Contributed Talk Sessions | Poster Sessions | All Posters | Search Papers

Poster Session A: Tuesday, August 12, 1:30 – 4:30 pm, de Brug & E‑Hall

Working Memory Reactivation Across Embedded Structures during Natural Language Comprehension

Jiaqi Li¹, Yali Pan, Hyojin Park, Huan Luo, Ole Jensen; ¹University of Oxford

Presenter: Jiaqi Li

A key feature of human language is recursion, involving hierarchical embedding of clauses– a process dependent on working memory (WM). During speech comprehension, listeners must maintain earlier linguistic elements for later integration. Even processing adjacent words into phrases requires WM resources (Desbordes et al., 2024), whereas the integration across embedded structures depends more on the flexibility of the WM system. However, the neural mechanisms underlying how WM supports the processing of complex recursive linguistic structures remain unclear. We constructed English sentences with embedded language structures and recorded magnetoencephalography (MEG) signals while 34 native speakers listened to these sentences. Neural decoding results demonstrate that during embedded structure processing, previously encoded information is stored in an activity-silent state until the non-adjacent verb of the main clause triggers reactivation. Source-level analysis reveals that the reactivation first occurs in prefrontal regions followed by activation in the temporal cortex. This study provides crucial insights into the temporal and spatial dynamics of WM functions required for unification operations across embedded structures.

Topic Area: Language & Communication

Extended Abstract: Full Text PDF