 42 children (aged 9-15 years) during a translation-recognition task before and after a regular instructional vocabulary unit and sub-grouped children by baseline proficiency. We quantified the canonical N400 (300-500 ms, centroparietal region of interest; ROI) alongside electrical-neuroimaging indices of ERP strength and topography (Global Field Power, Global Map Dissimilarity, ERP template-map preponderance), using adult template maps from a prior study as a normative reference. Behaviorally, children showed large gains in vocabulary accuracy and higher discrimination sensitivity (d') in post-assessment. The classical N400 incongruity effect did not increase from pre- to post-learning. By contrast, global and topographic measures were learning-sensitive: Global Field Power increased across sessions and interacted with proficiency, and Global Map Dissimilarity showed main effects of session and proficiency as well as their interaction, indicating proficiency-dependent topographic reorganization. Adult normative template maps backfitted well; within the N400 window, high-proficiency children more consistently occupied an adult-like configuration; later ([~]650-800 ms), patterns also differentiated by proficiency and congruency. Relative to adults, children exhibited more diffuse topographies and stronger dependence on individual proficiency. In classroom settings, childrens L2 vocabulary learning is therefore more sensitively indexed by scalp topography and global field dynamics than by a single N400 amplitude, positioning electrical neuroimaging as a proficiency-dependent, classroom-relevant marker of L2 acquisition.