Name: Liam Hart

Master’s Thesis Proposal Meeting

Data: Wednesday, December 10th, 2025

Time: 12:00PM – 2:00PM

Location: Virtual, Meeting link click here

Thesis Chair/Advisor:
Hsiao-Wen Liao, Ph.D. (Georgia Tech)

Thesis Committee Members:

Sashank Varma, Ph.D. (Georgia Tech)

Audrey Leroux, Ph.D. (Georgia Tech)

Title: Gist-Based Compensation for Age-Related Episodic Decline in Emotional Autobiographical Memory: A Natural Language Analysis of Life’s Highs and Lows 

Abstract:

 Age-related reductions in the episodic specificity of autobiographical memory are well established, whereas gist-based representations and semantic processing remain comparatively preserved in old age. Emerging evidence suggests that older adults may leverage intact semantic memory to compensate for diminished episodic retrieval, drawing on a broader range of semantic knowledge to facilitate detailed recall. Yet it remains unclear whether this compensatory mechanism extends to emotional memories. Socioemotional aging research consistently demonstrates an age-related positivity effect—older adults show a relative preference for positive over negative information, reflecting motivational priorities to maintain a positive emotional state. Theoretical accounts suggest that positive memories often rely more on relational, gist-based processing. Since detailed recall of positive memories align with these emotional goals, this compensatory semantic scaffolding may be especially prominent during the recall of positive events compared to negative ones. The proposed secondary data analysis study evaluates this possibility by analyzing high and low points in the life stories of older adults and testing whether the association between breadth of semantic content and episodic detail is moderated by valence. Topic modeling will be used to quantify semantic breadth, and a validated language model fine-tuned on Autobiographical Interview–scored narratives will provide automated estimates of episodic details. Multilevel models will assess the hypothesized compensatory relationship between semantic breadth and episodic detail, and whether this effect is moderated by memory valence. By integrating cognitive aging and socioemotional selectivity perspectives, this study aims to clarify how preserved semantic memory and emotional goals jointly shape detailed autobiographical remembering in old age.