Hippocampal Volume and Shape Analysis in an Older Adult Population

Abstract

This report presents a manual segmentation protocol for the hippocampus that yields a reliable and comprehensive measure of volume, a goal that has proven difficult with prior methods. Key features of this method include alignment of the images in the long axis of the hippocampus and the use of a three-dimensional image visualization function to disambiguate anterior and posterior hippocampal boundaries. We describe procedures for hippocampal volumetry and shape analysis, provide inter- and intra-rater reliability data, and examine correlates of hippocampal volume in a sample of healthy older adults. Participants were 40 healthy older adults with no significant cognitive complaints, no evidence of mild cognitive impairment or dementia, and no other neurological or psychiatric disorder. Using a 1.5 T GE Signa scanner, three-dimensional spoiled gradient recalled acquisition in a steady state (SPGR) sequences were acquired for each participant. Images were resampled into 1 mm isotropic voxels, and realigned along the interhemispheric fissure in the axial and coronal planes, and the long axis of the hippocampus in the sagittal plane. Using the BRAINS program (Andreasen et al., 1993 Andreasen , N. C. , Cizadlo , T. , Harris , G. , Swayze , V. , O'Leary , D. S. , et al. . ( 1993 ). Voxel processing techniques for the antemortem study of neuroanatomy and neuropathology using magnetic resonance imaging . Journal of Neuropsychiatry , 5 , 121 – 130 . [Google Scholar] ), the boundaries of the hippocampus were visualized in the three orthogonal views, and boundary demarcations were transferred to the coronal plane for tracing. Hippocampal volumes were calculated after adjusting for intracranial volume (ICV). Intra- and inter-rater reliabilities, measured using the intraclass correlation coefficient, exceeded .94 for both the left and right hippocampus. Total ICV-adjusted volumes were 3.48 (±0.43) cc for the left hippocampus and 3.68 (±0.42) for the right. There were no significant hippocampal volume differences between males and females (p > .05). In addition to providing a comprehensive volumetric measurement of the hippocampus, the refinements included in our tracing protocol permit analysis of changes in hippocampal shape. Shape analyses may yield novel information about structural brain changes in aging and dementia that are not reflected in volumetric measurements alone. These and other novel directions in research on hippocampal function and dysfunction will be facilitated by the use of reliable, comprehensive, and consistent segmentation and measurement methods.