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

Poster F in Poster Session F: Thursday, August 6, 1:45 – 3:30 pm, Kimmel Center, Shorin & Rosenthal Rooms

BCI-induced Dynamic Shifts of Valence Representations in the Primate Amygdala

Shaked Emily Leibovitz¹, Dean Halperin¹, Evgeniya Kornilova¹, Maya Sheffer¹, Liliya Iskhakova¹, Rony Paz¹; ¹Weizmann Institute of Science

Presenter: Shaked Emily Leibovitz

Adaptively identifying and responding to environmental threats is fundamental for survival. The amygdala plays a central role in encoding stimulus valence and guiding emotional learning, yet we still lack a comprehensive understanding of how valence is represented and adapts at the population level. Brain-computer interfaces (BCIs) enable direct manipulation of neural activity through volitional control of single neurons or populations guided by sensory feedback. Here, we developed an emotional BCI paradigm, providing real-time auditory feedback driven by the activity of a single amygdala neuron, guiding primates to shift spiking activity from an aversive-associated cue towards an appetitive-associated cue. Using multi-channel amygdala spike band power signals during memory acquisition, we constructed a low-dimensional neural subspace and developed a geometric framework that decomposes valence-specific trajectories into velocity components to quantify the directional dynamics of valence encoding. We demonstrate that successful BCI can systematically reshape population-level activity in the desired direction, namely, pushing aversive representations toward appetitive ones, a pattern apparent during BCI learning and preserved following the session. Our results provide insight into adaptive emotional learning and how it can be modulated by using the brain's natural learning mechanisms.

Topic Area: Methods, Tools, Theory & Neural Coding