Reduced white matter integrity correlated with cortico–subcortical gray matter deficits in schizophrenia

Schizophrenia Research 111 (2009) 78–85

Contents lists available at ScienceDirect

Schizophrenia Research j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / s c h r e s

Reduced white matter integrity correlated with cortico–subcortical gray matter deficits in schizophrenia Jun Miyata a,⁎, Kazuyuki Hirao a,1, Chihiro Namiki a,b, Hironobu Fujiwara a, Mitsuaki Shimizu a, Hidenao Fukuyama b, Nobukatsu Sawamoto b, Takuji Hayashi a, Toshiya Murai a a b

Department of Neuropsychiatry, Graduate School of Medicine, Kyoto University, Shogoin-Kawaharacho 54, Kyoto 606-8507, Japan Human Brain Research Center, Graduate School of Medicine, Kyoto University, Shogoin-Kawaharacho 54, Kyoto 606-8507, Japan

a r t i c l e

i n f o

Article history: Received 2 December 2008 Received in revised form 3 March 2009 Accepted 5 March 2009 Available online 10 April 2009 Keywords: Basal ganglia–thalamocortical circuits Diffusion tensor imaging Disconnection hypothesis Language processing networks Tract-based spatial statistics Voxel-based morphometry

a b s t r a c t Background: The pathology of schizophrenia is thought to involve multiple brain regions and the connections among them. Although a number of MRI studies have demonstrated gray matter reductions and abnormal white matter integrity in schizophrenia, to date no study has investigated their association in the whole brain. Methods: Twenty-seven schizophrenia patients and 33 healthy controls were recruited. Voxelwise group comparison of white matter fractional anisotropy (FA) was performed using tract-based spatial statistics (TBSS). Comparison of gray matter concentration (GMC) was performed using voxel-based morphometry (VBM). Voxelwise correlational analyses were performed for patients inside a significant GMC reduction mask created by VBM, using simple regression models with mean FA values of each significant TBSS cluster as explanatory variables. Results: TBSS revealed FA reduction in left prefrontal and occipital regions in the patients. Mean FA values of both areas revealed significant correlation with gray matter reduction in multiple cortical and subcortical areas, with overlapping but different patterns. Conclusions: Voxelwise correlational analysis of white and gray matter pathology, as performed here, further elucidated the pathophysiology of schizophrenia, and provided a novel view of the “disconnection hypothesis” of schizophrenia. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Volumetric magnetic resonance imaging (MRI) studies on patients with schizophrenia have demonstrated gray matter reductions in several cortical and subcortical areas, including the prefrontal cortex, the superior temporal gyrus (STG), medial temporal lobe structures, the thalamus, and basal ganglia (Shenton et al., 2001; Suzuki et al., 2005). Considering this widespread pathology, it is reasonable to hypothesize that the nature of this disorder resides in the disrupted

⁎ Corresponding author. Tel.: +81 75 751 3386; fax: +81 75 751 3246. E-mail address: [email protected] (J. Miyata). 1 Present address: Section of Neuroimaging, Department of Psychological Medicine and Psychiatry, Institute of Psychiatry, King's College London, PO67 De Crespigny Park London SE5 8AF, UK. 0920-9964/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2009.03.010

connectivity of the networks among these cortical and subcortical gray matters. Such a “disconnection hypothesis” (Friston, 1998) of schizophrenia was originally shown to be “functional” disconnectivity (Frith et al., 1995), and anatomical substrates of such disconnectivity were postulated to exist at the synaptic level in gray matter (McGlashan and Hoffman, 2000; Selemon and Goldman-Rakic, 1999), rather than in the white matter tracts that wire distant gray matter regions. Diffusion tensor imaging (DTI) provides information about white matter tracts and their organization based on water diffusion (Basser et al., 1994). DTI is thought to be more sensitive to subtle abnormalities in white matter than anatomical MRI. Fractional anisotropy (FA) is the most commonly used index and its reduction implies decreased white matter tract integrity. DTI studies on schizophrenia that examined FA in defined regions of interest, though not

