P. G. FAHEY1, B. CELII2, C. PAPADOPOULOS2, E. FROUDARAKIS2, T. MACRINA5, S. DORKENWALD5, N. L. TURNER5, P. ZHOU8, E. COBOS2, S. PAPADOPOULOS1, Z. DING3, D. YATSENKO2, E. Y. WALKER3, C. SMITH4, R. J. COTTON7, K. LEE5, J. ZUNG5, I. TARTAVULL5, D. IH5, D. BUNIATYAN5, W. SILVERSMITH6, J. WU6, N. KEMNITZ6, R. LU6,  S. POPOVYCH6, W. WONG6, A. WILSON6, J. BUCHANAN10 ,M. M. TAKENO10, A. L. BODOR10, D. J. BUMBARGER10, A., BLECKERT10, F., COLLMAN10, F. H. SINZ3, L. M. PANINSKI9, X. S. PITKOW3, N. M. DA COSTA10, R. REID11, H. SEUNG6, A. S. TOLIAS3, *J. REIMER2;

1Dept. of Neurosci., 3Neurosci., 4Dept of Neurosci., 2Baylor Col. of Med., Houston, TX; 6Princeton Neurosci. Inst., 5Princeton Univ., Princeton, NJ; 7Shirley Ryan Abilitylab, Chicago, IL; 8Columbia Univ., New York City, NY; 9Columbia Univ., New York, NY; 10Allen Inst. For Brain Sci., Seattle, WA; 11Neural Coding, Allen Inst. for Brain Sci., Seattle, WA  *Presenting Author

Although significant insight can be gleaned from analysis of neural activity or anatomy separately, answering many fundamental mechanistic questions in systems neuroscience requires information about both neural activity and connectivity in the same animal. Enormous strides have been made over the past decade in our ability to record activity from large populations of neurons distributed across multiple regions of the brain, and improvements in EM imaging techniques and the application of modern machine learning methods for segmentation of EM volumes have together enabled accurate reconstruction of increasingly large volumes of neural tissue. Here we describe an analysis of the connectivity of putative excitatory axons in a small volume of mouse primary visual cortex onto functionally-characterized excitatory cells. At Baylor College of Medicine, we recorded visual responses from neurons located in L2/3 of primary visual cortex of a 34 day-old male mouse expressing GCaMP6f in pyramidal cells under control of CamKII-Cre/tTA. The mouse was then shipped to the Allen Institute in Seattle where electron microscopy (EM) was performed on a ~200 x 150 x 100 micron volume, which was subsequently densely reconstructed at Princeton University. The EM data was subsequently used as ground-truth segmentation for a novel method to extract functional data developed at Columbia University. From this data we were able to analyze the projection patterns of tens of thousands of reconstructed axons onto hundreds of functionally-characterized neurons and isolated apical dendrites. Several methodologies have been used to find evidence for “like-to-like” connectivity (increased connectivity for cells with similar tuning preferences) via spine imaging (Iacaruso, et al 2017), combined in vivo imaging and in vitro multipatching (Ko et al. 2011; Cossell et al. 2015), and combined in vivo imaging with EM reconstruction (Bock et al. 2011; Lee et al. 2016). Here, we examine this question in this largest functionally-imaged and densely-reconstructed calcium imaging/EM data set collected to date in order to elucidate the principles of structure/function relationships including shared input and higher-order motifs.