Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Memorability of narrative elements is associated with hippocampal and widespread cortical activation: an fMRI study using a controlled naturalistic story retelling task
Poster Session A, Wednesday, September 30, 11:00 am - 1:00 pm, Wangari Maathai
This poster is part of the Sandbox Series.
Manson Cheuk-Man Fong1, Xavreila Sheung-Wa Ng1, Enoch Yee-Lok Wan1, Zirui Zou1, Wenbo Wang1, Fangfei Li1, Matthew King-Hang Ma1; 1The Hong Kong Polytechnic University
Naturalistic paradigms offer an optimal scenario for uncovering the dynamic interplay between language and higher-order cognitive functions, including memory. Traditional functional neuroimaging demonstrates that the frontoparietal network maintains a strictly capacity-limited working memory buffer. In supraspan conditions, as the information load exceeds this capacity, the medial temporal lobe (MTL) long-term memory system, especially the hippocampus (HPC), is rapidly engaged to bind the fragmented sensory information together into a long-term memory trace for later recall. In a naturalistic setting, working memory is constantly being pushed to its limits, making it an optimal scenario for observing the dynamic encoding of linguistic information. Recent research has found that, alongside the hippocampus, the medial prefrontal cortex (mPFC) and the posterior medial cortex (PMC) are recruited during encoding and recall of spoken stories and movies. However, to what extent these regions are modulated by different types of linguistic information and memory recall performance remains largely unknown. In the present ongoing work, using a controlled naturalistic story-retelling task that involves immediately recalling spoken story segments of varying lengths, we seek to explore this general question by first zooming into how syntactic information and group-wise memorability modulate the functional activation of these regions (HPC, mPFC, and PMC). For comparison purpose, auditory and visual sensory areas were also included as the regions of interest. Thirty-two native, Hong Kong Cantonese speakers, aged 18–28, were recruited to listen, rehearse covertly, and retell overtly two stories during fMRI scanning. Each story, in Cantonese, was presented in a separate block of six trials. Each trial included a 0.8-second fixation, a story segment presented in the listening period (8/16/24 s), a constant 8-second covert rehearsal period, and an overt recall period (9.6/19.2/28.8 s). Functional activation was analysed using Nipype and FSL. For first-level modelling, building upon a baseline model comprising story and memory load for each of the three periods, additional syntactic and memory predictors were included to yield three different GLMs, including a word class model (6 classes including noun/verb/adjective/adverb/pronoun/preposition), node count model (both top-down and bottom-up node counts as parametric modulators), as well as a memorability model (with memorability operationalized as a time-dependent vector of group-level recall performance of 132 key-points per story). Cortical and hippocampal ROIs were defined according to the HCP-extended and Julich atlas, respectively. Our analyses revealed significant memorability-related modulation throughout mPFC (10r/v/d/p, dorsal/ventral 24d, anterior 24p, anterior 32p), PMC (7m, dorsal 23ab, 31pd, ventral 31p), language Area (44/45/47, PeriSylvian Language Area, STG), sensory cortices (e.g., visual area V1-V8), and all subdivisions of bilateral hippocampus (cornu ammonis, subiculum, dentate gyrus, entorhinal cortex, etc.), FDR-ps < .05. In addition, each of these broad regions included subdivisions sensitive to the three syntactic variables in word class, top-down node-count, bottom-up node-count, FDR-ps < .05. Taken together, beyond effects of memory load, our preliminary analyses demonstrate that (1) memorability of narrative elements modulates a distributed memory network; and, (2) this memory network is strongly sensitive to syntactic information such as word class and syntactic structures.
Topic Areas: Syntax and Combinatorial Semantics, Computational Approaches