Machine learning to evaluate neuron density in brain sections

Abstract

Imaging applications often produce large numbers of data sets, which need to be processed in a uniform and unbiased manner to obtain precise information about the number and size of cells or cell densities in different regions of the brain. Machine learning is a novel method here introduced to adjust algorithms to the biological requirements and to evaluate cellular features of tissue samples in an automated manner. In this chapter we describe methods to prepare mouse brain tissue for subsequent image processing and data evaluation. We give information in a step-by-step manner how to choose and perform appropriate fixation protocols, decide for suitable sectioning, and give hints what to consider when performing immunofluorescence stainings. Furthermore, we introduce the Machine Learning-Based Image Segmentation (MLBIS) to determine neuronal cell density in brain slices. © 2014 Springer Science+Business Media New York.