Automated Detection of Soma Location and Morphology in Neuronal Network Cultures

Abstract

Automated identification of the primary components of a neuron and extraction of its sub-cellular features are essential steps in many quantitative studies of neuronal networks. The focus of this paper is the development of an algorithm for the automated detection of the location and morphology of somas in confocal images of neuronal network cultures. This problem is motivated by applications in high-content screenings (HCS), where the extraction of multiple morphological features of neurons on large data sets is required. Existing algorithms are not very efficient when applied to the analysis of confocal image stacks of neuronal cultures. In addition to the usual difficulties associated with the processing of fluorescent images, these types of stacks contain a small number of images so that only a small number of pixels are available along the z-direction and it is challenging to apply conventional 3D filters. The algorithm we present in this paper applies a number of innovative ideas from the theory of directional multiscale representations and involves the following steps: (i) image segmentation based on support vector machines with specially designed multiscale filters; (ii) soma extraction and separation of contiguous somas, using a combination of level set method and directional multiscale filters. We also present an approach to extract the soma’s surface morphology using the 3D shearlet transform. Extensive numerical experiments show that our algorithms are computationally efficient and highly accurate in segmenting the somas and separating contiguous ones. The algorithms presented in this paper will facilitate the development of a high-throughput quantitative platform for the study of neuronal networks for HCS applications.

Automated 3-D mapping of single neurons in the standard brain atlas using single brain slices

“…Then, circular structures, namely the labeled somas, were located using Hough transformation 30, 31 . AMaSiNe allows users to tune the parameters for cell detection, such as the diameter range of labeled somas, for flexible application of the algorithm for soma and/or neuropil detection 32, 33 . Because the 3-D reconstructed brain accurately fit into ARA, detected neurons in each slice were directly positioned on ARA, the common 3-D reference space.…”

Multiscale Analysis of Neurite Orientation and Spatial Organization in Neuronal Images

“…In fact, the segmentation and centerline tracing steps of the algorithm, which are adapted from our previous work in Jimenez et al (2013, 2015a) and Ozcan et al (2015), is already available both in the 2D and 3D settings. The extension of directional filtering to 3D is conceptually straightforward.…”

“…In case images contain additional structures such as somas or other blob-like objects, our algorithms can be applied after first removing such blob-like structures from the segmented images. This task can be addressed, for instance, by using a method for automated soma detection and segmentation recently developed by some of the authors (Ozcan et al, 2015). This method uses a geometric descriptor called ‘directionality ratio’ to automatically separate vessel-like structure from more isotropic ones and was successfully applied to automatically separate somas from neurites in confocal images of neuronal cultures and brain tissue.…”

“…This is the most computationally-intensive part of the algorithm but it needs to be run only once as long as the segmentation algorithm is applied to images of the same type, e.g., same cell type and microscope setting. The entire procedure, including the training stage, is fully automated and the algorithm was shown to perform very competitively even on challenging 2D and 3D datasets (Jimenez et al, 2013, 2015a; Ozcan et al, 2015). …”

“…From the segmented images (step 1 of the algorithm), seed points for the tracing routine are computed through an appropriately designed distance transform and then connected. The tracing routine is computationally very efficient and was shown to perform very competitively even with challenging neuronal imaging data (Jimenez et al, 2013, 2015a; Ozcan et al, 2015). …”