Altered resting-state cerebral blood fl ow and its connectivity in schizophrenia Jiajia Zhu a, 1 , Chuanjun Zhuo b, c, 1 , Wen Qin a, 1 , Yongjie Xu a , Lixue Xu a , Xingyun Liu a , Chunshui Yu a, * a Department of Radiology and Tianjin Key Laboratory of Functional Imaging, Tianjin Medical University General Hospital, Tianjin 300052, China b Tianjin Anning Hospital, Tianjin 300300, China c Department of Psychiatry Functional Neuroimaging Laboratory, Tianjin Mental Health Center, Tianjin Anding Hospital, Tianjin 300070, China a r t i c l e i n f o Article history: Received 29 October 2014 Received in revised form 28 February 2015 Accepted 5 March 2015 Keywords: Arterial spin labeling Cerebral blood fl ow Connectivity Magnetic resonance imaging Schizophrenia a b s t r a c t Small sample sizes and large inter-subject variations result in inconsistent fi ndings in resting-state ce- rebral blood fl ow (CBF) in schizophrenia. The CBF connectivity alterations in schizophrenia remain un- clear. Recently, three-dimensional pseudo-continuous arterial spin labeling (pcASL) imaging was performed to measure the resting-state CBF in 100 schizophrenia patients and 94 healthy comparison subjects. The normalized CBF was used to reduce the inter-subject variations. Both group comparisons in the CBF and correlations between the CBF alterations and clinical parameters were assessed. The CBF connectivity of the brain regions with regional CBF differences was also compared between the groups. Compared with the healthy controls, the schizophrenia patients exhibited increased CBF in the bilateral inferior temporal gyri, thalami and putamen and decreased CBF in the left insula and middle frontal gyrus and the bilateral anterior cingulate cortices and middle occipital gyri. In the schizophrenia patients, signi fi cant correlations were identi fi ed between the CBF and clinical parameters. Importantly, the schizophrenia patients exhibited CBF disconnections between the left thalamus and right medial su- perior frontal gyrus and between the left insula and left postcentral gyrus. Our results suggest that schizophrenia patients may exhibit both regional CBF abnormalities and de fi cits in CBF connectivity, which may underlie the clinical symptoms of schizophrenia. © 2015 Elsevier Ltd. All rights reserved. 1. Introduction Schizophrenia is a common and severe brain disorder charac- terized by abnormal mental activities and disturbed behaviors (Lewis and Lieberman, 2000). It has been associated with alter- ations in resting-state cerebral blood fl ow (CBF). Positron emission tomography (PET) and single photon emission computerized to- mography (SPECT) have traditionally been used to measure CBF changes in schizophrenia. Patients have exhibited increased or decreased resting-state CBF in multiple brain regions, especially the prefrontal cortex (Andreasen et al., 1997; Catafau et al., 1994; Kanahara et al., 2013, 2009; Kawasaki et al., 1993; Malaspina et al., 2004; Mathew et al., 1988; Rubin et al., 1994; Weinberger et al., 1986). Moreover, several resting-state CBF alterations have been associated with the core clinical symptoms of schizophrenia (Lahti et al., 2006; Yuasa et al., 1995). However, PET and SPECT techniques require the use of invasive radioactive tracers, which limits repeated examinations. The other limitations of the two techniques include the time-consuming, expensive image acquisi- tion and low spatial resolution. With the advantages of a noninvasive nature and short acqui- sition time, arterial spin labeling (ASL) magnetic resonance imaging (MRI) provides an alternative approach to the measurement of resting-state CBF using magnetically labeled arterial blood water as an endogenous tracer (Detre et al., 1992). Using this technique, several studies have demonstrated resting-state CBF changes in schizophrenia (Liu et al., 2012; Pinkham et al., 2011; Scheef et al., 2010; Walther et al., 2011), although one study failed to identify signi fi cant group differences (Horn et al., 2009). Furthermore, as- sociations between the altered resting-state ASL-CBF and clinical * Corresponding author. Department of Radiology, Tianjin Medical University General Hospital, No. 154, Anshan Road, Heping District, Tianjin 300052, China. E-mail address: chunshuiyu@tijmu.edu.cn (C. Yu). 