Search Abstracts | Symposia | Slide Sessions | Poster Sessions
The interplay between information status and prosody in the brain
Poster Session F, Friday, October 2, 2:45 - 4:45 pm, Wangari Maathai
Hatice Zora1,2, Laura Giglio3, Müge Özker Sertel1,2, Peter Hagoort1,2; 1Max Planck Institute for Psycholinguistics, 2Donders Institute for Brain, Cognition and Behaviour, Centre for Cognitive Neuroimaging, Radboud University, 3University of Zurich
Information Structure studies how speakers package information relative to their assumptions about the listener’s mental state and discourse context. Among its central dimensions is Information Status (IS), namely the distinction between ‘given’ information (shared or inferable in discourse) and ‘new’ information (previously unmentioned content). IS reflects how speakers mentally represent ongoing interaction and plays a key role in communication efficiency, yet its neural basis remains poorly understood. This study aims to contribute to a better understanding of how IS is represented and processed in the brain and how it interfaces with prosody, especially prominence. IS is typically reflected in prominence, which highlights elements by making them stand out from their surroundings. Prominence can be described both as a phonological representation (PP) and as an acoustic realisation (AP). PP involves categorical labels (pitch accents) that encode linguistic structure relying on phonological and morphosyntactic rules independently of their acoustic realisation. AP is conveyed through cues such as fundamental frequency (f0), duration and intensity. New information is assumed to be pitch accented and signalled with greater AP, while given elements are unaccented and expressed with reduced AP. Using fMRI data, this study examines neural mechanisms that mediate IS processing as well as its interplay with PP and AP. Native Dutch participants (n = 205) passively listened to spoken narratives. Narratives were first annotated for discourse referents, such as noun phrases, prepositional phrases, and their pronominal forms. Referents were then annotated for IS (given vs. new) using a language-independent annotation scheme. Subsequently, PP was coded (present vs. absent) for all referents, and AP was extracted with the Wavelet Prosody Toolkit by combining f0, energy, and duration into one measure. Binary IS, binary PP and continuous AP measures were entered as parametric modulators of each word in a GLM analysis. The results indicated that PP activated classical language areas such as superior temporal gyrus (STG), inferior frontal gyrus (IFG) and supplementary motor area (SMA) bilaterally, effectively modulating the response to all words. Words annotated for IS, in contrast, revealed distinct activation patterns such that new information elicited activation in the middle frontal gyrus (MFG) (specifically dorsolateral prefrontal cortex, DLPFC), inferior temporal gyrus (ITG), and supramarginal gyrus (SMG), whereas given information showed activation patterns similar to unannotated words. Notably, AP showed activation patterns similar to those for new information, but with greater angular gyrus (AG) activation. Overall, these findings suggest that IS does not appear to interface with PP, but aligns with AP. Because new information is cognitively demanding, as it requires discourse updating, AP likely facilitates its processing. The activation of regions associated with new information processing (DLPFC) and AP processing (AG) in pragmatic language functions including discourse management and attentional language functions, respectively, as reported in the literature, supports the consistency of these findings with the literature. We are currently investigating how these BOLD activations relate to different cortical networks such as Default Mode, Multiple Demand, and Attention Networks. To our knowledge, a cortical network model of IS processing has yet to be defined.
Topic Areas: Meaning: Discourse and Pragmatics, Prosody