J. Miyata et al. / Schizophrenia Research 111 (2009) 78–85

completely consistent, reveal subtle FA reductions in several areas including frontal white matter, the corpus callosum (CC), and the cingulum bundle (Kanaan et al., 2005; Kubicki et al., 2007; Walterfang et al., 2006). These findings suggest that regional white matter pathologies are also anatomical substrates in the pathology of connectivity in schizophrenia. As these gray and white matter abnormalities are subtle and multiregional, it is necessary to explore such abnormalities in the whole brain. Voxel-based morphometry (VBM; Ashburner and Friston, 2000) is a fully automated structural MRI analysis method that can reveal the patterns of regional gray matter volume alterations at the whole brain level, without any specific hypotheses about search areas. Consistent findings have been reported in VBM studies on schizophrenia (Honea et al., 2005). With respect to white matter pathologies, VBM style analysis has also been applied to DTI data (Buchsbaum et al., 2006; Kyriakopoulos et al., 2008), and the results show FA reductions in diverse areas. However, the methodological pitfalls of VBM style analysis seem to be more problematic when it is applied to DTI data; misregistration of white matter tracts between subjects can mistakenly be interpreted as a FA difference in the same tract. A smoothing procedure is necessary to compensate for such misregistration; however, the smoothing kernel sizes substantially affect the results of FA analyses (Jones et al., 2005). A recently developed technique, tract-based spatial statistics (TBSS; Smith et al., 2006), maps each subject's DTI data on a common white matter tract center (‘skeleton’) and does not need smoothing. This technique is considered more robust and suitable for whole brain DTI data analysis. As gray and white matter both constitute networks, with gray matter as nodes and white matter as wiring, their abnormalities in schizophrenia may be associated with each other. To date, only a limited number of studies have attempted to elucidate the covariation of gray and white matter abnormalities: Hulshoff Pol et al. (2001) used VBM to demonstrate gray matter volume reductions in several cortical and subcortical areas and volume increases in the caudate and globus pallidus in schizophrenia. They also applied VBM to white matter in the same sample, and revealed white matter volume reductions in the CC, the internal capsule, and the anterior commissure in schizophrenia (Hulshoff Pol et al., 2004). Moreover, they revealed correlations between gray matter volume changes and white matter volume reductions. Among the DTI studies, Douaud et al. (2007), using VBM for gray matter and TBSS for DTI data, superimposed both results, visualizing the anatomical association between regional gray matter reduction and FA reduction at the whole brain level. However, to date no study has examined the covariation between gray matter reduction and DTI abnormality in the whole brain using voxelwise correlational analysis. The present study aimed to elucidate the association between gray matter volume reductions and white matter FA abnormalities by using voxelwise correlational analysis in the whole brain. We hypothesized that regional FA reductions in the patients would have positive correlations with regional gray matter reductions. Considering the network pathology of schizophrenia, as predicted by the “disconnection hypothesis”, we expected such associations not only between gray and white matters that are located adjacent to each other, but also between those areas that are distant from each other.

79

2. Materials and methods 2.1. Subjects Twenty-seven schizophrenia patients, diagnosed by the patient edition of the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID; First et al., 1996), were studied. None of the patients had comorbid psychiatric disorders. Antipsychotic medication was prescribed to all patients. The Positive and Negative Syndrome Scale (PANSS; Kay et al., 1987) was used to assess the severity of clinical symptoms. Thirty-three healthy controls matched by age, sex, handedness and education levels with the patient group were recruited. The controls had no history of psychiatric illness as determined by using the non-patient edition of the SCID (First et al., 1998). There was no history of psychotic disorders among their first-degree relatives. Exclusion criteria for all individuals included a history of head trauma, neurological illness, serious medical or surgical illness and substance abuse. After receiving a complete description of the study, all participants gave written informed consent. The study design was approved by the Committee on Medical Ethics of Kyoto University. 2.2. MRI acquisition The DTI data were acquired using single-shot spin-echo echo-planar sequences and structural MRI data using three dimensional magnetization prepared rapid gradient echo (3DMPRAGE) sequences, on a 3.0-T MRI unit (Trio; Siemens, Erlangen, Germany) with a 40 mT/m gradient. A generalized autocalibrating partially parallel acquisition algorithm was applied for parallel imaging using a reduction factor of two, 24 additional autocalibrating phase-encoding steps in the center of k-space, and a 75% partial Fourier technique in the phaseencoding direction. Parameters for the DTI were as follows; TE 79 ms, TR 5200 ms, 128 × 128 matrix, FOV 220 × 220 mm, 40 contiguous axial slices of 3.0 mm thickness, 12 noncolinear axis motion probing gradient, b = 700 s/mm2. To enhance the signal-to-noise ratio, imaging was repeated four times. Parameters for the 3D-MPRAGE imaging were as follows; TE 4.38 ms, TR 2000 ms, inversion time 990 ms, 256 × 256 matrix, FOV 240 × 240 mm, 208 axial slices of 1.0 mm thickness. Axial slices were adjusted to be parallel to the anterior commissure– posterior commissure line in each subject. 2.3. Imaging data processing and statistical analysis 2.3.1. TBSS analysis of white matter DTI data processing was performed using FSL 3.3 (http:// www.fmrib.ox.ac.uk/fsl). All DTI source data were corrected for eddy currents and head motion, by registering each data to the first b = 0 image of the first repetition, with affine transformation. Averaged data of these 4 repetitions were created for each subject. FA maps were calculated by FSL DTIFit. The TBSS program, part of the FSL program, was used for voxelwise statistical analysis of FA data. Briefly, all subjects' FA maps were first co-registered using non-linear registration (Rueckert et al., 1999) to the most ‘typical’ subject's FA map, which minimized the total amount of warping required for all the subjects. This target FA map was affine-transformed into