1 These authors contributed equally to the article. Contents lists available at ScienceDirect Journal of Psychiatric 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 / p s y c h i r e s http://dx.doi.org/10.1016/j.jpsychires.2015.03.002 0022-3956/ © 2015 Elsevier Ltd. All rights reserved. Journal of Psychiatric Research 63 (2015) 28 e 35 symptoms have also been identi fi ed in schizophrenia (Pinkham et al., 2011). Although the decreased CBF in the frontal cortex has been repeatedly discerned in schizophrenia, the CBF changes in other brain regions differ largely across studies (Liu et al., 2012; Pinkham et al., 2011; Scheef et al., 2010; Walther et al., 2011). The small sample size and large inter-subject variations may account for the inconsistent fi ndings across investigations. Thus, studies that investigate normalized CBF to reduce inter-subject variations in a larger sample size are needed. As a re fl ection of neuronal activity, the regional CBFs of different brain regions are not independent. Instead, the CBFs of brain re- gions from the same functional network may change synchro- nously to ful fi ll the function of the network. In support of the hypothesis, the highest concurrent fl uctuations in CBF have been identi fi ed between homologous cortical regions, and the functional network constructed by CBF connectivity exhibits similar network properties to the networks constructed by anatomical or functional connectivity (Melie-Garcia et al., 2013). Recently, using a group- level independent component analysis on ASL-CBF data, Kindler and colleagues have found increased CBF connectivity within the default-mode network (DMN) (Kindler et al., 2015). However, the CBF connectivity alterations outside the DMN in schizophrenia remain largely unknown. The fi rst aim of this current study was to clarify the CBF alter- ation patterns in schizophrenia. We adopted a 3D pseudo- continuous arterial spin labeling (pcASL) technique that used fast spin echo acquisition and background suppression to provide robustness to motion and susceptibility artifacts and to improve the signal to noise ratio (SNR). We used normalized CBF to reduce the inter-subject difference and a large sample size (100 patients with schizophrenia and 94 healthy comparison subjects) to improve the statistical power. To exclude the effect of cortical atrophy on the CBF results, we also repeated the CBF comparisons after controlling for the regional gray matter volume (GMV). The second aim was to investigate the associations between CBF alterations and clinical parameters. The fi nal aim was to test whether the brain regions with altered CBF also exhibited CBF connectivity changes in schizophrenia. 2. Materials and methods 2.1. Subjects A total of 106 patients with schizophrenia and 94 healthy comparison subjects were included in our study. The individual patient diagnoses were con fi rmed using the Structured Clinical Interview for DSM-IV by trained psychiatrists. The inclusion criteria were age (16 e 60 years) and right-handedness. The exclusion criteria were MRI contraindications, a poor quality of the imaging data, the presence of a systemic medical illness (i.e., cardiovascular disease, diabetes mellitus) or central nervous sys- tem disorder (i.e., epilepsy) that would affect the study results, a history of head trauma (i.e., hematencephalon), or substance (i.e., hypnotics, alcohol) abuse within the previous 3 months or a lifetime history of substance abuse or dependence. Additional exclusion criteria for the comparison subjects included a history of any Axis I or II disorders and a psychotic disorder and fi rst- degree relative with a psychotic disorder. Six patients were excluded because of a poor quality of the imaging data. The fi nal sample included 100 schizophrenia patients and 94 comparison subjects (Table 1). There were no signi fi cant group differences in sex ( c 2 ¼ 2.017, P ¼ 0.156) or age ( t ¼ 0.198, P ¼ 0.844). Ninety-one patients were receiving atypical antipsychotic medications when the MRI examinations were performed, and the other 9 patients have never received any medications. The clinical symptoms of psychosis were quanti fi ed with the Positive and Negative Syn- drome Scale (PANSS) (Kay et al., 1987). The investigation was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of Tianjin Medical University General Hospital. The participants provided informed written consent to participate in the study. 2.2. MRI data acquisition MRI was performed using a 3.0-T MR system (Discovery MR750, General Electric, Milwaukee, WI, USA). Tight but comfortable foam padding was used to minimize head motion, and earplugs were used to reduce scanner noise. Sagittal 3D T1-weighted images were acquired using a brain volume (BRAVO) sequence (repetition time ¼ 8.2 ms; echo time ¼ 3.2 ms; inversion time ¼ 450 ms; fl ip angle ¼ 12 ; fi eld of view ¼ 256 mm 256 mm; matrix ¼ 256 256; slice thickness ¼ 1 mm, no gap; and 188 sagittal slices). The resting-state perfusion imaging was performed using a pcASL sequence with a 3D fast spin-echo acquisition and background suppression (repetition time ¼ 4886 ms, echo time ¼ 10.5 ms, post-label delay ¼ 2025 ms, spiral in readout of eight arms with 512 sample points; fl ip angle ¼ 111 ; fi eld of view ¼ 240 mm 240 mm; reconstruction matrix ¼ 128 128; slice thickness ¼ 4 mm, no gap; 40 axial slices; number of excitation ¼ 3; and 1.9 mm 1.9 mm in-plane resolution). The total acquisition time for the resting state ASL scan was 4 min and 44 s. During the ASL scans, all subjects were instructed to keep their eyes closed, relax and move as little as possible, think of nothing in particular, and not fall asleep. 