Cheng Huang, PhD – Department of Neuroscience, Washington University School of Medicine in St. Louis — CSB Seminar

CSB Departmental Seminar

Friday, January 9th @ 11:00 am

SPEAKER: Cheng Huang, PhD – Department of Neuroscience, Washington University School of Medicine in St. Louis

TITLE: Decoding dopamine-mediated learning algorithms in the Drosophila brain

ABSTRACT: My research program focuses on understanding the neural circuits underlying memory formation, storage, and retrieval. Using Drosophila as a model system, I am particularly interested in how dopamine signaling modulates neural plasticity and contributes to adaptive behaviors. We employ a multidisciplinary approach, combining advanced imaging techniques, quantitative behavioral assays, and computational modeling to dissect the complex dynamics of dopamine neurons during learning and memory processes. Our latest study reveals how innate valence information regulates memory dynamics and uncovers a mechanism by which short-term memory gates long-term memory formation through dopamine signaling. Specifically, we examined the MB circuit for learning aversive associations. First, we discovered a subset of dopamine neurons that encode negative innate sensory valence information. During learning, these dopamine neurons dynamically integrate this innate valence with learned associations (e.g., the association of the odor with an electric shock). This integrated valence signal then controls long-lasting neural plasticity in downstream neurons. To understand how this learned valence arises, we identified an inhibitory feedback pathway from short-term memory encoding neurons to the dopamine neurons. This pathway underlies the learned valence represented in the dopamine signal, ensuring that only consistent associations generate long-lasting memories. Further, we constructed a computational model using constraints derived from both our large-scale voltage imaging dataset and Drosophila connectome data. We uncovered the computational principle and benefits of dopamine-based valence integration in this learning network. Collectively, our results provide insights into how ecologically relevant information directly governs memory formation and demonstrate the power of integrative research approaches in studying complex brain processes.

HOST: Qian Lin

LOCATION: Cell and Systems Biology, 25 Harbord Street, Suite 432

LIVESTREAM LINK: https://csb.utoronto.ca/live-stream/