Schedule of Events | Search Abstracts | Invited Symposia | Symposia | Poster Sessions | Data Blitz

Postdoctorial Fellowship Award Winner Sketchpad Series 

Neural mechanisms of insight: nonlinear cortical representational change with hippocampal and catecholamine engagement

Poster Session A - Saturday, March 7, 2026, 3:00 – 5:00 pm PST, Fairview/Kitsilano Ballroom
Also presenting in Data Blitz Session 1 - Saturday, March 7, 2026, 10:30 am – 12:00 pm PST, Salon ABC.

Dr. Maxi Becker¹ (), Dr. Simon Davis, Dr. Roberto Cabeza; ¹Duke University

“Aha!” moments—sudden insights marked by a lasting memory trace—are central to creativity and learning, yet their mechanisms remain elusive. Our recent account proposes that “Aha!” reflects a sharp, non-linear update of an internal model when sustained prediction errors collapse, engaging catecholaminergic systems and hippocampal encoding to boost memory (Becker & Cabeza, 2025). We tested this account in two studies using hard-to-identify Mooney images. In a preregistered online experiment, we manipulated update steepness by showing participants a series of identical images going from degraded to clear. We varied only reveal timing for the same images (late rapid reveal vs gradual reveal). The steeper (non-linear) schedule reliably increased self-reported Aha! intensity. In a separate fMRI–pupillometry study with delayed memory, trial-wise insight modulated temporal trajectories in object-selective cortex and prefrontal control regions: we found a sustained positive ramp beginning ~2 s pre-response in both temporal multivoxel pattern change and model-based representational strength, indexing a more non-linear solution update. Pupil dilation exhibited a phasic, non-linear increase in the same window, consistent with noradrenergic arousal. Univariate analyses showed Aha!-scaled responses in hippocampus and canonical catecholamine-recipient areas (SN/VTA, Nucleus Accumbens, Locus coeruleus). Finally, structural equation models indicated that steep cortical update, hippocampal responses, and catecholaminergic proxies jointly accounted for Aha! intensity, with both Aha! and hippocampus predicting insight-related better memory. Together, convergent behavioral, neural, and pupillary evidence indicates that insight reflects a non-linear solution-model update synchronized with catecholaminergic signaling and hippocampal encoding, offering a mechanistic bridge between the phenomenology of “Aha!” and its mnemonic benefits.

Topic Area: LONG-TERM MEMORY: Episodic