Comparing basal dendrite branches in human and mouse hippocampal CA1 pyramidal neurons with Bayesian networks

Abstract

Pyramidal neurons are the most common cell type in the cerebral cortex. Understanding how they differ between species is a key challenge in neuroscience. A recent study provided a unique set of human and mouse pyramidal neurons of the CA1 region of the hippocampus, and used it to compare the morphology of apical and basal dendritic branches of the two species.The study found inter-species differences in the magnitude of the morphometrics and similarities regarding their variation with respect to morphological determinants such as branch type and branch order. We use the same data set to perform additional comparisons of basal dendrites. In order to isolate the heterogeneity due to intrinsic differences between species from the heterogeneity due to differences in morphological determinants, we fit multivariate models over the morphometrics and the determinants. In particular, we use conditional linear Gaussian Bayesian networks, which provide a concise graphical representation of the independencies and correlations among the variables. We also extend the previous study by considering additional morphometrics and by formally testing test whether a morphometric increases or decreases with the distance from the soma. This study introduces a multivariate methodology for inter-species comparison of morphology. mouse terminal branches 15 ). Besides missing existing differences, the converse is also possible and inter-species that were detected with a single test might be insignificant once we account for the determinants. The authors of Ref. 15 accounted for the determinants precisely by splitting the branches into terminal and non-terminal ones, and further down according to branch order, and then testing hypotheses of location difference (Kruskal-Wallis test) between pairs of obtained subsets of branches (e.g., mouse terminal branches of branch order two are as long as human terminal branches of the same branch order).Instead of splitting the branches according to the determinants and running multiple tests, we could specify a multivariate statistical model over the morphometrics and the morphological determinants. Models such as Bayesian networks (BNs) [16][17][18] can represent the probabilistic relationships among the variables of a domain, while algorithms for learning Bayesian networks from data can uncover such relationships. In the above example we would have three random variables -species, branch type and branch length-and learning a Bayesian network from the data would tell us that the species variable and the length variable are marginally independent yet dependent given branch type. Probabilistic dependence does not imply different locations (e.g., mean or median), and a morphometric with the same mean yet different variance in the two species is indeed dependent on the species variable. Thus, only after identifying a dependence between the species variable and a morphometric (with the Bayesian network) we would proceed to test for location difference between species. In addition, Bayesian networks can sim...