Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Common and distinct structural features of schizophrenia and bipolar disorder: The European Network on Psychosis, Affective disorders and Cognitive Trajectory (ENPACT) study

  • Eleonora Maggioni,

    Roles Data curation, Formal analysis, Investigation, Methodology, Software, Writing – original draft, Writing – review & editing

    Affiliation Department of Neurosciences and Mental Health, IRCCS Fondazione Ca’ Granda Ospedale Maggiore Policlinico, University of Milan, Milan, Italy

  • Benedicto Crespo-Facorro,

    Roles Conceptualization, Data curation, Investigation, Resources, Supervision, Writing – review & editing

    Affiliations Department of Psychiatry, University Hospital Marqués de Valdecilla, School of Medicine, University of Cantabria-IDIVAL, Santander, Spain, CIBERSAM, Centro Investigación Biomédica en Red Salud Mental, Santander, Spain

  • Igor Nenadic,

    Roles Conceptualization, Data curation, Investigation, Resources, Supervision, Writing – review & editing

    Affiliations Department of Psychiatry and Psychotherapy, Jena University Hospital, Jena, Germany, Department of Psychiatry and Psychotherapy, Philipps University Marburg / Marburg University Hospital UKGM, Marburg, Germany

  • Francesco Benedetti,

    Roles Conceptualization, Data curation, Investigation, Resources, Supervision, Writing – review & editing

    Affiliation Department of Clinical Neurosciences and Centro di Eccellenza Risonanza Magnetica ad Alto Campo, Scientific Institute and University Vita-Salute San Raffaele, Milan, Italy

  • Christian Gaser,

    Roles Data curation, Investigation, Methodology, Software, Writing – review & editing

    Affiliations Department of Psychiatry and Psychotherapy, Jena University Hospital, Jena, Germany, Department of Neurology, Jena University Hospital, Jena, Germany

  • Heinrich Sauer,

    Roles Data curation, Investigation, Supervision, Writing – review & editing

    Affiliation Department of Psychiatry and Psychotherapy, Jena University Hospital, Jena, Germany

  • Roberto Roiz-Santiañez,

    Roles Data curation, Investigation, Resources, Supervision, Writing – review & editing

    Affiliations Department of Psychiatry, University Hospital Marqués de Valdecilla, School of Medicine, University of Cantabria-IDIVAL, Santander, Spain, CIBERSAM, Centro Investigación Biomédica en Red Salud Mental, Santander, Spain

  • Sara Poletti,

    Roles Data curation, Investigation, Resources, Writing – review & editing

    Affiliation Department of Clinical Neurosciences and Centro di Eccellenza Risonanza Magnetica ad Alto Campo, Scientific Institute and University Vita-Salute San Raffaele, Milan, Italy

  • Veronica Marinelli,

    Roles Data curation, Investigation, Resources, Writing – review & editing

    Current address: Department of Pathophysiology and Transplantation, University of Milan, Milan, Italy

    Affiliation Department of Experimental and Clinical Medical Sciences (DISM), University of Udine, Udine, Italy

  • Marcella Bellani,

    Roles Data curation, Investigation, Resources, Writing – review & editing

    Affiliation Section of Psychiatry, Azienda Ospedaliera Universitaria Integrata Verona, Verona, Italy

  • Cinzia Perlini,

    Roles Data curation, Investigation, Resources, Writing – review & editing

    Affiliation Department of Neurosciences, Biomedicine and Movement Sciences, Section of Clinical Psychology, University of Verona, Verona, Italy

  • Mirella Ruggeri,

    Roles Data curation, Investigation, Resources, Writing – review & editing

    Affiliation Department of Neurosciences, Biomedicine and Movement Sciences, Section of Psychiatry, University of Verona, Verona, Italy

  • A. Carlo Altamura,

    Roles Data curation, Investigation, Supervision, Writing – review & editing

    Affiliation Department of Neurosciences and Mental Health, IRCCS Fondazione Ca’ Granda Ospedale Maggiore Policlinico, University of Milan, Milan, Italy

