Search Papers | Poster Sessions | All Posters
Poster A9 in Poster Session A - Tuesday, August 6, 2024, 4:15 – 6:15 pm, Johnson Ice Rink
State-dependent Online Reactivations for Different Learning Strategies in Virtual Foraging
Sangkyu Son1,2, Maya Wang3, Benjamin Hayden4, Seng Bum Yoo1,2,4; 1Sungkyunkwan University, 2Center for Neuroscience Imaging Research, 3National Institution of Mental Health, 4Baylor College of Medicine
Learners often seek multiple objectives paired with different strategies, and the brain needs to (re)activate representation fit for each objective. We questioned if neural reactivation, typically viewed as stereotyped during the learning process, is the subject of control to promote different types of navigational learning. We trained macaques to forage in a first-person virtual maze. Two behavioral repertories emerged from the low-level features; one seeks reward (explore-like) and the other for information (exploit-like). While alternating among two objectives, the orbitofrontal (OFC) and retrosplenial cortices (RSC) preplayed the future optimal path and goal itself, specifically when prioritizing reward. When prioritizing information, both cortices strategically devalued the uninformative paths with reduced reactivation. Meanwhile, the reactivation of the fastest path that leads to the goal was reinforced when prioritizing reward. The artificial agent foraging in the identical maze confirmed that RSC and OFC devaluating the uninformative path and reinforcing the reward-optimal path promotes the reward rate. These results highlight that neural ensemble adaptively aids the learning process as per the need of each moment.
Keywords: virtual-reality navigation retrosplenial cortex orbitofrontal cortex neural reactivation