2.3. CBF calculation The pcASL difference images were calculated after the subtrac- tion of the label images from the control images. The CBF maps were subsequently derived from the ASL difference images. The detailed calculation procedures have been described in a previous study (Xu et al., 2010). SPM8 software was used to coregister the CBF images of the 94 comparison subjects to a PET-perfusion template in the MNI space using non-linear transformation. The MNI-standard CBF template was de fi ned as the mean coregistered CBF image of the 94 healthy comparison subjects. The CBF images of all participants, including the patients and controls, were subse- quently coregistered to the MNI-standard CBF template. Each cor- egistered CBF map was removed of non-brain tissue and spatially smoothed with a Gaussian kernel of 8 mm 8 mm 8 mm FWHM. The CBF of each voxel was normalized by dividing the mean CBF of the whole brain (Aslan and Lu, 2010). Table 1 Demographic and clinical characteristics of the schizophrenia patients and com- parison subjects. Characteristics Schizophrenia patients Comparison subjects P value Number of subjects 100 94 Age (years) 33.6 ± 8.6 33.3 ± 10.4 0.844 Sex (female/male) 43/57 50/44 0.156 Antipsychotic dosage (mg/d) (chlorpromazine equivalents) 453.2 ± 342.9 e Duration of illness (months) 122.9 ± 98.7 e PANSS Positive score 17.0 ± 7.8 e Negative score 20.1 ± 9.0 e Total score 71.3 ± 22.7 e The data are shown as the mean ± SD. PANSS, The Positive and Negative Syndrome Scale. J. Zhu et al. / Journal of Psychiatric Research 63 (2015) 28 e 35 29 2.4. GMV calculation The GMV of each voxel was calculated using Statistical Parametric Mapping (SPM8; http://www. fi l.ion.ucl.ac.uk/spm/software/spm8/). The structural MR images were segmented into gray matter (GM), white matter and cerebrospinal fl uid using the standard uni fi ed segmentation model. After an initial af fi ne registration of the GM concentration map into the Montreal Neurological Institute (MNI) space, the GM concentration images were nonlinearly warped using the diffeomorphic anatomical registration through the exponentiated Lie algebra (DARTEL) technique (Ashburner, 2007) and were resam- pled to a voxel size of 1.5 mm 1.5 mm 1.5 mm. The GMV of each voxel was obtained by multiplying the GM concentration map by the non-linear determinants derived from the spatial normalization step. Finally, the GMV images were smoothed with a Gaussian kernel of 8 mm 8 mm 8 mm full-width at half maximum (FWHM). After spatial preprocessing, the normalized, modulated, and smoothed GMV maps were used for statistical analysis. 2.5. Normalized CBF analyses The group differences in the normalized CBF were compared in a voxelwise manner using a two-sample t -test with age and sex as the nuisance variables. Multiple comparisons were corrected using a family-wise error (FWE) method with a corrected threshold of P < 0.05. For each subject, the normalized CBF of each cluster with a signi fi cant group difference was extracted and used for region of interest (ROI)-based analyses. We used Cohen's d to describe the effect size (ES) of each ROI-based comparison. To exclude the effect of GMV on the CBF comparisons, we also repeated the voxel-based CBF analyses with the GMV of each voxel as a covariate of no interest. 2.6. Correlations between CBF and clinical parameters The ROI-based correlation analyses with antipsychotic dosage in chlorpromazine equivalents, duration of illness and PANSS scores were performed for the patient group using a partial correlation analysis with age and sex as the nuisance covariates. For these correlation analyses, a signi fi cant threshold was set at P < 0.05. 2.7. CBF connectivity analyses Characterizing CBF concurrent changes across subjects between pairs of brain regions by computing the correlation coef fi cient is able to provide a CBF connectivity measure among these brain re- gions (Melie-Garcia et al., 2013). To test whether the brain regions with altered CBF also had altered CBF connectivity in schizophrenia, the clusters with signi fi cant group differences in the CBF were selected as the seed ROIs. The CBF value of each ROI of each subject was extracted from an individual CBF map. For each group, multiple regression models were used to calculate the CBF connectivity between each seed ROI and all other voxels of the whole brain across individuals using sex and age as confounding covariates. These statistical analyses were used to identify the voxels whose CBF values were positively or negatively correlated with the CBF value of each seed ROI in each group. Multiple comparisons were corrected using a FWE method ( P < 0.05). For each ROI, the CBF connectivity maps of the two groups were merged into a spatial mask where the CBF of each voxel was correlated with the CBF of the ROI in either of the two groups. For any pair of voxels, the CBF correlation may have different slopes in the two groups, which re fl ects the difference in CBF connectivity between the groups. To map the voxels that expressed a signi fi cantly different CBF corre- lation with each seed ROI between the schizophrenia patients and the healthy comparison subjects, speci fi c T contrasts were estab- lished within the spatial mask of the CBF connectivity map of the ROI after controlling for age and sex. Multiple comparisons were corrected using a FWE method ( P < 0.05). 