Frontiers in Neuroscience (Nov 2012)

Robust automated amygdala segmentation via multi-atlas diffeomorphic registration

  • Jamie eHanson,
  • Jung eSuh,
  • Brendon eNacewicz,
  • Matt eSutterer,
  • Amy eCayo,
  • Diane eStodola,
  • Cory eBurghy,
  • Hongzhi eWang,
  • Brian eAvants,
  • Paul eYushkevich,
  • Marilyn eEssex,
  • Seth ePolllak,
  • Richard J Davidson

DOI
https://doi.org/10.3389/fnins.2012.00166
Journal volume & issue
Vol. 6

Abstract

Read online

Here, we describe a novel method for volumetric segmentation of the amygdala from MRI images collected from 35 human subjects. This approach is adapted from open-source techniques employed previously with the hippocampus (Suh et al., 2011; Wang et al., 2011a; Wang et al., 2011b). Using multi-atlas segmentation and machine learning-based correction, we were able to produce automated amygdala segments with high dice (Mean= 0.918 for the left amygdala; 0.916 for the right amygdala) and Jaccard coefficients (Mean= 0.850 for the left; 0.846 for the right) compared to rigorously hand-traced volumes. This automated routine also produced amygdala segments with high intra-class (consistency=.830, absolute agreement =.819 for the left; consistency=. 786, absolute agreement =. 783 for the right) and bivariate (r =.831 for the left; r =0.797 for the right) correlations compared to hand-drawn amygdala. Our results are discussed in relation to results from other cutting-edge segmentation techniques, as well as commonly- available approaches to amygdala segmentation (e.g., Freesurfer). We believe this new technique has broad application to research with large sample sizes for which amygdala quantification might be needed.

Keywords