80

J. Miyata et al. / Schizophrenia Research 111 (2009) 78–85

the 1 × 1 × 1 mm Montréal Neurological Institute (MNI) 152 space. For each of all the other subjects, non-linear and affine transformations were combined and applied to the original FA map, resulting in a standard-space version FA map. All transformed FA images were averaged to create a mean FA image, which was then thinned (skeletonized) to create a mean FA skeleton, taking only the centers of white matter tracts. The skeleton was thresholded by FA ≥ 0.20 to ensure that gray matter regions were excluded from the analyses. Voxel values of each subject's FA map were projected onto the skeleton by searching the local maxima along the perpendicular direction from the skeleton. Statistical analysis was undertaken using skeletonized FA data. As a smoothing procedure was not applied to the TBSS data, voxelwise permutation-based nonparametric inference (Nichols and Holmes, 2001) was performed using the FSL randomize program, which does not rely on Gaussian distribution of data. For group comparisons, an analysis of covariance (ANCOVA) design was used, with age and gender as nuisance covariates. Both covariates were centered (demeaned), fed into a nuisance design matrix and regressed out of the data before implementing the permutation tests. Both control-patient and patient-control contrasts were tested with 5000 permutations. Multiple comparisons were corrected with a cluster-forming threshold of t = 3 and a cluster-wise significance level of p b 0.05. The fiber tracts corresponding to the clusters were identified with reference to the Johns Hopkins University DTI-based White Matter Atlas (http://cmrm.med.jhmi.edu; Mori et al., 2005). The mean FA values within each significant cluster were calculated for each patient, using the command-line utilities of FSL (AVWUTILS), and used for correlational analyses in Section 2.3.3. 2.3.2. VBM analysis of gray matter Structural MRI data were processed and analyzed using statistical parametric mapping 2 (SPM2; http://www.fil.ion. ucl.ac.uk/spm) software, running on Matlab 6.5.2 (The MathWorks, Natick, MA, U.S.A.). The optimized VBM method was performed for voxelwise group comparison of gray matter reduction; as described in detail by Ashburner and Friston (2000) and Good et al. (2001). We used an extension of SPM, the VBM tools written by Christian Gaser (http:// dbm.neuro.uni-jena.de/vbm). Briefly, a study-specific whole brain template and gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) prior images were created from all subjects. Using these customized template and prior images, each participant's original brain image was spatially normalized, segmented into GM, WM, and CSF, and resampled at a resolution of 1 × 1 × 1 mm, according to the optimized protocol. In this study, we used only unmodulated data, i.e. Jacobian modulation was not applied. This was because in our earlier study, which had a sample population with similar characteristics, we found a significant group difference with unmodulated data, not with modulated data (Yamada et al., 2007). The GM images were smoothed with a Gaussian kernel of 12 mm full width at half maximum. To identify the brain area where patients showed reductions in gray matter concentration (GMC) relative to controls, an ANCOVA design was used for voxelwise parametric analysis, with age, gender, and global brain volume as nuisance covariates. The statistical significance level was thresholded

for correction of multiple comparisons using a false discovery rate (FDR; Genovese et al., 2002) of 0.05. Small clusters were excluded using an extent threshold of 800 contiguous voxels, which was approximately the size of a 12-mm-diameter sphere. The result of this group comparison was used as an inclusion mask for the following voxelwise correlational analyses. 2.3.3. Correlational analyses between gray and white matter To investigate the correlation between FA and GMC abnormalities, voxelwise simple regression analyses were performed for schizophrenia patients using SPM2. As the initial exploration of the mean FA values revealed them to be normally distributed, parametric statistical analyses were used. Design matrices were created for each TBSS cluster, modeling the mean FA value as the explanatory variable. The above mentioned mask was used to include only those voxels with significant group differences. The statistical significance level was thresholded for correction of multiple comparisons by an FDR of 0.05. Small clusters were excluded using an extent threshold of 800 contiguous voxels. 3. Results 3.1. Demographic and clinical data Demographic data and clinical data are shown in Table 1. The patient group consisted mainly of chronic patients with relatively mild symptoms. Most were taking atypical antipsychotics. 3.2. Imaging data The TBSS analysis revealed two significant clusters of FA reduction in patients compared with controls (Fig. 1). One

Table 1 Demographic, clinical, and neuropsychological characteristics of the participants.

Age (years) Sex (male/female) Handedness (right/left) Education (years) Age of onset (years) Duration of illness (years) Medication (mg/day, HPD equivalent) a Atypical/typical/both b PANSS c total score Positive scale Negative scale General psychopathology scale

Schizophrenia (n = 27)

Control (n = 33)

Statistics

Mean

S.D.

Mean

S.D.

p

38.5 14/13 26/1 14.1 25.1 13.9 12.0

8.9

7.8

2.6 6.1 9.6 8.3

37.1 16/17 32/1 14.2 – – –

2.8 – – –

N.S. N.S. N.S. N.S. – – –

17 / 4 / 6 68.5 16.3 18.4 33.7

18.9 6.3 7.6 9.8

– – – – –

– – – – –

– – – – –

a Haloperidol (HPD) equivalents were calculated according to the practice guidelines for the treatment of patients with schizophrenia (Inagaki and Inada, 2006; Lehman et al., 2004). b Atypical = patients who were taking atypical antipsychotics. Typical = patients taking typical antipsychotics. Both = patients taking both typical and atypical antipsychotics. c PANSS = Positive and Negative Syndrome Scale (Kay et al., 1987).