3. Results 3.1. Group differences in resting-state normalized CBF In the voxel-based analysis, the CBF differences between the schizophrenia patients and the healthy comparison subjects are shown in Fig. 1 and Table 2. Compared with the healthy com- parison subjects, the schizophrenia patients exhibited increased CBF in the bilateral inferior temporal gyri (ITG) (left: ES ¼ 0.67; right: ES ¼ 0.69), thalami (Th) (left: ES ¼ 0.57; right: ES ¼ 0.63), and putamen (Put) (left: ES ¼ 0.59; right: ES ¼ 0.65) ( P < 0.05, FWE corrected). In contrast, these patients had signi fi cantly decreased CBF in the left insula (Ins) (ES ¼ 0.51) and middle frontal gyrus (MFG) (ES ¼ 0.61) and the bilateral anterior cingulate cortices (ACC) (ES ¼ 0.58) and middle occipital gyri (MOG) (left: ES ¼ 0.58; right: ES ¼ 0.64) ( P < 0.05, FWE corrected). The distribution of the brain regions with signi fi cant differences in the CBF with GMV correction ( P < 0.05, FWE corrected) is shown in Supplementary Figure S1. After GMV correction, the schizophrenia patients exhibited signi fi cantly increased CBF in the bilateral ITG, Th and Put and decreased CBF in the bilateral ACC and MOG and left MFG, which were the same as the regions without GMV correction. However, the original cluster with reduced CBF in the left insula disappeared and a new cluster with increased CBF in the left middle temporal gyrus (MTG) was identi fi ed after GMV correction. 3.2. Correlations between CBF and clinical parameters The signi fi cant correlations between the CBF changes and the illness-relevant demographic characteristics (duration of illness Fig. 1. Brain regions with signi fi cant differences in the normalized CBF between the schizophrenia patients and healthy comparison subjects. Voxel-based analysis in- dicates the brain regions with signi fi cant group differences in the normalized CBF ( P < 0.05, FWE corrected). The warm color represents the signi fi cantly increased CBF in the patients with schizophrenia. The cold color denotes that the CBF was signi fi cantly decreased in the schizophrenia patients. Abbreviations: ACC, anterior cingulate cortex; CBF, cerebral blood fl ow; FWE, family-wise error; Ins, insula; ITG, inferior temporal gyrus; L, left; MFG, middle frontal gyrus; MOG, middle occipital gyrus; Put, putamen; R, right; Th, thalamus. J. Zhu et al. / Journal of Psychiatric Research 63 (2015) 28 e 35 30 and antipsychotic dosage) or clinical symptoms (PANSS scores) are depicted graphically in Fig. 2. The antipsychotic dosage was posi- tively correlated with the CBF of the left Put ( pr ¼ 0.274, P ¼ 0.006) and negatively correlated with the CBF of the left MOG ( pr ¼ 0.235, P ¼ 0.020) in the schizophrenia patients. The CBF of the left ITG exhibited a positive correlation with the duration of illness in the patients ( pr ¼ 0.308, P ¼ 0.002). Regarding the clinical symptoms, the PANSS negative scores were correlated with the CBF values in the left ITG ( pr ¼ 0.261, P ¼ 0.009) and insula ( pr ¼ 0.238, P ¼ 0.018). No signi fi cant correlations were identi fi ed between the CBF values and the PANSS positive scores. 3.3. CBF connectivity patterns A total of 11 ROIs were de fi ned as the seed regions that had signi fi cant CBF differences between the schizophrenia patients and the healthy comparison subjects. The CBF connectivity maps ( P < 0.05, FWE corrected) of each ROI from the two groups are displayed in Fig. 3. The left ITG exhibited CBF connectivity with nearly regions in the controls, but it also exhibited connectivity with the medial prefrontal cortex and inferior frontal cortex in the patients; however, the right ITG had a similar connectivity pattern in both groups. The bilateral thalami demonstrated similar positive connectivity in both groups but had completely different negative connectivity patterns. The bilateral putamen had similar positive connectivity in both groups but exhibited more extensive negative connectivity in the patients. The left insula had similar connectivity in both groups. The ACC demonstrated similar positive connectivity in both groups but had completely different negative connectivity patterns. The left MFG exhibited similar positive connectivity in both groups. The bilateral MOG had similar positive connectivity in both groups but exhibited more extensive negative connectivity in the patients. 