Hippocampal Shape in Schizophrenia: A Deformation-Based Morphometric Analysis

Andrew J. Saykin, Laura A. Flashman, Li Shen, John Ashburner, Molly Sparling, Annette Donnelly, Fillia Makedon, David Isecke, James C. Ford, Vasileios Megalooikonomou, Thomas W. McAllister

Brain Imaging Laboratory, Departments of Psychiatry & Radiology, Dartmouth Medical School, Lebanon, NH 03756 USA
New Hampshire Hospital, Concord, NH USA
Department of Computer Science, Dartmouth College, Hanover, NH 03755 USA
Functional Imaging Laboratory, Wellcome Department of Cognitive Neurology, London, UK
¶Department of Computer and Information Sciences, Temple University, Philadelphia, PA 19122 USA

Abstract
Introduction:
Volumetric MR studies in schizophrenia indicate 4-5% reduction of the hippocampus relative to controls (1). Morphological analysis of shape characteristics may provide greater specificity regarding hippocampal changes and could potentially detect abnormalities in the absence of volume differences. Recently developed methods apply warping and deformation field analysis to brain landmarks and region of interest (ROI) data (2-5), and these methods have been used for analysis of hippocampal shape differences in schizophrenia (6). Here, we contrasted shape information between patients and controls, controlling for volume, and examined several approaches for shape characterization.

Method:
Participants were 39 medicated patients meeting DSM-IV criteria for schizophrenia or schizoaffective disorder and 24 healthy controls. Volumetric data was acquired on a 1.5T GE scanner as a T1-weighted SPGR coronal series (TR=24, TE=8, flip=40, NEX=1, slices= 124 x 1.5mm, no skip, in-plane resolution= .9375 mm², FOV=24cm). ROIs were manually segmented based on a standardized approach (ICC reliabilities >.9). Tracing was performed using BRAINS (7) after realignment to the plane of the hippocampus and resampling to isotropic 1mm voxels. Traces of the hippocampus were rendered as 3D surface models and transformed to solid ROI object maps using Matlab-6. A symmetrical hippocampal template was constructed from 24 registered hippocampi from 12 healthy controls (Fig 1). Preprocessing of all ROI maps included smoothing (FWHM=3mm), and rigid body registration plus scaling to match the volume of the template. Shape differences were examined with several methods including DBM (2). High dimensional warping was used to obtain deformation fields (x,y,z) mapping points between each standardized ROI map and the template. Jacobian determinant maps reflecting local shape differences (stretching, shearing and rotation) were computed from the deformations.

Results:
Fig 2 shows the result of a linear model analysis comparing groups on the Jacobian determinant maps using SPM99. The between group contrast images for the Jacobian maps indicated local shape differences in the head and particularly posterior tail region (Fig 3) suggesting differential curvature. The Chi-square method detected a similar pattern of local differences. Although the group mean images were very similar, inspection of individual hippocampal ROIs indicated considerable shape variability across individuals, particularly within the patient group.

Conclusions:
Curvature and orientation of the posterior hippocampus appears to be a promising variable for further investigation, including skeletonization approaches. Additional ongoing work is addressing the characterization, classification and mining of individual differences in shape features that could potentially be useful for diagnostic and treatment monitoring purposes. This heterogeneity may also be related to brain activity patterns and could be used to examine structure-function relationships.

We thank NARSAD, the Ira DeCamp Foundation & NH Hospital for support.

References:
1. M.D. Nelson, A.J. Saykin, L.A. Flashman et al. Arch Gen Psychiat 55, 433-440 (1998).
2. J. Ashburner, et al., Human Brain Mapping 9, 212-225 (2000).
3. F.L. Bookstein, Medical Image Analysis 1, 225-243 (1997).
4. C. Davatzikos et al., J Comput Assist Tomogr 20, 88-97 (1996).
5. A.W. Toga, Brain warping (Academic Press, 1998).
6. J. Csernansky et al., PNAS 95, 11406-11411 (1998).
7. N.C. Andreasen, et al., J Neuropsych Clin Neurosci 4, 125-133 (1992).