The overall goal of the Neural Engineering Section is to develop and apply cutting-edge optical imaging technologies to study neural circuit changes that lead to long-term drug addiction and relapse. We developed a custom miniature fluorescence microscope (miniScope) that enables large-scale in vivo calcium imaging in freely moving mice. Combined with a Gradient Index (GRIN) lens implanted into a mouse brain, the miniScope imaging system allows us to record calcium activity from hundreds of neurons in deep brain regions while mice freely perform complex behaviors. Using this in vivo imaging technique, we recorded medium spiny neuron (MSN) activity from the dorsal striatum of freely moving mice. We showed that both the direct and indirect pathway MSNs display similar neural cluster organization that is functionally and structurally stable. Using machine-learning algorithms, we showed that cluster activities can be used to predict locomotion relevant behavioral states and locomotion velocities. We further showed that acute cocaine administration alters striatum network dynamics in both direct and indirect pathways. We next applied miniScope and GRIN lens to image neural activity in the prelimbic region (PrL) of the medial prefrontal cortex (mPFC) during mouse social exploration. We identified distinct and dynamic ON and OFF neural ensembles that encode social exploration. We further showed that dysfunctions in these neural ensembles are associated with abnormal social exploration elicited by the psychedelic drug phencyclidine (PCP). In our most recent study, we developed deep behavior mapping (DBM), a self-supervised learning method for fine-grained analysis of behavioral microstates obtained from behavior video data. We combined DBM with miniScope in vivo calcium imaging in freely moving mice to show that neurons in the PrL represent different behaviors, especially the sequence of behaviors involved in lever pressing for food reward and various sequences of exploratory behaviors. Our findings showed that PrL neurons are associated with the most important events in a behavioral sequence, and encompass the entire sequence from start to finish. Furthermore, PrL neurons that were not previously tuned to any behavior came to represent newly learned behaviors. Together, our results underscore the unique advantage of combining in vivo deep brain calcium imaging in freely behaving mice with advanced computational analyses to identify neural coding mechanisms of animal behavior. Ongoing studies in our Section focus on neural coding mechanisms of reward seeking in the nucleus accumbens (NAc) in mice and rats. Additionally, we continue with the development of our new generation miniScope system, which incorporates a dual LED and liquid lens design. This design enables dual-color imaging to study functional interactions between different local neural populations (e.g., direct and indirect pathway MSNs in NAc, pyramidal neurons and interneurons in the prefrontal cortex, neuron versus glia network, etc.), or combined optical imaging with optogenetic manipulation of neural activity in real time. Finally, we will continue to develop new behavior mapping methods to perform fine-grained behavior analyses in freely behaving mice and rats, which will allow us to go beyond rate-based study of operant behavior and other freely moving complex behaviors in rodents. Combining new generation miniScope imaging with more detailed mapping of complex behaviors will enable us to stay at the forefront of elucidating how alterations in spatially and temporally organized neural activity result in long-term drug addiction and relapse in animal models. Relevance to NIDA mission: Our proposed mechanistic studies into neural coding of reward seeking, and the development of cutting-edge tools enabling study of neural coding mechanisms, all fit with NIDAs mission of advance science on the causes and consequences of drug use and addiction. In this regard, the development of our miniScope system is highlighted in the 2016-2020 NIDA Strategic Plan. Importantly, miniature microscope in vivo imaging in freely behaving animals was chosen by Nature Methods as Method of The Year 2018.