3.4. Group differences in CBF connectivity Group differences in CBF connectivity followed the same pattern: the healthy comparison subjects exhibited signi fi cant connectivity, whereas the schizophrenia patients did not exhibit any signi fi cant connectivity (Fig. 4). Compared with the healthy comparison subjects ( pr ¼ 0.610, P < 0.001), the schizophrenia patients exhibited decreased negative CBF connectivity Table 2 Brain regions with signi fi cant group differences in normalized CBF. Regions Brodmann areas Cluster size (Voxels) Peak t values Coordinates in MNI (x, y, z) Schizophrenia patients > Healthy comparison subjects Left inferior temporal gyrus 21, 38 172 6.17 38, 2, 50 Right inferior temporal gyrus 38 163 6.07 40, 6, 48 Left thalamus e 227 5.40 20, 28, 2 Right thalamus e 350 6.28 18, 18, 8 Left putamen e 326 6.06 28, 6, 6 Right putamen e 652 5.99 30, 2, 2 Schizophrenia patients < Healthy comparison subjects Left insula 13 220 6.06 34, 22, 2 Anterior cingulate cortex 24, 32 1638 6.80 4, 36, 24 Left middle frontal gyrus 10 456 5.87 30, 60, 4 Left middle occipital gyrus 18 127 5.86 14, 104, 12 Right middle occipital gyrus 18, 19 665 6.21 26, 100, 8 Abbreviations: CBF, cerebral blood fl ow; MNI, Montreal Neurological Institute. Fig. 2. Correlations between the normalized CBF values and clinical parameters. Abbreviations: CBF, cerebral blood fl ow; Ins, insula; ITG, inferior temporal gyrus; L, left; MOG, middle occipital gyrus; PANSS, The Positive and Negative Syndrome Scale; Put, putamen. J. Zhu et al. / Journal of Psychiatric Research 63 (2015) 28 e 35 31 ( pr ¼ 0.121, P ¼ 0.234) between the seed ROI of the left thalamus and the right medial part of the superior frontal gyrus (Med_SFG) (Brodmann area [BA] 32; peak coordinate: x/y/z ¼ 4/26/46; t ¼ 4.90; and cluster size ¼ 232 voxels) ( P < 0.05, FWE corrected). Compared with the healthy comparison subjects ( pr ¼ 0.547, P < 0.001), the schizophrenia patients also exhibited decreased positive CBF connectivity ( pr ¼ 0.060, P ¼ 0.560) between the seed ROI of the left insula and the left postcentral gyrus (Post_CG) (BA 3; Fig. 3. The CBF connectivity maps of each ROI from the two groups ( P < 0.05, FWE corrected). Abbreviations: ACC, anterior cingulate cortex; CBF, cerebral blood fl ow; FWE, family- wise error; Ins, insula; ITG, inferior temporal gyrus; L, left; MFG, middle frontal gyrus; MOG, middle occipital gyrus; Put, putamen; R, right; Th, thalamus. Fig. 4. Group differences in the CBF connectivity. Compared with the healthy controls (white line), the schizophrenia patients (yellow line) exhibited decreased CBF connectivity between the left thalamus and the right medial superior frontal gyrus and between the left insula and the left postcentral gyrus ( P < 0.05, FWE corrected). Scatter plots demonstrate the CBF connectivity of each group. The solid line represents signi fi cant CBF connectivity, whereas the dotted line represents non-signi fi cant CBF connectivity. Abbreviations: CBF, cerebral blood fl ow; HC, healthy controls; Ins, insula; L, left; Med_SFG, medial superior frontal gyrus; Post_CG, postcentral gyrus; R, right; SZ, schizophrenia patients; Th, thalamus. (For interpretation of the references to color in this fi gure legend, the reader is referred to the web version of this article.) J. Zhu et al. / Journal of Psychiatric Research 63 (2015) 28 e 35 32 peak coordinate: x/y/z ¼ 48/-16/28; t ¼ 4.55; and cluster size ¼ 53 voxels) ( P < 0.05, FWE corrected). The other 9 seed ROIs with signi fi cant group differences in CBF did not exhibit any sig- ni fi cant differences in the CBF connectivity between the two groups ( P > 0.05, FWE corrected). 4. Discussion In the present study, we used a 3D-pcASL technique to investi- gate the normalized CBF and CBF connectivity changes in 100 pa- tients with schizophrenia. These patients had increased CBF in the bilateral ITG, thalami, and putamen and decreased CBF in the left insula and MFG and the bilateral ACC and MOG. Both antipsychotic medication and the duration of illness were correlated with the regional CBF. The CBFs of the left insula and ITG were correlated with the PANSS negative scores. More importantly, the CBF con- nectivities of the left thalamus and insula were also impaired in schizophrenia. We identi fi ed decreased CBF in the left MFG, a cortical region of the frontal lobe, which is consistent with prior fi ndings in schizo- phrenia using PET, SPECT, or ASL-MRI (Andreasen, O'Leary, 1997; Kanahara et al., 2013; Kanahara et al., 2009; Kawasaki et al., 1993; Kindler et al., 2015; Malaspina et al., 2004; Pinkham et al., 2011; Rubin et al., 1994; Scheef et al., 2010; Walther et al., 2011; Weinberger et al., 1986). Because the frontal lobes are highly evolved and regulate complex behaviors in healthy subjects, hypofrontality may contribute to the clinical symptoms in schizo- phrenia (Gur and Gur, 1995). In agreement with previous reports (Kanahara et al., 2009; Scheef et al., 