Rapid semi-automated segmentation and analysis of neuronal morphology and function from confocal image data

Confocal microscopy combined with cellular labeling techniques can be an effective method for imaging the morphology of a cell as well as various functional characteristics in vivo. Current analysis methods are manual, and therefore, time-consuming and prone to error. Through the development of custom algorithms and application design, the analysis process can be improved to decrease analysis time and increase reproducibility. Utilizing off-the-self PC hardware and software, a custom application was designed that would provide useful three-dimensional (3D) segmentation and analysis tools to analyze confocal image data of neurons. Techniques such as dynamic thresholding, adaptive filtering, and morphological processing were implemented to provide a robust and efficient analysis package. The automated method was compared with the standard manual method using two metrics - reproducibility and overall time necessary for analysis. The semi-automated method was more time efficient with very high reproducibility. Additional studies are necessary to further assess and improve upon the automated technique.