Keynotes | K&Ts | GACs | Talks | Posters | Search

Poster C in Poster Session C: Wednesday, August 5, 9:30 – 11:15 am, Kimmel Center, Shorin & Rosenthal Rooms

A multidimensional framework for episodic memory reveals age-related changes in underlying cognitive processes

Soroush Mirjalili¹, Sahana Ram², Sydney Lamphier², Ivy Liang³, Priscilla Pham², Audrey Duarte²; ¹University of Oregon, ²University of Texas at Austin, ³University of Texas Health Center at Houson

Presenter: Soroush Mirjalili

Episodic memory is inherently multidimensional, supported by the coordinated engagement of multiple cognitive processes. However, most prior work has examined these processes in isolation, leaving unclear how they co-occur, fluctuate over time, and interact to shape behavior. This limitation is particularly relevant for understanding cognitive aging, where memory impairment may arise from changes in the engagement, variability, and coordination of underlying processes. Recent machine learning approaches have identified neural “brain states” during encoding that predict subsequent memory performance, but these methods typically treat episodic memory as a unidimensional construct. Here, we propose a process-based framework in which trial-by-trial fluctuations in visual perception, sustained attention, and selective attention are quantified to better predict memory success and explain its underlying mechanisms. We recorded EEG while 43 younger and 32 older adults performed perception, attention, and episodic memory tasks. Classifiers trained to distinguish high versus low performance states in the perception and attention tasks were transferred to decode process engagement during memory encoding at millisecond resolution. This multidimensional approach significantly improved prediction of subsequent memory compared to traditional models. Each process explained unique variance in successful encoding, demonstrating independent contributions. Moreover, successful encoding was characterized by greater temporal variability in process engagement and stronger synchrony between processes, highlighting the importance of both flexibility and coordination. Critically, older adults exhibited reduced within-trial variability, indicating less flexible engagement of cognitive processes, while maintaining intact synchrony. Across participants, greater variability and synchrony predicted better memory performance. These findings suggest that age-related memory impairment may stem from reduced variability in the engagement of perceptual and attentional processes, consistent with neural flexibility accounts of aging. Together, these findings establish a general process decoding framework for tracking dynamic cognitive interactions, offering broad insights into complex cognition, including, but not limited to, the mechanisms underlying age-related memory impairment.

Topic Area: Memory, Learning & Knowledge Structures