2010), we also identi fi ed decreased CBF in the ACC, which serves to integrate cognitive and emotional processes in support of goal-directed behaviors (Ridderinkhof et al., 2004). The hypoperfusion in the ACC may be related to the dif fi culty in cognitive and emotional integration in schizophrenia (Fornito et al., 2009). In the present study, we identi fi ed decreased CBF in the left insula, which is involved in the multisensory integration and processing of cognitive and emotional information (Wylie and Tregellas, 2010). Our fi nding of the association between decreased CBF in the left insula and increased PANSS negative scores suggests that the hypoperfusion in the insula may contribute to cognitive de fi cits in schizophrenia. Furthermore, both the dorsal ACC and the anterior insula are critical components of the salience network (SN) (Seeley et al., 2007). The SN serves to identify salient stimuli and to interact with the DMN and central executive network (CEN) to guide behaviors (Menon and Uddin, 2010). Thus, the decreased CBF in the SN regions may be related to functional de fi cits, which is consistent with the fi ndings that the reduced SN connectivity with the DMN/CEN is correlated with psychotic symptoms in schizophrenia (Manoliu et al., 2014). The decreased CBF in the occipital lobe is consistent with a previous fi nding of occipital atrophy in chronic schizo- phrenia patients (Onitsuka et al., 2007), which may explain the de fi cits in visual perception in schizophrenia (Butler et al., 2001, 2005). An alternative explanation for the decreased CBF in the oc- cipital lobe may result from the side effects of antipsychotic med- ications, which is supported by our fi nding of the negative correlation between the antipsychotic dosage and CBF in the left MOG in schizophrenia. Consistent with previous fi ndings (Andreasen, O'Leary, 1997; Liu et al., 2012; Malaspina et al., 2004; Mathew et al., 1988; Pinkham et al., 2011; Scheef et al., 2010), we also identi fi ed increased CBF in the bilateral ITG, thalami, and putamen in schizophrenia. The ITG is involved in processes of visual perception (Ishai et al., 1999) and multimodal sensory integration (Mesulam, 1998) and has exhibited volumetric atrophy in schizophrenia (Onitsuka et al., 2004). Although schizophrenia patients had an increased CBF in the ITG, the increased CBF in the left ITG was found to be related to more serious negative symptoms and longer duration of illness. These fi ndings suggest that the clinical implication for the increased CBF in the ITG is unclear and needs to be further investigated. The thalamus is a critical node of multiple functional circuits, such as the frontostriatal and cerebellothalamocortical circuits, both of which are impaired in schizophrenia (Byne et al., 2009). The pu- tamen, which is a component of the striatum, is involved in motor, cognitive, and emotional functions. It contains rich dopaminergic neurons, the function of which has been consistently demonstrated to be impaired in schizophrenia (Perez-Costas et al., 2010). More importantly, we demonstrated that the increased CBF in the left putamen was correlated with an increased antipsychotic dosage, which has also been demonstrated in previous PET studies (Lahti et al., 2003, 2009). Moreover, a recent study with a similar pcASL sequence identi fi ed a correlation between the CBF and chlor- promazine equivalents in the same location as our study (Kindler et al., 2013). These fi ndings suggest that the increased CBF in the putamen is more likely a re fl ection of antipsychotic drug effects. After GMV correction, the schizophrenia patients exhibited signi fi cantly increased CBF in the bilateral ITG, Th and Put and decreased CBF in the bilateral ACC and MOG and left MFG, which were the same as the regions without GMV correction. These fi ndings indicate that altered CBFs in these regions are independent of GMV changes. However, the original cluster with reduced CBF in the left insula disappeared and a new cluster with increased CBF in the MTG was identi fi ed after GMV correction, which suggests that the altered CBF in the two regions is related to GM atrophy, either secondary to or compensatory for GM atrophy. Connectivity alterations in schizophrenia have been extensively investigated by MRI techniques, such as the structural connectivity derived from structural MRI, the anatomical connectivity derived from diffusion tensor MRI, and the functional connectivity derived from functional MRI. However, only a pioneer study has investi- gated the CBF connectivity within the DMN in schizophrenia (Kindler et al., 2015). Although both CBF connectivity and BOLD connectivity measure functional correlations between pairs of brain regions, they are calculated using different methods and indicate different physiological meanings. The CBF connectivity was measured by computing the correlation coef fi cient of the CBF between pairs of brain