  • Vaibhav A. Diwadkar,

    Roles Conceptualization, Data curation, Investigation, Methodology, Software, Supervision, Writing – review & editing

    Affiliation Department of Psychiatry & Behavioral Neuroscience, Wayne State University, Detroit, MI, United States of America

  • Paolo Brambilla ,

    Roles Conceptualization, Data curation, Investigation, Methodology, Project administration, Resources, Supervision, Visualization, Writing – review & editing

    paolo.brambilla1@unimi.it

    Affiliations IRCCS Scientific Institute “E. Medea”, Bosisio Parini, Lecco, Italy, Department of Pathophysiology and Transplantation, University of Milan, Milan, Italy

  •  [ ... ],
  • ENPACT group

    Membership of the ENPACT group is provided in the Acknowledgments.

  • [ view all ]
  • [ view less ]

Abstract

Introduction

Although schizophrenia (SCZ) and bipolar disorder (BD) share elements of pathology, their neural underpinnings are still under investigation. Here, structural Magnetic Resonance Imaging (MRI) data collected from a large sample of BD and SCZ patients and healthy controls (HC) were analyzed in terms of gray matter volume (GMV) using both voxel based morphometry (VBM) and a region of interest (ROI) approach.

Methods

The analysis was conducted on two datasets, Dataset1 (802 subjects: 243 SCZ, 176 BD, 383 HC) and Dataset2, a homogeneous subset of Dataset1 (301 subjects: 107 HC, 85 BD and 109 SCZ). General Linear Model analyses were performed 1) at the voxel-level in the whole brain (VBM study), 2) at the regional level in the anatomical regions emerged from the VBM study (ROI study). The GMV comparison across groups was integrated with the analysis of GMV correlates of different clinical dimensions.

Results

The VBM results of Dataset1 showed 1) in BD compared to HC, GMV deficits in right cingulate, superior temporal and calcarine cortices, 2) in SCZ compared to HC, GMV deficits in widespread cortical and subcortical areas, 3) in SCZ compared to BD, GMV deficits in insula and thalamus (p<0.05, cluster family wise error corrected). The regions showing GMV deficits in the BD group were mostly included in the SCZ ones. The ROI analyses confirmed the VBM results at the regional level in most of the clusters from the SCZ vs. HC comparison (p<0.05, Bonferroni corrected). The VBM and ROI analyses of Dataset2 provided further evidence for the enhanced GMV deficits characterizing SCZ. Based on the clinical-neuroanatomical analyses, we cannot exclude possible confounding effects due to 1) age of onset and medication in BD patients, 2) symptoms severity in SCZ patients.

Conclusion

Our study reported both shared and specific neuroanatomical characteristics between the two disorders, suggesting more severe and generalized GMV deficits in SCZ, with a specific role for insula and thalamus.

1. Introduction

The neurobiological relationship between schizophrenia (SCZ) and bipolar disorder (BD) has been the subject of many debates but is still largely unknown [13]. Currently, the dimensional approach for diagnosing mental disorders has revitalized the question of shared characterization of SCZ and BD in a dimensional continuum view, as opposed to a categorical dichotomous vision. This has been supported by the neurobiological research that has shown that SCZ and BD exhibit shared genetic [4, 5] and neurocognitive determinants [6], besides clinical features [7]. The investigation of the neuroanatomical characteristics of the two disorders can therefore help to further delineate their common and different pathophysiological bases, opening the door to the development of tailored instruments with higher diagnostic specificity.

Meta-analyses of Magnetic Resonance Imaging (MRI) studies showed that both SCZ and BD are characterized by a global brain volume reduction, as well as by ventricular enlargement compared to healthy controls (HC) [1, 8, 9]. Voxel-Based Morphometry (VBM) meta-analyses reported gray matter volume (GMV) deficits in insula, thalamus, cingulate cortex, medial frontal gyrus, middle and superior temporal gyri in SCZ [8, 10], and in anterior cingulate cortex, inferior frontal gyrus, insula, middle and superior temporal gyri and pole and claustrum in BD [8, 11]. These studies suggest overlapping areas of GMV reduction among the two disorders, but also provide evidence for the larger extent of the deficits in SCZ than in BD [2, 8]. Conversely, only a minority of MRI studies directly compared BD and SCZ [3, 1216]. A recent review on voxel-based comparisons between SCZ and BD found consistent evidence of GM deficits in SCZ compared to BD [17]. Region-based studies on SCZ vs. BD found in BD increased volume in the right amygdala and decreased volume in the bilateral ventricles compared to SCZ [1].

