Automated detection of intercellular signaling in astrocyte networks using the converging squares algorithm

Abstract

Intercellular calcium waves in central nervous system astrocyte networks underline the principle mechanism of cell signaling in astrocyte syncsytiums, which putatively contribute to the modulation of neuronal signaling and metabolic regulation. In support of carrying out systems level analyses of astrocyte networks, we have optimized and validated the converging squares image segmentation algorithm to automatically detect the relative spatial locations of all cells in a visible network as a preliminary step towards analyzing the dynamics of astrocyte intracellular calcium transients, which are the signals that mediate intercellular calcium waves. We used the temporal derivatives of pixel intensities as the data source for the algorithm. The method works by converging progressively smaller squares until the signal peak is reached. It is robust to noise and performs comparably to manual cell signal identification, but is much faster and efficient. This is the first reported application of this algorithm to glial networks that we are aware of.