Cortical encoding of acoustic and linguistic rhythms reflects L2 narrative comprehension

Speech comprehension is a multistage process involving both acoustic encoding and linguistic processing. Accumulating evidence has demonstrated that low-frequency cortical activity can track perceived linguistic units (e.g., words) on top of basic acoustic features (e.g., speech envelope). However, it remains unclear how the neural tracking of acoustic and linguistic information relates to second language (L2) speech comprehension in narrative contexts. Here, we investigate neural tracking of narrative speech for L2 listeners using electroencephalography (EEG). Notably, we introduce amplitude modulation (AM) cues aligned with word rhythm onto the basic envelope of speech and employ a frequency-tagging paradigm to measure neural responses to word and AM rhythm separately. When narrative speech was presented to L2 listeners during a speech comprehension task, reliable neural tracking of word and AM rhythm was observed in low-frequency cortical activity. While the introduction of AM cues enhances both comprehension performance and word-tracking responses, listeners with high versus low comprehension performance exhibit differences in their word-tracking responses rather than AM-tracking responses. Furthermore, the power and phase associated with word-tracking responses jointly reflect individual comprehension performance of L2 listeners. Our results indicate that bottom-up acoustic cues and top-down linguistic knowledge predominantly modulate the low-frequency neural tracking of linguistic units, which contributes to speech comprehension in a nonnative language.