Spoken language comprehension relies on context-dependent neural weighting of syntax and prosody

Deciphering spoken language hinges on neural weighting of syntax and prosody that is influenced by context. Psycholinguistic research suggests prosody plays a crucial role in this comprehension process. Nevertheless, the general nature of interactions between prosody and syntax in spoken language processing remains uncertain due to controlled sentence constructions typically used in studies.

To tackle this issue, an amalgamation of experimental psycholinguistics and naturalistic neuroscience was employed. In this study, 26 participants (19 of whom were female) listened to narrative speech recorded via MEG (magnetoencephalography) recordings. Linguistic annotations indicated that strong prosodic boundaries often align with syntactic phrase closures, with the intensity of prosodic boundaries scaling according to the number of phrases closed.

Through time-resolved regression with temporal response functions, the relationship between these features and delta-band neural activity during continuous speech processing was modeled. The findings reveal a reciprocal interaction between syntax and prosody. Prosodic marking impacts neural processing of syntax, with phrase closures being encoded more robustly when accompanied by strong prosodic boundaries.

Conversely, syntactic structure influences prosodic encoding, with boundary strength primarily encoded neurally at points of syntactic phrase closure. These results underscore the mutual constraining and context-dependent encoding of syntax and prosody, underscoring their joint role in guiding structure building during naturalistic spoken language comprehension.