A Probabilistic Atlas of Cortical Gray Matter Changes in Monozygotic Twins Discordant for Schizophrenia

7th International Conference on Functional Mapping of the Human Brain, Brighton, England, June 2001.

Introduction:

Cortical gray matter reduction is a robust correlate of schizophrenia, but whether this loss is uniformly distributed or follows a preferential topography remains unclear. Individual differences in the patterning and shape of sulci and gyri in the cerebral cortex may obscure disease-related regional differences in gray matter, particularly when volumes are derived from ROIs using cortical landmarks, as in nearly all prior studies of schizophrenia. In this study we utilized a recently created algorithm for cortical pattern matching [1] to develop a probabilistic atlas of gray matter density differences between patients with schizophrenia and their non-ill monozygotic (MZ) co-twins. The pattern matching procedure ensured that regional variation in gray matter was assessed in the same anatomical reference locations across subjects. Comparison of MZ twins discordant for schizophrenia enabled isolation of the disease-specific differences in gray matter by eliminating variation due to random genetic influences [2].

Methods:

High resolution 3D (MPRAGE) MRI scans were performed on 10 pairs of MZ twins discordant for schizophrenia (5 male, 5 female pairs), and 10 pairs of MZ twins without schizophrenia or a related diagnosis, who were matched to the index pairs on age (48.2+/-3.4 years), gender, handedness, and parental social class. Trained operators drew 36 major gyral/sulcal boundaries on each subject's MRI scan, and variability in gyral/sulcal patterning and shape was then modeled using elastic deformation mappings to the control twins' average configuration. Maps of gray matter differences associated with schizophrenia were generated within each index pair and were subsequently elastically realigned for averaging across the 10 pairs, with voxel-wise computation of probability for deviation from the null distribution. To control for multiple comparisons, permutation testing was used to determine whether the gray matter differences within the primary regions affected in the voxel-wise map deviated from the expected distribution based on simulations involving random pairings of unrelated subjects.

Results:

The averaged within-pair difference map revealed a regionally-specific pattern of gray matter loss associated with schizophrenia, with bilateral reductions of cortical thickness in the range of 5 to 8% in the dorsolateral prefrontal cortex (BA46/9), superior parietal cortex, and superior temporal gyrus. A nearly identical regional distribution was evident in the probability map thresholded at a voxel-wise alpha of p<.05. In permutation tests conducted to control for multiple comparisons, the gray matter reductions in the prefrontal and parietal areas were found to be significant (p<.05) bilaterally, but that in the superior temporal gyrus was significant only in the left hemisphere. Notably, there were no differences in gray matter between discordant co-twins in terms of primary sensory and motor areas.

Conclusion:

This study is the first to apply a pattern matching procedure and probabilistic atlasing approach to anatomical characterization of the gray matter changes that differentiate schizophrenia patients from their own non-ill monozygotic co-twins. The results indicate that the disease phenotype is associated with a regional pattern of gray matter loss that impacts heteromodal association areas in the prefrontal, temporal, and parietal lobe regions preferentially, with a sparing of primary sensory and motor areas.

References: [1]: Thompson PM et al., Cerebral Cortex 11:1-16(2001). [2] Cannon TD et al., Am J Hum Gen 67:369-382(2000).

Related Publications

(back to main list)

Contact Information