A Novel Tool for Neurite Tracing and Analysis in Fluorescence Microscopy
E. Meijering, M. Jacob, J.-C.F. Sarria, P. Steiner, H. Hirling, M. Unser
Proceedings of the EMBO Workshop on Advanced Light Microscopy, Third International Meeting of the European Light Microscopy Initiative (ELMI'03), Barcelona (Barcelona), Spain, June 11-13, 2003, in press.
Please do not bookmark the In Press papers as content and presentation may differ from the published version.
In a few seconds, you should be redirected to the published version. The preprint version is still available here
For the investigation of the proteins and mechanisms involved in neurite formation, accurate and user-friendly morphometric image analysis tools are indispensable. In the past, a number of automated tools have been developed for the tracing and quantification of neurite outgrowth in three-dimensional confocal microscopy image stacks. In two-dimensional fluorescence microscopy images of cell cultures, however, fully automated neurite tracing is seriously impeded by ambiguities caused by gross intensity discontinuities and the crossing of neurites. It is not surprising, therefore, that manual delineation using rudimentary drawing tools still is the most frequently used approach to neurite tracing, as our recent study of over one hundred papers related to neurite outgrowth published in the last two years has revealed. To facilitate neurite analysis in fluorescence microscopy, we have developed a novel tracing technique that is able to cope with the mentioned problems. Contrary to existing techniques, which are based mostly on recursive vectorial tracking, our technique is based on a powerful optimization algorithm capable of yielding very plausible tracings even under extreme noise conditions and intensity discontinuities. In addition, it does not depend on any hard thresholds for its operation. The algorithm contains only a few parameters which can be kept fixed for a wide range of images. The software tool built around the tracing algorithm is computer-platform independent and provides a number of functionalities facilitating quantitative and statistical analysis. In this presentation we highlight the principles underlying our neurite tracing tool and present the findings of a currently ongoing validation study showing the precise improvement in accuracy and reproducibility as well as the reduction in user interaction and variability compared to fully manual tracing. The presentation is concluded with a demonstration of the functionalities of the tool.