regions across a group of individuals; however, the BOLD connectivity was obtained by measuring the temporal correlations of the BOLD signal fl uctuations between pairs of brain areas in a single individual (Biswal et al., 1995). For a group of individuals, only one value of CBF connectivity could be obtained, whereas multiple values (a value for an individual) of BOLD con- nectivity could be obtained. Physiologically, CBF connectivity re- fl ects concertedness in metabolism or perfusion between brain regions; however, BOLD connectivity represents temporal syn- chronization in the neural activity between brain regions. Compared with the BOLD connectivity, which is in fl uenced by multiple physiological parameters, such as the CBF, cerebral blood volume (CBV) and cerebral metabolic rate of oxygen (Buxton et al., 2004), the CBF connectivity is only modulated by regional CBF and has a more de fi nite physiological implication (Melie-Garcia et al., 2013). The disadvantage of the CBF connectivity is that it cannot be used to compare the connectivity difference between any two subjects because the CBF connectivity is a group-level measure rather than an individual-level measure. Here, we investigated the CBF connectivity alterations of brain regions that had signi fi cant changes in the regional CBF in schizophrenia. We identi fi ed a CBF disconnection between the left insula and sensorimotor area in schizophrenia. The functional connectivity of the insula with the sensorimotor area may be associated with sensorimotor integration (Cauda et al., 2011). Thus, the disconnection between the two J. Zhu et al. / Journal of Psychiatric Research 63 (2015) 28 e 35 33 regions may be related to the functional de fi cits in sensorimotor integration in schizophrenia. A recent fMRI BOLD connectivity study has identi fi ed reduced functional coupling between the insula and postcentral gyrus (Ebisch et al., 2014), which is consis- tent with our fi ndings. A reduced CBF connectivity was also iden- ti fi ed between the left thalamus and the right medial superior frontal gyrus in the patient group, which is consistent with a pre- vious resting-state functional connectivity study that indicates a prefrontal-thalamic disconnection in schizophrenia (Woodward et al., 2012). There are several limitations in our study. First, most of our patients exhibited chronic schizophrenia with mixed symptoms and have received medications for a long time, which may in fl u- ence our interpretation. Investigations using medication-naïve, fi rst-episode schizophrenia patients with more homogeneous clinical features may facilitate a more accurate characterization of the pathogenesis of schizophrenia. Second, the CBF connectivity was calculated at a group level by analyzing the inter-regional CBF correlations across subjects, i.e., a group only had one value of CBF connectivity. This measure cannot be used to investigate its cor- relations with demographic or clinical parameters. The develop- ment of the ASL technique with a higher temporal resolution and the calculation of the CBF connectivity as the resting-state func- tional connectivity may help to resolve this issue. Third, only brain regions with signi fi cant group differences in CBF were selected as the ROIs to perform the CBF connectivity analysis, which may miss CBF connectivity alterations between regions with normal regional CBF values. A data-driven method of CBF connectivity analysis throughout the whole brain is needed. Finally, we did not collect information regarding the years of education for the subjects, which may be different between the two groups and may affect our results. In conclusion, we used a noninvasive 3D-pcASL technique to investigate the resting-state normalized CBF and CBF connectivity changes under a strict threshold (FWE correction) in a large sample size of schizophrenia patients. We identi fi ed altered CBF in multiple cortical and subcortical regions, which may underlie de fi cits in multiple systems in schizophrenia and account for the symptom- atic diversity of the disorder. We also identi fi ed diminished CBF connectivity of the left thalamus and the left insula in schizo- phrenia for the fi rst time, which may further support the hypoth- esis that schizophrenia is a connectivity disorder from the perspective of CBF connectivity. Both the regional CBF alterations and de fi cits of CBF connectivity in our fi ndings highlight the need to investigate the underlying neuropathology of schizophrenia from the perspective of both regional and inter-regional properties of resting-state CBF. Role of funding source This work was supported by grants from the National Basic Research Program of China (973 program, 2011CB707801); Natural Science Foundation of China (91332113 and 81271551) and Tianjin Key Technology R & D Program (14ZCZDSY00018). Contributors Jiajia Zhu and Chunshui Yu designed the study, collected data, wrote the protocol and the draft of the manuscript. Chuanjun Zhuo undertook neurological, psychopathological and psychometric as- sessments. Wen Qin performed image processing and statistical analyses. Yongjie Xu and Lixue Xu operated the magnetic resonance imaging (MRI) machine. Xingyun Liu managed literature searches. All authors contributed to and have approved the fi nal manuscript. Con fl ict of interest All authors declare that they have no con fl icts of interest. Acknowledgments We would gratefully thank for the support and assistance from Zhenyu Zhou Ph.D., and Ziheng, Zhang Ph.D. of GE Healthcare China Research Team. