Speaker: Jason Moore, Ph.D., Postdoctoral Fellow, Basu Lab, NYU School of Medicine
Topic: “Algorithmic Segmentation of Dendrites in Mouse Hippocampus”
Dendrites support non-linear integration of synaptic inputs to facilitate important brain-wide functions such as plasticity and feature selective tuning. Hippocampal neurons known as “place cells” are tuned to specific locations in an environment, but the tuning properties of hippocampal dendrites are largely unexplored.
In-vivo two photon calcium imaging offers high resolution, subcellular compartment-specific access to activity in somata, dendrites, axons and even synaptic boutons. However, a major hurdle in processing such data is the inability to use existing automated segmentation algorithms to detect and segment these fine-scale regions of interests.
Here I describe continuing efforts to develop an extension to the CaImAn analysis package to initialize regions of interest (ROIs) that cover all active dendrites and soma in the field of view. The algorithm is optimized to have minimal space and time requirements, and closely matches manually labeled datasets.