Cortical speech-in-noise processing reflects subcortical syllable encoding and lexical attention

Speech-in-noise performance varies among individuals with similar hearing sensitivity. Such variability may manifest as varying cortical representations of speech in noise, which may stem from peripheral sensory fidelity and the efficacy of top-down attention. The present study examined how subcortical speech-syllable encoding and lexical attention shape cortical speech-in-noise representation, as measured by neural signal-to-noise ratio (SNR) based on the amplitude ratio of auditory-evoked responses to the target speech onset and the noise onset. The result indicated that neural SNR significantly correlated with behavioral speech-in-noise performance in the moderate-difficulty acoustic condition. The regression model showed that subcortical syllable encoding fidelity and lexical attentional performance independently predicted the neural SNR, suggesting converging contributions from both bottom-up sensory encoding and top-down attention. In addition, syllable encoding, lexical attention, and neural SNR collectively accounted for individual variability in behavioral performance, indicating a shared, partially overlapped contribution from sensory, attentional, and cortical measures to behavioral speech-in-noise perception. Our findings highlight that individual neural SNR serves as a behaviorally relevant cortical speech-in-noise representation reflecting both sensory and attentional mechanisms.