DOI: 10.1002/alz.079485 ISSN: 1552-5260

Antemortem MRI detects atrophy related to hippocampal sclerosis pathology in the oldest‐old but side matters: The 90+ Study

Davis C. Woodworth, Janice X. Li, Sina Faramarzi, Dana E. Greenia, Syed A. Bukhari, Thomas J. Montine, Claudia H. Kawas, María M. M. Corrada, S. Ahmad Sajjadi
  • Psychiatry and Mental health
  • Cellular and Molecular Neuroscience
  • Geriatrics and Gerontology
  • Neurology (clinical)
  • Developmental Neuroscience
  • Health Policy
  • Epidemiology



Hippocampal sclerosis of aging (HS) is the neuropathological finding of neuronal death and gliosis in the hippocampus. HS is more prevalent, and associated with dementia, in the oldest‐old (those ≥90 years), but cannot be diagnosed during life. HS is often unilateral and can be missed if only one brain hemisphere is assessed, which is standard neuropathological practice but may hamper the development of biomarkers for HS. Here we examined hippocampal volumes from antemortem MRI in relation to bilaterally‐assessed HS at autopsy, accounting for side, in oldest‐old participants from The 90+ Study.


We included participants with both autopsy and antemortem MRI data. All participants had neuropathological assessment of HS for both hemispheres. We performed hippocampal segmentation of antemortem 3D‐T1w MRIs using FreeSurfer, adjusting by intracranial volume. We compared total average (average of left and right), left, and right, hippocampal volumes between those with and those without HS. For HS participants we combined the hippocampal volumes from each hemisphere with HS into a single group and from each hemisphere without HS into a separate group, and compared these to the average hippocampal volumes from those without HS. We used Welch’s T‐tests for comparisons and report Cohen’s d effect sizes.


Participant characteristics (N = 89) are shown in Table‐1. All HS was unilateral (Right = 6, Left = 4). Average total hippocampal volume was not significantly different between those with and those without HS (d = 0.51, P = 0.11, Figure‐1a). However, right hippocampal volumes were significantly related to right‐sided HS with a large effect size (P = 0.002, d = 1.02) and left hippocampal volumes had a large effect size for left‐sided HS (P = 0.17, d = 0.90, Figure‐1c), whereas this was not the case when HS was on the opposite side (P>0.3, d<0.4, Figure‐1b,c). Combined HS participant hippocampal volumes for the side with HS (d = 0.98, P = 0.003), but not the side without HS (d = ‐0.14, P = 0.7), were different from volumes in participants without HS (Figure‐1d).


Our results suggest HS is detectable by atrophy on antemortem MRI, but most strongly when using hippocampal volumes from the same side as HS. Researchers should consider assessing and documenting the presence of HS in both hemispheres to accelerate development of neuroimaging biomarkers for HS.

More from our Archive