Graph‐based unsupervised segmentation algorithm for cultured neuronal networks' structure characterization and modeling

Abstract

Large scale phase-contrast images taken at high resolution through the life of a cultured neuronal network are analyzed by a graph-based unsupervised segmentation algorithm with a very low computational cost, scaling linearly with the image size. The processing automatically retrieves the whole network structure, an object whose mathematical representation is a matrix in which nodes are identified neurons or neurons' clusters, and links are the reconstructed connections between them. The algorithm is also able to extract any other relevant morphological information characterizing neurons and neurites. More importantly, and at variance with other segmentation methods that require fluorescence imaging from immunocytochemistry techniques, our non invasive measures entitle us to perform a longitudinal analysis during the maturation of a single culture. Such an analysis furnishes the way of individuating the main physical processes underlying the selforganization of the neurons' ensemble into a complex network, and drives the formulation of a phenomenological model yet able to describe qualitatively the overall scenario observed during the culture growth. V C 2014 International Society for Advancement of Cytometry Key terms neuron image segmentation; cultured neuronal network; neurite tracing; complex networks; network topology analysis; automated tracing; high throughput; neuronal morphology; light microscopy; connectome reconstruction ALONG the past decades, cultured neuronal networks (CNNs) have constituted a fundamental tool for scientists, as one of the benchmark models for the study of the central nervous system. They, indeed, allow performing very well controlled laboratory experiments, thus providing a systematic way to approach fundamental questions, as for example, unveiling the principles and mechanisms underlying memory, connectivity, and even information processing of their in vivo counterparts (1-5).CNNs have also important practical applications, when computer-connected to a real or a simulated robot (to create what is called a hybrot (6-8) or an animat (9,10), respectively), in that scientists are then endowed with the possibility of studying some basic neuronal processes in realistic contexts, such as learning and plasticity. Possibly, the most relevant advantage of CNNs is the unique option they offer of following the footprints of the self (or induced) organization of the network's functionality and dynamics (usually by means of a multi-electrode array [MEA] or calcium fluorescence, recording the CNN electrophysiological data, or inducing electrical stimulations in given spatial positions) together with the monitoring and tracking of the structural organization of the neurons' connectivity along the entire course of the culture's growth (11-13).Although culturing neurons on top of a MEA equipped chamber implicates, in general, only mild constraints, following the development of the culture's structure is

Network-based features for retinal fundus vessel structure analysis

“…All the methods make use of the result of an unsupervised segmentation algorithm that was adapted from the one proposed in [43]. We start by filtering the photograph to enhance the contrast between the vessels and the background, and then we perform the Graph-based segmentation algorithm as proposed in [43], further details on the segmentation process can be found in the supporting information S1 Appendix. We refer to the raw segmentation result as a binary image whose pixels are 1 if they belong to a vessel and 0 if they belong to the background.…”

“…Here we propose a new method that uses concepts inspired in network science [40–42]. We use the segmentation algorithm proposed in [43] to extract, from each digital fundus image, a tree-like graph where the nodes represent branching (bifurcation) points and endpoints, while the links represent vessel segments that connect two nodes. The graphs obtained are characterized by using the concept of node-distance distribution (NDD) [44], which is the fraction of nodes that are at distance d (shortest path) from a given node.…”

Quantitative Determination of Cellular-and Neurite Motility Speed in Dense Cell Cultures

“…Segmentation of neurites in hippocampal cultures have been obtained by several groups on highly resolved neurons to analyze neurite development and 3D structures with conventional algorithms (Meijering, 2010; de Santos-Sierra et al, 2015) or convolution neuronal network (CNN) software implementations (Kandaswamy et al, 2013; Li et al, 2017; Spoerer et al, 2017). Very successful results were paid for by long computation times (de Santos-Sierra et al, 2015) (in the order of minutes per image) or sophisticated neuronal network software, running on large computer systems (Li et al, 2017). One simple segmentation strategy was the transformation of the neuronal network into a binary skeleton, a technique that we tested with SynoQuant's build in grayscale skeletonization algorithm and compared it to binary skeleton algorithm of “Ne.Mo” Software (Billeci et al, 2013).…”

Quantitative phase imaging for label‐free cytometry

“…QPI yields optical path difference maps associated with the specimen of interest and, as such, it is sensitive to both local thickness and the refractive index of the sample. Several QPI related publications have previously appeared in Cytometry Part A, paving the way for this new field of applications (4)(5)(6). A collection of manuscripts in this special issue attests to the fact that QPI is becoming a prominent technique complementary to traditional cytometry technologies and indispensable in dynamic label-free live-cell analysis applications.…”