Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Social-Cognitive to Memory System Connectivity during Learning Supports Word Retention
Poster Session F, Friday, October 2, 2:45 - 4:45 pm, Wangari Maathai
Katherine Trice1, Zhenghan Qi1; 1Northeastern University
Behavioral work has shown that more active social-cognitive engagement during initial mapping results in better retention (Trice et al., 2024, 2026). While emerging behavioral and neural evidence in non-linguistic literature further substantiates potential privileging of social information in memory (Lin et al., 2019; Harvey et al., 2007; Jimenez & Meyer, 2024), no studies have directly examined whether functional connectivity between the social cognitive system and the memory system is associated with word learning and retention success. Twenty-three adults learned novel words during fMRI, mapping words where referent ambiguity was resolved using lexical cues (lexical inference: LI), where referent ambiguity was resolved using social cognition (pragmatic inference: PI), and where there was unambiguous mapping (direct mapping: DM). Retention was tested about 20 minutes later. Individualized fROIs for social cognition and memory were individually defined using the theory-of-mind localizer movie-watching task (Richardson et al., 2018) and a post-learning word retrieval task (remembered vs. forgotten). Directed functional connections from social-cognition to memory fROIs during learning were determined using dynamic graphical modeling (Schwab et al., 2018), getting the optimal combination of social-cognitive fROIs, if any, whose activation timeseries predicted a given memory fROI timeseries. Subjects remembered PI and LI words significantly better than DM words (PI: M=0.66, CI_95%=0.06; LI: M=0.71, CI_95%=0.05; DM: M=0.50, CI_95%=0.06; chance: 0.25; ps<0.05). There was no significant difference between PI and LI. Social-cognitive-to-memory functional connections during novel word learning were seen for a majority of subjects stemming from the VMPFC, MMPFC, PC, and LTPJ (Memory fROI Connections: N_VMPFC=10/11, N_MMPFC=2/11, N_PC=3/11; N_LTPJ=2/11; ps-adj<0.05). These connections were seen for only novel words in the MMPFC and PC and were seen for both novel and known words, in similar social situations, in the VMPFC and LTPJ. While these connections were not altered by learning context, a new connection from PC arose for solely PI when context was considered (p-adj<0.05). Connections in the two inferential-learning contexts were more predictive of individuals’ behavioral retention success than connections in the DM condition (Significant Correlations: N_PI=8, N_LI=10, N_DM=2; p-adj<0.05). In a majority of these significant LI or PI cases, this prediction was positive – a connection predicting better rather than worse retention – while with DM the trend was evenly split (PI: 88% positive; LI: 70% positive; DM: 50% positive). Findings were further supported by converging evidence from significant correlations in BOLD activation in univariate analyses between social-cognitive regions reliably engaged for novel over known words during learning and bilateral hippocampi (PI: r_DMPFC=0.35, p_DMPFC=0.02, r_LTPJ=0.40, p_LTPJ=0.008; LI: r_LTPJ=0.42, p_LTPJ=0.004; DM: p_LTPJ>0.05). To conclude, functional connections from the social-cognitive to the memory system exist during word learning, especially in situations that require inference, which likely support longer-term word retention. This is substantiated by two separate methods of neural analyses showing alignment and direct relationships with behavioral data. ROI Abbreviations: VMPFC=Ventromedial Prefrontal Cortex, MMPFC=Medial Prefrontal Cortex, PC=Precuneus, LTPJ=Left Temporoparietal Junction, DMPFC=Dorsomedial Prefrontal Cortex
Topic Areas: Language Development/Acquisition, Meaning: Lexical Semantics