The diffeomorphometry of temporal lobe structures in preclinical Alzheimer's disease

Michael I. Miller, Laurent Younes, J. Tilak Ratnanather, Timothy Brown, Huong Trinh, Elizabeth Postell, David S. Lee, Mei Cheng Wang, Susumu Mori, Richard O'Brien, Marilyn Albert

Research output: Contribution to journalArticlepeer-review


This paper examines morphometry of MRI biomarkers derived from the network of temporal lobe structures including the amygdala, entorhinal cortex and hippocampus in subjects with preclinical Alzheimer's disease (AD). Based on template-centered population analysis, it is demonstrated that the structural markers of the amygdala, hippocampus and entorhinal cortex are statistically significantly different between controls and those with preclinical AD. Entorhinal cortex is the most strongly significant based on the linear effects model (p <.0001) for the high-dimensional vertex- and Laplacian-based markers corresponding to localized atrophy. The hippocampus also shows significant localized high-dimensional change (p <.0025) and the amygdala demonstrates more global change signaled by the strength of the low-dimensional volume markers. The analysis of the three structures also demonstrates that the volume measures are only weakly discriminating between preclinical and control groups, with the average atrophy rates of the volume of the entorhinal cortex higher than amygdala and hippocampus. The entorhinal cortex thickness also exhibits an atrophy rate nearly a factor of two higher in the ApoE4 positive group relative to the ApoE4 negative group providing weak discrimination between the two groups.

Original languageEnglish (US)
Pages (from-to)352-360
Number of pages9
JournalNeuroImage: Clinical
StatePublished - 2013

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Neurology
  • Clinical Neurology
  • Cognitive Neuroscience


Dive into the research topics of 'The diffeomorphometry of temporal lobe structures in preclinical Alzheimer's disease'. Together they form a unique fingerprint.

Cite this