Cortical intercorrelations of frontal area volumes in schizophrenia Serge A. Mitelman, * Monte S. Buchsbaum, Adam M. Brickman, and Lina Shihabuddin Department of Psychiatry, Neuroscience-PET Laboratory, Box 1505, Mount Sinai Medical Center, One Gustave L. Levy Place, New York, NY 10029, USA Received 6 July 2004; revised 3 May 2005; accepted 5 May 2005 Available online 28 June 2005 Background: Abnormal regional volume intercorrelations between selected cortical areas in schizophrenia patients were previously reported in several MRI studies. Methods: A detailed analysis of frontal gray and white matter volume correlations with Brodmann’s area volumes in the rest of the cortex was undertaken in normal subjects (n = 42) and patients with schizophrenia (n = 106), divided into good-outcome (n = 52) and poor-outcome (Kraepelinian; n = 54) subtypes. Results: Frontal gray matter volumes were correlated with temporal lobe volumes in schizophrenics but not in normal subjects. Some frontal – parietal and frontal – occipital correlations showed a similar pattern. In comparison to normal subjects, schizophrenia patients showed weaker or absent intercorrelations intrafrontally, specifically between left motor– premotor and eye-movement areas (4, 6, 8) and dorsolateral area 44, as well as between left areas 9 and 46 vs. area 24 (cingulate gyrus). Poor outcome among patients with schizophrenia was associated with weaker correlations between left frontal area 9 and both medial and lateral temporal cortices, as compared to normal subjects or good-outcome patients. Conclusions: There appears to be a structural component in the task or symptom-related dysfunctional interactions between the frontal and more posterior cortical regions with preferential pathological involve- ment of frontotemporal and more limited involvement of frontoparietal and fronto-occipital systems in schizophrenia. Impaired regional associations within the frontal lobe, between left motor – premotor and Broca’s areas, may play a role in language processing deficits in schizophrenia, while frontocingulate disconnection may result in working memory disturbances. Poor outcome may be associated with more widespread disconnections between prefrontal vs. cingulate and temporal regions in the left hemisphere, consistent with a disruption along the course of the left cingulum or uncinate bundles. D 2005 Elsevier Inc. All rights reserved. Keywords: MRI; Schizophrenia; Volume correlations; Frontal lobe; Brodmann; Poor outcome Introduction Disrupted cortical connectivity has been hypothesized as the pathophysiologic mechanism of cognitive dysfunction in schizo- phrenia, described variously as reticulo-frontal disconnection syndrome (Goldberg et al., 1989), psychosis circuitry (Bunney et al., 1995), dysplastic net (Bullmore et al., 1997), and devel- opmental reduction of synaptic connectivity (McGlashan and Hoffman, 2000). Functional rather than structural (Friston, 1998), specifically frontotemporal disconnection has usually been empha- sized (Goodman, 1989; Friston and Frith, 1995; Weinberger and Lipska, 1995; Frith, 1996). Indeed, several electrophysiological (Peled et al., 2001; Winterer et al., 2003; Pae et al., 2003) and neuroimaging studies showed reduced frontotemporal functional connectivity associated with specific cognitive tasks (Katz et al., 1996; Ragland et al., 1998, 2001; Meyer-Lindenberg et al., 2001), verbal productions (Ford et al., 2002), smooth pursuit eye- movements (Tregellas et al., 2004), active disease state (Erkwoh et al., 1999), or active hallucinatory experiences (Lawrie et al., 2002) in patients with schizophrenia. In an EEG coherence study, Winterer et al. (2003) reported a dynamic, task-dependent frontotemporal relationship with both increased and reduced effective connectivity as a function of information-processing demands. Pae et al. (2003), using low-resolution electromagnetic tomography (LORETA), found reduced current density of auditory P300 generators in schizophrenia patients (as compared to normal controls) in the left medial temporal lobe and at the parietotemporal junction, whereas the current density was relatively increased in the prefrontal and orbitofrontal cortex. Reduced frontoparietal (Paulus et al., 2002; Kim et al., 2003) and dorsolateral prefrontal – anterior cingulate (Spence et al., 2000) functional connectivity has also been reported, with a proposed failure of anterior cingulate modulatory role in the frontotemporal integration in schizophrenia (Fletcher et al., 1999). It has been argued that structural disconnection may underlie some of these functional dissociations in schizophrenia (McGuire and Frith, 1996; Woodruff et al., 1997; Gaser et al., 2004) and reduced fractional anisotropy in the left uncinate and arcuate fasciculi, suggestive of frontotemporal and frontoparietal disconnections, was demonstrated in diffusion-tensor imaging studies (Kubicki et al., 2002; Burns et al., 2003). In addition, frontotemporal dementias may sometimes clinically 1053-8119/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2005.05.024 * Corresponding author. Fax: +1 212 423 0819. E-mail address: serge.mitelman@mssm.edu (S.A. Mitelman). Available online on ScienceDirect (www.sciencedirect.com). www.elsevier.com/locate/ynimg NeuroImage 27 (2005) 753 – 770