ИСТИНА

Due to recent refinement and expanding availability of imaging hardware and genetically encoded dyes, there is an upsurge of fluorescent imaging data reporting Ca 2+ activity in astrocytes. Following this wave of data abundance, an interest to imaging data processing and analysis methods is widely expressed in the literature of the last two years. We suggest open-source python-based solutions for motion correction, estimation of dynamic baseline fluorescence level and detection of Ca 2+ events. Motion correction is based on a multi-template scheme with combined local-global optical flow estimation. Baseline fluorescence level follows a two-step coarse-fine level asymmetric smoothing scheme. Next, noise suppression and event detection is done using local dimensionality reduction and selective grouping of time signals from nearby locations. We further discuss approaches to interpreting the observed spatiotemporal patterns of Ca 2+ events, from analysing distributions of such features as event duration and covered area, to information-theory based analysis of activity converted to symbolic form, to fitting simplified distributed dynamical systems directly to the data.