Deformation-based brain morphometry in rats

Abstract

Magnetic resonance imaging (MRI)-based morphometry provides in vivo evidence for macro-structural plasticity of the brain. Experiments on small animals using automated morphometric methods usually require expensive measurements with ultra-high field dedicated animal MRI systems. Here, we developed a novel deformation-based morphometry (DBM) tool for automated analyses of rat brain images measured on a 3-Tesla clinical whole body scanner with appropriate coils. A landmark-based transformation of our customized reference brain into the coordinates of the widely used rat brain atlas from Paxinos and Watson (Paxinos Atlas) guarantees the comparability of results to other studies. For cross-sectional data, we warped images onto the reference brain using the low-dimensional nonlinear registration implemented in the MATLAB software package SPM8. For the analysis of longitudinal data sets, we chose high-dimensional registrations of all images of one data set to the first baseline image which facilitate the identification of more subtle structural changes. Because all deformations were finally used to transform the data into the space of the Paxinos Atlas, Jacobian determinants could be used to estimate absolute local volumes of predefined regions-of-interest. Pilot experiments were performed to analyze brain structural changes due to aging or photothrombotically-induced cortical stroke. The results support the utility of DBM based on commonly available clinical whole-body scanners for highly sensitive morphometric studies on rats.