A novel method for analyzing images of live nerve cells

Analysis of images from live-cell experiments is a central activity to studying the effects of stimulation on neuronal behavior. Image analysis techniques currently used to study these effects rely for the most part on the salience of the neuronal structures within the image. In both fluorescent and electron microscopy, neuronal structures are enhanced and therefore easy to distinguish in an image. Unlike images obtained via fluorescent or electron microscopy, however, images produced via transmission microscopy (e.g., bright field, phase contrast, DIC) are significantly more difficult to analyze because there is little contrast between the object-of-interest and the image background. This difficulty is amplified when a time-dependent sequence of images are to be analyzed, because of the corresponding large data sets. To address this problem, we introduce a novel approach to the analysis of images of live cells captured via transmission microscopy that takes advantage of commercially available software and the Fourier transform. Specifically, our approach utilizes several morphological functions in MATLAB to enhance the contrast of the cells with respect to the background, which is followed by 2-D Fourier analysis to generate a spectrum from which the orientation and alignment of cells and their processes can be measured. We show that this method can be used to simplify the interpretation of complex structure in images of live neurons obtained via transmission microscopy and consequently, discover trends in neurite development following different types of stimulation. This approach provides a consistent and reliable tool for analyzing changes in cell structure that occurs during live-cell experiments.

► An image processing technique using morphological operations in MATLAB is described. ► This technique improves detection of neurites in transmission microscopy images. ► Analysis of neurite changes in complex images is simplified using this technique. ► Sequential images can be batch processed through automation of this technique. ► This technique accelerates the analysis of live-cell images of neurons.