J. Miyata et al. / Schizophrenia Research 111 (2009) 78–85

Fig. 1. The tract-based spatial statistics (TBSS) group comparison of white matter Fractional anisotropy (FA). Clusters of significant FA reduction in patients are shown on the mean FA map (p b 0.05, cluster level correction at t = 3): a = axial slice, b = sagittal slice, c = coronal slice. The left prefrontal cluster (blue) and left occipital cluster (red) are overlaid on the FA skeleton (yellow). Left and right orientation is displayed according to neurological convention, to preserve consistency with results from the statistical parametric mapping (SPM) 2 program.

cluster was located in the left prefrontal region adjacent to the minor forceps of the CC. By reference to the Johns Hopkins University white matter atlas (Mori et al., 2005), this cluster was identified to contain callosal inter-hemispheric fibers, fibers from the anterior cingulate cortex (ACC), the anterior thalamic radiation, and a small part of the corticopontine tract. The other cluster was located in the left occipital region adjacent to the splenium of the CC, which, according to the atlas, contains the inferior longitudinal fasciculus (ILF), the posterior thalamic radiation, the inferior occipitofrontal fasciculus (IOFF), and the corticopontine tract. There were no significant clusters of FA increase in patients compared with controls.

81

Fig. 2 shows the result of VBM group comparisons. A significant GMC reduction in patients was revealed, for the most part, in frontal and temporal lobes and in subcortical areas, including bilateral dorsolateral prefrontal cortex (DLPFC), ventrolateral prefrontal cortex (VLPFC), medial prefrontal cortex (MPFC), ACC, orbitofrontal cortex (OFC), STG, middle temporal gyrus (MTG), medial temporal lobe structures, insula, striatum, and thalamus. The results of FA and GMC correlational analyses are shown in Table 2. MNI coordinates were transformed into Talairach coordinates (Talairach and Tournoux, 1988) using the mni2tal.m Matlab script written by Matthew Brett (http://imaging.mrc-cbu.cam.ac.uk/imaging/MniTalairach). Mean FA of the left prefrontal cluster revealed significant positive correlations with GMC in 7 areas: (1) left insula; (2) left temporal pole (TP) and inferior frontal gyrus (IFG); (3) bilateral anterior striatum, medial thalamus, subcallosal area, left entorhinal cortex, right amygdala, insula, and TP; (4) right IFG; (5) left hippocampus and parahippocampal gyrus; (6) left dorsal ACC; (7) and left inferior temporal gyrus (ITG), MTG, and STG (Fig. 3). The height threshold was t = 2.34, which corresponded to a correlation coefficient r = 0.424. Mean FA of the left occipital cluster showed significant positive correlations with GMC in 8 regions: (1) left TP, IFG, and insula; (2) left MTG, STG, and Heschl's gyrus; (3) Broca's speech area; (4) bilateral subcallosal area, caudate head, right

Table 2 Correlational analyses between mean FA of each TBSS cluster and gray matter concentration in patients. Anatomical region

BA

X Correlation with L prefrontal cluster L insula L TP, IFG 38/47 LR Ant striatum, Med 25/34/38 thalamus, subcallosal area, L entorhinal cortex, R amygdala, insula, TP R IFG 11 L hippocampus, 30 parahippocampal gyrus L dorsal ACC 24/32 L ITG, MTG, STG 20/21/22 Correlation with L occipital cluster L TP, IFG, insula 38/47 L MTG, STG, Heschl's gyrus 21/22/41 Broca's speech area 44/45 LR subcallosal area, 25/28/38 caudate head, R TP, amygdala, L entorhinal cortex LR rostral ACC 32 R insula L Med OFC 11 L pulvinar

Fig. 2. Areas of significant gray matter concentration reduction in schizophrenia patients. FDR corrected p b 0.05 (height threshold t = 2.70), extent threshold = 800 voxels.

Talairach coordinate Y

Cluster size

Z

Z

– 28 – 39 –6

10 18 5

8 –19 0

2489 3851 35254

3.93 3.88 3.8

47 – 28

49 –42

–18 4

1117 1164

3.72 3.61

–9 – 61

27 –20

22 – 17

1095 2817

3.11 3.05

–33 – 58 – 54 –2

18 –7 21 6

–24 –7 17 –8

10851 4209 1098 29560

4.35 4.11 3.98 3.72

–4 37 –1 –6

39 –22 56 –23

11 12 –13 6

1324 1046 817 982

3.61 2.96 2.8 2.79

Coordinates for the peak voxels in the clusters are displayed. The Montréal Neurological Institute (MNI) coordinates were transformed into Talairach coordinates. Abbreviations: FA = fractional anisotropy, TBSS = tract-based spatial statistics, BA = Brodmann area, L = left, R = right, TP = temporal pole, IFG = inferior frontal gyrus, Ant = anterior, Med = medial, ACC = anterior cingulate cortex, ITG = inferior temporal gyrus, MTG = middle temporal gyrus, STG = superior temporal gyrus, OFC = orbitofrontal cortex.

82

J. Miyata et al. / Schizophrenia Research 111 (2009) 78–85

TP, amygdala, and left entorhinal cortex; (5) bilateral rostral ACC; (6) right insula; (7) left medial OFC; (8) and left pulvinar (Fig. 4). The height threshold was t = 2.32, corresponding r = 0.421. The significant areas revealed by both correlational analyses showed considerable overlapping. However, the max-

imum intensity projections of Figs. 3 and 4 showed distinct correlational patterns between these two analyses: the prefrontal TBSS cluster showed strong associations, predominantly with the subcortical gray matter, while the occipital TBSS cluster strongly correlated with the left temporal lobe and its surrounding areas. These differential patterns were more evident at the more conservative threshold of FDR corrected p b 0.025 (height threshold t = 3.25 (r = 0.545) for prefrontal and t = 3.19 (r = 0.538) for occipital TBSS cluster, respectively) (Fig. 5).

