Morphological analysis of isolated astrocytes using phase-contrast microscopy

A new method of data collection and analysis for creating digital doubles of cellular systems is presented. To collect data, experiments were performed to visualize astrocyte growth for 18 days from the second day of cultivation. The obtained images were labeled and a Mask R-CNN neural network was trained based on them. From the detected astrocytes, features such as lengths of all astrocyte outgrowths $L_t$, number of nodes $Nb_t$ and total astrocyte area $V_t$ were extracted. The obtained parameters were determined as functions of cultivation time. The obtained time dependencies will be used to create a digital twin of isolated astrocyte cells.