Presentations

Automated reconstruction of astrocyte morphology with the fast marching method

Brazhe A.R.

Lomonosov Moscow State University, faculty of Biology, department of biophysics SRC Institute of bioorganic chemistry RAS, department of metabolism and redox-biology

Morphological remodeling of astrocytes plays a significant role in the functional state of the central nervous system and in the genesis of neuropathologies. The morphological heterogeneity of astrocytes across various regions of the nervous system is also a subject of current research interest. This underscores the need for automated methods for reconstruction and analysis of astrocyte morphological features from fluorescent confocal imaging data.

We propose an algorithm for morphological reconstruction based on the fast marching method, followed by the construction of merging gradient descent paths, which collectively model the astrocyte structure as a directed acyclic graph. The resulting reconstructions can be directly utilized in morphology description algorithms based on methods from applied topology. The applicability of the method is demonstrated by describing astrocyte morphological changes during postnatal development.

The work is supported by RSF grant #23-44-00103