4. Discussion This is, to the best of our knowledge, the first study to use voxelwise correlational analysis to directly explore the association between regional FA reduction and gray matter reduction in schizophrenia at the whole brain level. Multiple white-gray associations of abnormalities were demonstrated. Interestingly, while these associations were revealed to be spatially overlapped, the correlational patterns were distinct between the two analyses, indicating the co-occurrence of gray and white matter abnormalities within specific networks. Such associations of gray and white matter abnormalities could only be determined by applying voxelwise correlational analysis to multimodal imaging. To date, a number of studies have reported neuropathological abnormalities of schizophrenia that are thought to explain macroscopic findings of gray matter reduction, particularly a reduction of neuronal size in regional cortical gray matter (Rajkowska et al., 1998), a reduction of neuron number in the thalamus (Pakkenberg, 1990), a decrease of presynaptic markers (Glantz and Lewis, 1997) and a reduction of dendritic spine density (Garey et al., 1998). These are fairly consistent findings. On the other hand, recent neuropathological and neurogenetic studies, though not completely consistent, are suggestive of reduced size and density of oligodendrocytes and of disrupted myelination gene expression in schizophrenia (Davis et al., 2003; Walterfang et al., 2006). Our results, at the macroscopic level, can be considered to integrate these findings, and suggest the existence of possible coupled mechanisms that might generate gray and white matter pathologies in schizophrenia. Our TBSS results revealed two significant clusters of FA reduction, one in the left prefrontal white matter and the other in the left occipital region. Similar locations were reproducibly reported in earlier VBM style studies (Agartz et al., 2001; Ashtari et al., 2007). Both clusters were identified as containing interhemispheric, corticocortical, or cortico-subcortical fibers. Fig. 3. Correlational analysis between the mean FA of the left prefrontal TBSS cluster and gray matter concentration in patients. FDR corrected p b 0.05 (height threshold t = 2.34), extent threshold = 800 voxels. a: Overall distribution of gray matter areas with significant correlation with the mean FA of the left prefrontal cluster is shown by the maximum intensity projection. Top b: significant clusters in the bilateral anterior striatum, thalamus, insula, left superior temporal gyrus, and left hippocampus are displayed. Middle b: bilateral temporal pole, left inferior temporal gyrus, entorhinal cortex, and right amygdala are displayed. Bottom b: left dorsal anterior cingulate cortex and bilateral inferior frontal gyrus are displayed. Abbreviations: FA = fractional anisotropy, TBSS = tract-based spatial statistics.

J. Miyata et al. / Schizophrenia Research 111 (2009) 78–85

83

Fig. 5. Correlational patterns of both TBSS clusters with GMC are shown at FDR corrected p b 0.025 and an extent threshold of 800 voxels. a: correlation between the mean FA of the left prefrontal TBSS cluster and GMC (height threshold t = 3.25). b: correlation between the mean FA of the left occipital TBSS cluster and GMC (height threshold t = 3.19). Abbreviations: FA = fractional anisotropy, TBSS = tract-based spatial statistics, GMC = gray matter concentration.

Fig. 4. Correlational analysis between the mean FA of the left occipital TBSS cluster and gray matter concentration in patients. FDR corrected p b 0.05 (height threshold t = 2.32), extent threshold= 800 voxels. a: overall distribution of gray matter areas with significant correlation with the mean FA of the left occipital cluster is shown by the maximum intensity projection. Top b: significant clusters in the bilateral insula, left Heschl's gyrus, Broca's speech area, ACC, pulvinar, and right caudate head are displayed. Middle b: bilateral temporal pole, left entorhinal cortex, and right amygdala are displayed. Bottom b: subcallosal gyrus, rostral ACC, medial orbitofrontal cortex, and left pulvinar are displayed. Abbreviations: FA = fractional anisotropy, TBSS = tractbased spatial statistics, ACC = anterior cingulate cortex.

On the other hand, our VBM group comparisons revealed gray matter reductions, mostly in bilateral prefrontal, temporal, and subcortical areas. This pattern of gray matter reduction was basically in accordance with the results of VBM studies on schizophrenia with large sample sizes (Honea et al., 2008; Hulshoff Pol et al., 2001; Koutsouleris et al., 2008; Meda et al., 2008); thus these supported the validity of our mask. The results of two correlational analyses between FA and gray matter reduction revealed considerable overlapping across a wide range of cortical and subcortical areas; bilateral insula, TP, caudate head, subcallosal area, left IFG, entorhinal cortex, MTG, STG, pulvinar, and right amygdala, even though the two TBSS clusters were remotely located. These results indicated that there is a global covariation of pathology among widespread white and gray matter regions. However,

