161. Frontostriatal Miswiring in Schizophrenia

• Published in: Schizophrenia bulletin Vol. 43; no. suppl_1; p. S82
• Main Authors: Levitt, James, Kubicki, Marek, McCarley, Robert, Shenton, Martha, Rathi, Yogesh
• Format: Journal Article
• Language: English
• Published: US Oxford University Press 01.03.2017
• Subjects: Abstracts
• ISSN: 0586-7614; 1745-1701
• DOI: 10.1093/schbul/sbx021.219

Background: Schizophrenia may be based on abnormal brain connectivity. We assessed the frontostriatal network which can be parsed into functionally segregated, parallel and functionally integrative, converging white matter pathways. It has been proposed that segregated tracts allow for refining skills, whereas integrative tracts allow for new, reward-based learning. We developed a novel method, using MR diffusion weighted imaging (MR DWI) tractography, enabling us to label the striatal surface area into 2 types; that receiving segregated and that receiving integrative pathways. Methods: We used MR DWI tractography to calculate frontostriatal pathway streamline counts, an estimate of fiber counts, and FA between the cortex and striatum in 27 chronic schizophrenia (CSZ) and 26 matched healthy control (HC) subjects. All images were acquired on a 3T Siemens scanner. ROIs were acquired using structural MRI and then registered to the MR DWI images. Using 2-tensor tractography, we extracted streamlines connecting 4 distinct frontal regions of interest (ROIs) with the striatum (limbic cortex, 2 associative cortex subregions [dorsolateral and ventrolateral prefrontal cortex], and sensorimotor cortex). For each surface voxel on the striatum, we thus obtained streamline counts of 4 tracts projecting from 4 distinct frontal ROI origins with their destination endpoints on the surface voxels of the striatum. We chose a threshold proportion of 0.7 such that if a given surface voxel exceeded the threshold from a single cortical ROI source, it was labeled as segregated; if not, it was labeled as integrative. Results: First, mixed model ANOVA showed a group difference for total striatal surface voxel number ( P  = .045). F/u t-tests showed that CSZs had fewer integrative surface voxels in the left ( P  = .007), but not right, hemisphere ( P  = .2). More specifically, it was the integrative voxels in the left hemisphere (LH) Limbic ( P  = .007) and LH Associative ( P  = .01) striatal subregions that were reduced in CSZ. Second, using a mixed model ANOVA, we examined the relative amount of striatal surface area receiving segregated vs integrative input, i.e., a relative segregated vs integrative (RSI) surface area quotient. This showed a significant group by hemisphere interaction for RSI ( P  = .04). F/u t-tests showed a left, but not right, hemisphere increase in RSI quotient in CSZ versus HCs ( P  = .006; P  = .5). Conclusion: First, CSZs had significantly fewer integrative surface voxels in the left, but not in the right, hemisphere, driven by group differences in LH Limbic and associative striatal subregions, which is of interest as such voxels serve an integrative function. Second, CSZs had significantly higher RSI quotient scores in the left, but not right, hemisphere indicating an imbalance in the ratio of segregated to integrative surface areas on the striatum, lateralized to left hemisphere. These findings support brain miswiring of the frontostriatal network in CSZ.