Neurocognitive Signatures of Lexical-Class Perception in Noise Among L2 Listeners

This study investigated the neurocognitive foundations of lexical category resilience in second language (L2) speech perception under adverse listening conditions. Forty Saudi Arabic learners of English at beginner and advanced proficiency levels completed word-recognition and sentence-recognition tasks in speech-shaped noise while behavioral accuracy and EEG measures were recorded. Mixed-effects models revealed a consistent hierarchy of perceptual resilience adjectives ranked highest, followed by nouns, adverbs, and verbs across proficiency levels, with stops and affricates providing stronger perceptual anchors than approximants and trills. Neurophysiologically, content words elicited larger PMN and P300 amplitudes, indicating more robust phonological mapping and attentional resource allocation. Advanced learners showed stronger N100 and PMN responses, reflecting earlier and more automatic processing, whereas beginner learners relied on later attention-mediated components (P300, N400). Time–frequency analyses identified theta and gamma synchronization as neural markers of successful perception, especially for high-resilience categories. Working memory and phonological awareness correlated strongly with both behavioral and neural measures, particularly for low-resilience categories and less proficient learners. These findings integrate phonetic, linguistic, and cognitive factors to provide a comprehensive account of how lexical category, phonetic salience, and proficiency interact in L2 speech perception under noise, offering implications for L2 pedagogy, automatic speech recognition systems, and neurocognitively informed language training. Nonetheless, the focus on Saudi Arabic learners and the lack of control for certain psycholinguistic variables (e.g., imageability, concreteness) limit generalizability, highlighting the need for studies with diverse populations and more comprehensive stimuli measures.