84

J. Miyata et al. / Schizophrenia Research 111 (2009) 78–85

such global associations may be relatively weak, as most of these overlapping areas disappeared when a higher statistical threshold was applied. The left prefrontal TBSS cluster had associations with 7 cortical and subcortical areas of gray matter reduction, with strong associations predominantly with subcortical gray matter. Of note, fibers contained in this TBSS cluster (anterior thalamic radiation and fibers from the ACC) and GMC reduction areas (the ACC, anterior striatum, and the thalamus) are components of basal ganglia-thalamocortical circuits (Alexander et al., 1990). These circuits, filtering primary sensory input and modulating cortical information processing (Ongür and Price, 2000), are considered to have a crucial role in the pathophysiology of schizophrenia (Carlsson and Carlsson, 1990). On the other hand, the left occipital TBSS cluster had associations with 8 cortical and subcortical areas of gray matter reduction, with strong associations predominantly in the left temporal lobe and its surrounding areas. Some of these structures are connected with language processing: Broca's speech area and primary auditory cortex of Heschl's gyrus are crucial components of the perisylvian language network. Recent studies also suggest that the semantic processing network could be constituted by the IOFF (Duffau et al., 2005) and the ILF (Mandonnet et al., 2007; Vigneau et al., 2006). The former connects occipital and posterior temporal cortices to the IFG and the DLPFC. The latter connects occipitotemporal regions with the TP, relayed by the uncinate fasciculus, which further connects the TP with orbitofrontal areas. Abnormal language processing has also been considered a crucial component of the underlying pathophysiology of schizophrenia (Crow, 1998). Our correlational analyses of multimodal imaging are the first to reveal the co-occurrence and covariation of gray and white matter abnormalities involved in language processing. Unfortunately, it was difficult to ascertain exactly which fiber tracts were responsible for FA reduction in each TBSS cluster, because of the relatively coarse resolution of the original DTI image and cluster level inference. Studies with finer spatial resolution and with a larger number of motion probing gradient directions, which would enable the use of probabilistic tractography (Behrens et al., 2003; Behrens et al., 2007), would be required to add more detail to our findings. In conclusion, our study, which used voxelwise correlational analysis in the whole brain, is the first to demonstrate direct associations between regional FA reduction in two white matter areas and multiple cortical/subcortical gray matter reductions in schizophrenia. The spatially overlapping, but with distinct patterns of association between gray and white matter abnormalities, supports the involvement of basal ganglia-thalamocortical circuits and language processing networks in the pathophysiology of schizophrenia. When abnormal connectivity is considered, correlational analyses at the whole brain level could be useful for elucidating the pathophysiology of schizophrenia. Role of funding source This work was supported by a grant-in-aid for scientific research from the Japan Society for the Promotion of Science and the Ministry of Education, Culture, Sports, Science and Technology, Japan [20691401 to Toshiya Murai]; a grant from the Ministry of Health, Labor and Welfare, Japan [20E-3 to Toshiya Murai]; a research grant from the Mitsubishi Pharma Research Foundation; and a research grant from the Research Group for Schizophrenia

sponsored by Astellas Pharma Inc. These agencies had no further role in the study design, the collection, analysis and interpretation of data, the writing of the report, or in the decision to submit the paper for publication. Contributors Authors Miyata and Murai designed the study and wrote the protocol, under the supervision of Author Hayashi. Author Miyata managed the literature searches and analyses, and undertook the statistical analyses. Authors Hirao, and Namiki supervised voxel based morphometry. Authors Fujiwara and Shimizu organized the recruitment and clinical assessment of the participants. Authors Fukuyama and Sawamoto supervised the MRI data acquisition and diffusion tensor processing. Author Miyata wrote the first draft of the manuscript. All authors contributed to and have approved the final manuscript. Conflict of interest All authors declare that they have no conflicts of interest.

Acknowledgments The authors wish to extend their gratitude to Makiko Yamada, Miho Yoshizumi, Teruyasu Saze, and Ryosaku Kawada for their assistance in data acquisition and processing, and, most of all, to the patients and volunteers for participating in the study.