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpsychires.2015.03.002. References Andreasen NC, O'Leary DS, Flaum M, Nopoulos P, Watkins GL, Boles Ponto LL, et al. Hypofrontality in schizophrenia: distributed dysfunctional circuits in neuroleptic-naive patients. Lancet 1997;349:1730 e 4. Ashburner J. A fast diffeomorphic image registration algorithm. Neuroimage 2007;38:95 e 113. Aslan S, Lu H. On the sensitivity of ASL MRI in detecting regional differences in cerebral blood fl ow. Magn Reson Imaging 2010;28:928 e 35. Biswal B, Yetkin FZ, Haughton VM, Hyde JS. Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn Reson Med 1995;34:537 e 41. Butler PD, Schechter I, Zemon V, Schwartz SG, Greenstein VC, Gordon J, et al. Dysfunction of early-stage visual processing in schizophrenia. Am J Psychiatry 2001;158:1126 e 33. Butler PD, Zemon V, Schechter I, Saperstein AM, Hoptman MJ, Lim KO, et al. Early- stage visual processing and cortical ampli fi cation de fi cits in schizophrenia. Arch Gen Psychiatry 2005;62:495 e 504. Buxton RB, Uludag K, Dubowitz DJ, Liu TT. Modeling the hemodynamic response to brain activation. Neuroimage 2004;23(Suppl. 1):S220 e 33. Byne W, Hazlett EA, Buchsbaum MS, Kemether E. The thalamus and schizophrenia: current status of research. Acta Neuropathol 2009;117:347 e 68. Catafau AM, Parellada E, Lomena FJ, Bernardo M, Pavia J, Ros D, et al. Prefrontal and temporal blood fl ow in schizophrenia: resting and activation technetium-99m- HMPAO SPECT patterns in young neuroleptic-naive patients with acute disease. J Nucl Med 1994;35:935 e 41. Cauda F, D'Agata F, Sacco K, Duca S, Geminiani G, Vercelli A. Functional connectivity of the insula in the resting brain. Neuroimage 2011;55:8 e 23. Detre JA, Leigh JS, Williams DS, Koretsky AP. Perfusion imaging. Magn Reson Med 1992;23:37 e 45. Ebisch SJ, Mantini D, Northoff G, Salone A, De Berardis D, Ferri F, et al. Altered brain long-range functional interactions underlying the link between aberrant self- experience and self-other relationship in fi rst-episode schizophrenia. Schiz- ophr Bull 2014;40:1072 e 82. Fornito A, Yucel M, Dean B, Wood SJ, Pantelis C. Anatomical abnormalities of the anterior cingulate cortex in schizophrenia: bridging the gap between neuro- imaging and neuropathology. Schizophr Bull 2009;35:973 e 93. Gur RC, Gur RE. Hypofrontality in schizophrenia: RIP. Lancet 1995;345:1383 e 4. Horn H, Federspiel A, Wirth M, Muller TJ, Wiest R, Wang JJ, et al. Structural and metabolic changes in language areas linked to formal thought disorder. Br J Psychiatry 2009;194:130 e 8. Ishai A, Ungerleider LG, Martin A, Schouten JL, Haxby JV. Distributed representation of objects in the human ventral visual pathway. Proc Natl Acad Sci U S A 1999;96:9379 e 84. Kanahara N, Sekine Y, Haraguchi T, Uchida Y, Hashimoto K, Shimizu E, et al. Orbi- tofrontal cortex abnormality and de fi cit schizophrenia. Schizophr Res 2013;143: 246 e 52. Kanahara N, Shimizu E, Sekine Y, Uchida Y, Shibuya T, Yamanaka H, et al. Does hypofrontality expand to global brain area in progression of schizophrenia?: a cross-sectional study between fi rst-episode and chronic schizophrenia. Prog Neuropsychopharmacol Biol Psychiatry 2009;33:410 e 5. Kawasaki Y, Maeda Y, Suzuki M, Urata K, Higashima M, Kiba K, et al. SPECT analysis of regional cerebral blood fl ow changes in patients with schizophrenia during the Wisconsin Card Sorting Test. Schizophr Res 1993;10:109 e 16. Kay SR, Fiszbein A, Opler LA. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr Bull 1987;13:261 e 76. Kindler J, Homan P, Jann K, Federspiel A, Flury R, Hauf M, et al. Reduced neuronal activity in language-related regions after transcranial magnetic stimulation therapy for auditory verbal hallucinations. Biol Psychiatry 2013;73:518 e 24. Kindler J, Jann K, Homan P, Hauf M, Walther S, Strik W, et al. Static and dynamic characteristics of cerebral blood fl ow during the resting state in schizophrenia. Schizophr Bull 2015;41:163 e 70. J. Zhu et al. / Journal of Psychiatric Research 63 (2015) 28 e 35 34 Lahti AC, Holcomb HH, Weiler MA, Medoff DR, Tamminga CA. Functional effects of antipsychotic drugs: comparing clozapine with haloperidol. Biol Psychiatry 2003;53:601 e 8. Lahti AC, Weiler MA, Holcomb HH, Tamminga CA, Carpenter WT, McMahon R. Correlations between rCBF and symptoms i