Title: Dopamine signatures of excessive and compulsive cocaine and fentanyl use

Speaker: Ke Chen (Wang Lab)

Abstract: Excessive and compulsive drug use despite adverse consequences is a hallmark of addiction, yet individuals differ markedly in their vulnerability. Drugs of abuse alter endogenous dopamine (DA) signaling, but shared principles linking DA dynamics to compulsive use across individuals and drug classes are unclear. Here, we monitored DA release in nucleus accumbens (NAc) medial shell during cocaine or fentanyl self-administration, with or without punishment, in large mouse cohorts. Contingent cocaine and fentanyl self-administration evoked complex and individually distinct DA dynamics, yet a robust negative correlation emerged across both drugs: high takers showed lower drug-evoked DA signals. Under punished drug taking, cocaine and fentanyl produced distinct DA signatures of compulsivity. A computational model grounded in the Actor-Critic temporal-difference (TD) learning framework with considerations ofinternal states, action cost and drug-specific effects captured the observed diversity in DA dynamics across conditions, unifying NAc DA as encoding TD reward prediction errors in addiction.

Title: Machine learning-guided rhodopsin engineering enables sensitive all-optical voltage imaging and optogenetics

Speaker:Miranda Dawson (Fan Lab)

Abstract:  Understanding how neural circuits change during learning and disease requires tools that can measure fast synaptic voltage signals with high spatial and temporal resolution. Genetically encoded voltage indicators (GEVIs) enable optical recording of membrane potential dynamics from genetically defined neurons, but current sensors lack the sensitivity and robustness needed to reliably resolve subthreshold synaptic events on single trials in vivo during behavior. We develop and apply next-generation rhodopsin-based GEVIs optimized for brightness, voltage sensitivity, and kinetics. Using a machine learning-guided protein engineering framework, candidate indicators are computationally prioritized across multiple performance parameters and experimentally benchmarked using all-optical electrophysiology in neuronal cultures. Top-performing variants are integrated with two-photon optogenetic approaches to establish an all-optical platform capable of resolving unitary synaptic excitation and inhibition in intact neural circuits during behavior. By enabling direct optical measurement of synaptic signaling during behavior, this work overcomes a critical technical barrier in systems neuroscience and provides new tools for investigating circuit plasticity mechanisms underlying learning, memory, and neurological disorders.