References Agartz, I., Andersson, J.L., Skare, S., 2001. Abnormal brain white matter in schizophrenia: a diffusion tensor imaging study. Neuroreport 12, 2251–2254. Alexander, G.E., Crutcher, M.D., DeLong, M.R., 1990. Basal ganglia-thalamocortical circuits: parallel substrates for motor, oculomotor, “prefrontal” and “limbic” functions. Prog. Brain Res. 85, 119–146. Ashburner, J., Friston, K.J., 2000. Voxel-based morphometry — the methods. Neuroimage 11, 805–821. Ashtari, M., Cottone, J., Ardekani, B.A., Cervellione, K., Szeszko, P.R., Wu, J., Chen, S., Kumra, S., 2007. Disruption of white matter integrity in the inferior longitudinal fasciculus in adolescents with schizophrenia as revealed by fiber tractography. Arch. Gen. Psychiatry 64, 1270–1280. Basser, P.J., Mattiello, J., LeBihan, D., 1994. MR diffusion tensor spectroscopy and imaging. Biophys. J. 66, 259–267. Behrens, T.E., Berg, H.J., Woolrich, M.W., Smith, S.M., Wheeler-Kingshott, C.A., Boulby, P.A., Barker, G.J., Sillery, E.L., Sheehan, K., Ciccarelli, O., Thompson, A.J., Brady, J.M., Matthews, P.M., 2003. Non-invasive mapping of connections between human thalamus and cortex using diffusion imaging. Nat. Neurosci. 6, 750–757. Behrens, T.E., Berg, H.J., Jbabdi, S., Rushworth, M.F., Woolrich, M.W., 2007. Probabilistic diffusion tractography with multiple fibre orientations: what can we gain? Neuroimage 34, 144–155. Buchsbaum, M.S., Friedman, J., Buchsbaum, B.R., Chu, K.W., Hazlett, E.A., Newmark, R., Schneiderman, J.S., Torosjan, Y., Tang, C., Hof, P.R., Stewart, D., Davis, K.L., Gorman, J., 2006. Diffusion tensor imaging in schizophrenia. Biol. Psychiatry 60, 1181–1187. Carlsson, M., Carlsson, A., 1990. Schizophrenia: a subcortical neurotransmitter imbalance syndrome? Schizophr. Bull. 16, 425–432. Crow, T.J., 1998. Schizophrenia as a transcallosal misconnection syndrome. Schizophr. Res. 30, 111–114. Davis, K.L., Stewart, D.G., Friedman, J.I., Buchsbaum, M., Harvey, P.D., Hof, P.R., Buxbaum, J., Haroutunian, V., 2003. White matter changes in schizophrenia: evidence for myelin-related dysfunction. Arch. Gen. Psychiatry 60, 443–456. Douaud, G., Smith, S., Jenkinson, M., Behrens, T., Johansen-Berg, H., Vickers, J., James, S., Voets, N., Watkins, K., Matthews, P.M., James, A., 2007. Anatomically related grey and white matter abnormalities in adolescent-onset schizophrenia. Brain 130, 2375–2386. Duffau, H., Gatignol, P., Mandonnet, E., Peruzzi, P., Tzourio-Mazoyer, N., Capelle, L., 2005. New insights into the anatomo-functional connectivity of the semantic system: a study using cortico-subcortical electrostimulations. Brain 128, 797–810. First, M., Spitzer, R., Gibson, M., Williams, J., 1996. Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I). American Psychiatric Publishing, Washington D.C. First, M., Spitzer, R., Gibson, M., Williams, J., 1998. Structured Clinical Interview for DSM-IV Axis I Disorders - Non patient Edition (With Psychotic Screen). American Psychiatric Publishing, Washington D.C. Friston, K.J., 1998. The disconnection hypothesis. Schizophr. Res. 30, 115–125.

J. Miyata et al. / Schizophrenia Research 111 (2009) 78–85 Frith, C.D., Friston, K.J., Herold, S., Silbersweig, D., Fletcher, P., Cahill, C., Dolan, R.J., Frackowiak, R.S., Liddle, P.F., 1995. Regional brain activity in chronic schizophrenic patients during the performance of a verbal fluency task. Br. J. Psychiatry 167, 343–349. Garey, L.J., Ong, W.Y., Patel, T.S., Kanani, M., Davis, A., Mortimer, A.M., Barnes, T.R., Hirsch, S.R., 1998. Reduced dendritic spine density on cerebral cortical pyramidal neurons in schizophrenia. J. Neurol. Neurosurg. Psychiatry 65, 446–453. Genovese, C.R., Lazar, N.A., Nichols, T., 2002. Thresholding of statistical maps in functional neuroimaging using the false discovery rate. Neuroimage 15, 870–878. Glantz, L.A., Lewis, D.A., 1997. Reduction of synaptophysin immunoreactivity in the prefrontal cortex of subjects with schizophrenia. Regional and diagnostic specificity. Arch. Gen. Psychiatry 54, 943–952 [corrected and republished article. originally appeared in 2000. Arch. Gen. Psychiatry. 54, 660-669]. Good, C.D., Johnsrude, I.S., Ashburner, J., Henson, R.N., Friston, K.J., Frackowiak, R.S., 2001. A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage 14, 21–36. Honea, R., Crow, T.J., Passingham, D., Mackay, C.E., 2005. Regional deficits in brain volume in schizophrenia: a meta-analysis of voxel-based morphometry studies. Am. J. Psychiatry 162, 2233–2245. Honea, R.A., Meyer-Lindenberg, A., Hobbs, K.B., Pezawas, L., Mattay, V.S., Egan, M.F., Verchinski, B., Passingham, R.E., Weinberger, D.R., Callicott, J.H., 2008. Is gray matter volume an intermediate phenotype for schizophrenia? A voxel-based morphometry study of patients with schizophrenia and their healthy siblings. Biol. Psychiatry 63, 465–474. Hulshoff Pol, H.E., Schnack, H.G., Mandl, R.C., van Haren, N.E., Koning, H., Collins, D.L., Evans, A.C., Kahn, R.S., 2001. Focal gray matter density changes in schizophrenia. Arch. Gen. Psychiatry 58, 1118–1125. Hulshoff Pol, H.E., Schnack, H.G., Mandl, R.C., Cahn, W., Collins, D.L., Evans, A.C., Kahn, R.S., 2004. Focal white matter density changes in schizophrenia: reduced inter-hemispheric connectivity. Neuroimage 21, 27–35. Inagaki, A., Inada, T., 2006. Translation table of psychotropic drugs. Keio University, Tokyo. Jones, D.K., Symms, M.R., Cercignani, M., Howard, R.J., 2005. The effect of filter size on VBM analyses of DT-MRI data. Neuroimage 26, 546–554. Kanaan, R.A., Kim, J.S., Kaufmann, W.E., Pearlson, G.D., Barker, G.J., McGuire, P.K., 2005. Diffusion tensor imaging in schizophrenia. Biol. Psychiatry 58, 921–929. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The Positive and Negative Syndrome Scale (PANSS) for schizophrenia. Schizophr. Bull. 13, 261–276. Koutsouleris, N., Gaser, C., Jäger, M., Bottlender, R., Frodl, T., Holzinger, S., Schmitt, G.J., Zetzsche, T., Burgermeister, B., Scheuerecker, J., Born, C., Reiser, M., Möller, H.J., Meisenzahl, E.M., 2008. Structural correlates of psychopathological symptom dimensions in schizophrenia: a voxelbased morphometric study. Neuroimage 39, 1600–1612. Kubicki, M., McCarley, R., Westin, C.F., Park, H.J., Maier, S., Kikinis, R., Jolesz, F.A., Shenton, M.E., 2007. A review of diffusion tensor imaging studies in schizophrenia. J. Psychiatr. Res. 41, 15–30. Kyriakopoulos, M., Vyas, N.S., Barker, G.J., Chitnis, X.A., Frangou, S., 2008. A diffusion tensor imaging study of white matter in early-onset schizophrenia. Biol. Psychiatry 63, 519–523. Lehman, A.F., Lieberman, J.A., Dixon, L.B., McGlashan, T.H., Miller, A.L., Perkins, D.O., Kreyenbuhl, J., American Psychiatric Association Steering Committee on Practice Guidelines, 2004. Practice guideline for the treatment of patients with schizophrenia, second ed. . Am. J. Psychiatry, vol. 161, pp. 1–56.

