Explore a comprehensive tutorial on calcium imaging data cell extraction techniques, focusing on the CNMF-E method developed for analyzing microendoscopic data. Learn about the challenges of extracting single-neuronal activity from microendoscopic recordings and how CNMF-E improves upon existing methods. Dive into the matrix factorization approach, which accurately separates background fluctuations and simultaneously demixes and denoises neuronal signals. Compare CNMF-E with independent components analysis and constrained nonnegative matrix factorization approaches, and discover its superior performance in detecting well-isolated neural signals, especially in noisy data regimes. Follow along with a hands-on demonstration of the MATLAB version of CNMF-E, applicable to both 1-photon microendoscopic and 2-photon imaging data. Gain insights into the method's framework, including background modeling, optimization problems, and the overall pipeline for cell extraction.
Overview
Syllabus
Install CNMF-E
Outline
Calcium Imaging
Cell Extraction Problem
Two Approaches
ROI analysis
Matrix Factorization Approach
CNMF framework
Microendoscope
Microendoscopic Data
Model of the background
Three subproblems
Optimization problem
Interventions
Initialization
Pipeline
Taught by
MITCBMM