Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Poster Slams
Semantic processing requires communication between semantic and domain general systems
Poster A8 in Poster Session A, Thursday, October 6, 10:15 am - 12:00 pm EDT, Millennium Hall
Curtiss Chapman1, Gesa Hartwigsen1; 1Max Planck Institute for Human Cognitive and Brain Sciences
How do brain networks coordinate to process meaning? Neuroimaging studies have provided evidence for a distributed contribution of semantic areas, but the interaction of specialized areas for semantic processing with the rest of the brain is less clear. Some research has begun to illuminate these interactions1, but limitations in scope have caused evidence to be accumulated for interactions of only a few semantic regions. Other studies suggest that domain general networks such as the multiple demand network (MDN) are recruited even for basic semantic processing2, but it remains unclear which networks are important and under which conditions they are recruited. The current study explored how semantic networks interact with domain general networks during semantic processing. Interactions between semantic and domain general networks were investigated with correlational psychophysiological interaction analyses (cPPI). Forty-one participants performed two semantic tasks and a baseline task during fMRI. In the semantic tasks, participants decided whether two simultaneously presented pictures were related to each other; in the baseline task, participants performed visual judgments on scrambled images. Ttask-related data were submitted to an independent components analysis (ICA). Non-noise components were compared with network templates for the semantic system,2 MDN,3 and the 17-network map of Yeo et al.4 using the Jaccard similarity index. Eight components with highest similarity to semantic and domain-general networks were submitted to a cPPI, which derived a partial correlation between components for each pair of regions and each task condition. We then used linear mixed models to determine which components showed different functional coupling during semantic tasks compared to the baseline task, using Bonferroni correction to assess significance across the 28 comparisons. Results showed that our primary semantic component increased functional coupling with MDN and dorsal attention components during semantic processing compared to baseline. Furthermore, MDN and dorsal attention components increased functional coupling with components overlapping the default mode network (DMN). Interestingly, semantic-DMN component pairs and DMN-DMN component pairs showed stronger functional coupling during baseline than semantic tasks. Our results provide evidence that communication between semantic and domain general networks is necessary for semantic processing. Such coupling may be necessary for retrieval and comparison of semantic representations. Domain-general networks showed similar patterns of coupling with semantic and DMN components, consistent with theories suggesting that the DMN has an explicit role in declarative—perhaps specifically semantic—memory.5-6 The observed stronger coupling between semantic and DMN components during non-semantic processing is surprising. This result may suggest that the particular networks have partially distinct functions—and thus less functional coupling—during controlled semantic processing, whereas during non-semantic processing, the regions have more similar functions and are thus more strongly functionally coupled. References: 1 Palacio N. & Cardenas F. (2019). Rev Neurosci 30(8): 889-902. 2 Jackson R.L. (2021). NeuroImage 224: 117444. 3 Fedorenko E. et al. (2013). Proc Natl Acad Sci USA 110(41): 16616-21. 4 Yeo B.T. et al. (2011). J Neurophysiol 106: 1125-1165. 5 Stark C.E. & Squire L.R. (2001). Proc Natl Acad Sci USA 98(22): 12760-6. 6 Spreng R.N. et al. (2009). J Cogn Neurosci 21(3): 489-510.
Topic Areas: Meaning: Lexical Semantics, Control, Selection, and Executive Processes