Search

The recent development of cellular imaging techniques and the application of genetically encoded sensors of neuronal activity led to significant methodological progress in neurobiological studies. These methods often result in complex and large data sets consisting of image stacks or sets of multichannel fluorescent images. The detection of synapses, visualized by fluorescence labeling, is one major challenge in the analysis of these datasets, due to variations in synapse shape, size, and fluorescence intensity across the images. For their detection, most labs use manual or semi-manual techniques that are time-consuming and error-prone. We developed SynEdgeWs, a MATLAB-based segmentation algorithm that combines the application of an edge filter, morphological operators, and marker-controlled watershed segmentation. SynEdgeWs does not need training data and works with low user intervention. It was superior to methods based on cutoff thresholds and local maximum guided approaches in a realistic set of data. We implemented SynEdgeWs in two automatized routines that allow accurate, direct, and unbiased identification of fluorescently labeled synaptic puncta and their consecutive analysis. SynEval routine enables the analysis of three-channel images, and ImgSegRout routine processes image stacks. We tested the feasibility of ImgSegRout on a realistic live-cell imaging data set from experiments designed to monitor neurotransmitter release using synaptic phluorins. Finally, we applied SynEval to compare synaptic vesicle recycling evoked by electrical field stimulation and chemical depolarization in dissociated cortical cultures. Our data indicate that while the proportion of active synapses does not differ between stimulation modes, significantly more vesicles are mobilized upon chemical depolarization.

The recent development of cellular imaging techniques and the application of genetically encoded sensors of neuronal activity led to significant methodological progress in neurobiological studies. These methods often result in complex and large data sets consisting of image stacks or sets of multichannel fluorescent images. The detection of synapses, visualized by fluorescence labeling, is one major challenge in the analysis of these datasets, due to variations in synapse shape, size, and fluorescence intensity across the images. For their detection, most labs use manual or semi-manual techniques that are time-consuming and error-prone. We developed SynEdgeWs, a MATLAB-based segmentation algorithm that combines the application of an edge filter, morphological operators, and marker-controlled watershed segmentation. SynEdgeWs does not need training data and works with low user intervention. It was superior to methods based on cutoff thresholds and local maximum guided approaches in a realistic set of data. We implemented SynEdgeWs in two automatized routines that allow accurate, direct, and unbiased identification of fluorescently labeled synaptic puncta and their consecutive analysis. SynEval routine enables the analysis of three-channel images, and ImgSegRout routine processes image stacks. We tested the feasibility of ImgSegRout on a realistic live-cell imaging data set from experiments designed to monitor neurotransmitter release using synaptic phluorins. Finally, we applied SynEval to compare synaptic vesicle recycling evoked by electrical field stimulation and chemical depolarization in dissociated cortical cultures. Our data indicate that while the proportion of active synapses does not differ between stimulation modes, significantly more vesicles are mobilized upon chemical depolarization., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2022 Guhathakurta, Akdaş, Fejtová and Weiss.)