85

Mandonnet, E., Nouet, A., Gatignol, P., Capelle, L., Duffau, H., 2007. Does the left inferior longitudinal fasciculus play a role in language? A brain stimulation study. Brain 130, 623–629. McGlashan, T.H., Hoffman, R.E., 2000. Schizophrenia as a disorder of developmentally reduced synaptic connectivity. Arch. Gen. Psychiatry 57, 637–648. Meda, S.A., Giuliani, N.R., Calhoun, V.D., Jagannathan, K., Schretlen, D.J., Pulver, A., Cascella, N., Keshavan, M., Kates, W., Buchanan, R., Sharma, T., Pearlson, G.D., 2008. A large scale (N = 400) investigation of gray matter differences in schizophrenia using optimized voxel-based morphometry. Schizophr. Res. 101, 95–105. Mori, S., Wakana, S., Nagae-Poetscher, L., van Zijl, P.C.M., 2005. MRI Atlas of Human White Matter. Elsevier, Amsterdam. Nichols, T.E., Holmes, A.P., 2001. Nonparametric permutation tests for functional neuroimaging: a primer with examples. Hum. Brain Mapp. 15, 1–25. Ongür, D., Price, J.L., 2000. The organization of networks within the orbital and medial prefrontal cortex of rats, monkeys and humans. Cereb. Cortex 10, 206–219. Pakkenberg, B., 1990. Pronounced reduction of total neuron number in mediodorsal thalamic nucleus and nucleus accumbens in schizophrenics. Arch. Gen. Psychiatry 47, 1023–1028. Rajkowska, G., Selemon, L.D., Goldman-Rakic, P.S., 1998. Neuronal and glial somal size in the prefrontal cortex: a postmortem morphometric study of schizophrenia and Huntington disease. Arch. Gen. Psychiatry 55, 215–224. Rueckert, D., Sonoda, L., Hayes, C., Hill, D., Leach, M., Hawkes, D., 1999. Nonrigid registration using free-form deformations: application to breast MR images. IEEE. Trans. Med. Imag. 18, 712–721. Selemon, L.D., Goldman-Rakic, P.S., 1999. The reduced neuropil hypothesis: a circuit based model of schizophrenia. Biol. Psychiatry 45, 17–25. Shenton, M.E., Dickey, C.C., Frumin, M., McCarley, R.W., 2001. A review of MRI findings in schizophrenia. Schizophr. Res. 49, 1–52. Smith, S.M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T.E., Mackay, C.E., Watkins, K.E., Ciccarelli, O., Cader, M.Z., Matthews, P.M., Behrens, T.E., 2006. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 31, 1487–1505. Suzuki, M., Zhou, S.Y., Takahashi, T., Hagino, H., Kawasaki, Y., Niu, L., Matsui, M., Seto, H., Kurachi, M., 2005. Differential contributions of prefrontal and temporolimbic pathology to mechanisms of psychosis. Brain 128, 2109–2122. Talairach, J., Tournoux, P., 1988. Co Planar Stereotaxic Atlas of the Human Brain. Thieme Medical Publishers, New York. Vigneau, M., Beaucousin, V., Hervé, P.Y., Duffau, H., Crivello, F., Houdé, O., Mazoyer, B., Tzourio-Mazoyer, N., 2006. Meta-analyzing left hemisphere language areas: phonology, semantics, and sentence processing. Neuroimage 30, 1414–1432. Walterfang, M., Wood, S.J., Velakoulis, D., Pantelis, C., 2006. Neuropathological, neurogenetic and neuroimaging evidence for white matter pathology in schizophrenia. Neurosci. Biobehav. Rev. 30, 918–948. Yamada, M., Hirao, K., Namiki, C., Hanakawa, T., Fukuyama, H., Hayashi, T., Murai, T., 2007. Social cognition and frontal lobe pathology in schizophrenia: a voxel-based morphometric study. Neuroimage 35, 292–298.