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

When Stability Meets Change: Neural Dissociation of Statistical Learning and Adaptive Flexibility in the Motor Domain

Poster Session E - Monday, March 9, 2026, 2:30 – 4:30 pm PDT, Fairview/Kitsilano Ballroom

Erik Chihhung Chang¹ (), Chiao-En Chan¹, Denise Hsien Wu¹; ¹National Central University, Taiwan

Statistical learning (SL) enables efficient pattern extraction, yet real-world environments demand flexibility when regularities shift. Applying functional magnetic resonance imaging (fMRI) with an interleaved-dual-structure serial reaction time task (IDS-SRTT), we investigated the neural substrates underlying this fundamental tension between learning stability and adaptive flexibility. Twenty participants learned an initial motor sequence before gradually transitioning to a novel sequence across sessions, with proportions systematically shifting from 100:0 to 0:100. Behaviorally, individuals demonstrating superior initial learning showed reduced flexibility when adapting to the new structure, confirming a trade-off between exploitation and exploration. Neuroimaging results revealed distinct neural signatures for learning versus switching. When comparing structured to random trials during initial SL, we observed repetition suppression effects, with reduced activation in cortical motor areas, cerebellum, and the frontoparietal network (FPN) and the salience network (SN), which reflected successful pattern consolidation. Conversely, during rule-switching phases, both FPN and SN exhibited increased activation, with critical differences in their temporal dynamics to highlight functional dissociation between these networks in detecting and adapting to environmental changes. The SN showed peak activation during early switching when the new rule first emerged, consistent with conflict detection and prediction error signaling. In contrast, FPN activation remained elevated throughout all switching phases, suggesting sustained goal-directed control, inhibition of outdated patterns, and working memory updating. Our findings provide the first integrated neural account of the learning-flexibility trade-off in motor sequence learning, demonstrating that a common cognitive control system centered on FPN and SN dynamically reconfigures to balance stability and change.

Topic Area: LONG-TERM MEMORY: Skill Learning