Multiscale vision model highlights spontaneous glial calcium waves recorded by 2-photon imaging in brain tissue

Intercellular glial calcium waves (GCW) constitute a signaling pathway which can be visualized by fluorescence imaging of cytosolic Ca2+ changes. Reliable detection of calcium waves in multiphoton imaging data is challenging because of low signal-to-noise ratio. We modified the multiscale vision model (MVM), originally employed to detect faint objects in astronomy data to process stacks of fluorescent images. We demonstrate that the MVM identified and characterized GCWs with much higher sensitivity and detail than pixel thresholding. Origins of GCWs were often associated with prolonged secondary Ca2+ elevations. The GCWs had variable shapes, and secondary GCWs were observed to bud from the primary, larger GCW. GCWs evaded areas shortly before occupied by a preceding GCW instead circulating around the refractory area. Blood vessels uniquely reshaped GCWs and were associated with secondary GCW events. We conclude that the MVM provides unique possibilities to study spatiotemporally correlated Ca2+ signaling in brain tissue. ► We modify multiscale vision model (MVM) to process stacks of fluorescent images. ► The modified MVM was applied to characterization of spontaneous glial calcium waves. ► The method is more sensitive and robust than ad hoc smooth-and-threshold approach. ► We demonstrate non-linear nature of glial calcium waves (GCWs). ► Blood vessels reshape GCWs and are associated with generation of secondary GCWs.