However, the meta-analyses that provided evidence for anatomical brain differences in patients with SCZ and BD were limited by the heterogeneity of the primary studies in terms of 1) clinical sample, 2) applied MR methodology and 3) post-processing techniques.

To provide robust information on the neuroanatomical correlates of SCZ and BD, the present work performs voxel-based and region-based analyses on a large dataset of 802 MR images recorded from HC, SCZ and BD patients across four European Research Centers, 1) the Jena University Hospital (Jena, Germany) (JUH), 2) the University Hospital Marqués de Valdecilla (Santander, Spain) (UHMV), 3) the Scientific Institute and University Vita-Salute San Raffaele (Milan, Italy) (UVSSR), 4) the University Hospital of Verona, (Verona, Italy) (VUH). These Centers, together with the University of Milan, Milan (Italy) (UNIMI), promote the European Network on Psychosis, Affective disorders and Cognitive Trajectory (ENPACT), whose purpose is to share and integrate clinical, demographic and neuroimaging data on BD and SCZ patients, as well as HC, in order to identify large-scale neuroanatomical similarities and differences associated with the two disorders. To our knowledge, this is one of the first studies comparing such a large sample of SCZ, BD and HC subjects in terms of GMV through voxel-based and region-based approaches.

2. Materials and methods

2.1. Subjects

Eight hundred and thirty-one subjects, 259 SCZ patients, 187 BD patients and 385 HC, were recruited across the four Centers. Preliminarily to the MRI acquisition, they signed a written informed consent to the protocol, in accordance with the Declaration of Helsinki and approved by the local ethical committee guidelines. After a quality check procedure, 802 subjects were selected for the analysis.

2.1.1. Recruitment.

In each Center, the diagnoses of BD and SCZ were assessed using the Structured Clinical Interview for Axis-I DSM Disorders (SCID) [18] and confirmed by the clinical consensus of an expert psychiatrist. In most of the cases, SCZ patients were scored using the Brief Psychiatric Rating Scale (BPRS), which was combined with the Positive and Negative Syndrome Scale (PANSS) (in Milan and Verona) or the Scales for Assessment of Positive/Negative symptoms (SANS/SAPS) (in Jena and Santander). In the majority of BD patients, depressive and manic symptoms were investigated using the Hamilton Rating Scale for Depression (HRSD) [19] and the Young Mania Rating Scale (YMRS) [20]. HC were recruited within the local communities of Jena, Milan, Santander and Verona through flyers and word of mouth. Subjects with personal or family history of psychiatric illnesses, personal history of substance or alcohol abuse, mental retardation or neurological disorders were excluded from the study.

2.2. MRI data acquisition

Structural T1-weighted images were recorded in the four Centers using 3.0 Tesla MRI scanners with the following parameters. JUH: Magnetization Prepared Rapid Gradient Echo (MPRAGE) sequence (matrix 256x256x192, 1 mm3 voxel), Siemens Tim Trio scanner (Siemens, Erlangen, Germany). UVSSR: MPRAGE sequence (matrix 256x256x220, 0.9x0.9x0.8 mm3 voxel), Philips Gyroscan Intera MR scanner (Philips, Best, the Netherlands). UHMV: T1-Fast Field Echo sequence (matrix 256x256x160, 0.94x0.94x1 mm3 voxel), Philips Intera scanner (Philips, Best, the Netherlands). VUH: MPRAGE sequence (matrix 256x256x160, 1 mm3 voxel), Magnetom Allegra Syngo scanner (Siemens, Erlangen, Germany).

2.3. MRI data processing

2.3.1. Pre-processing.

In a first quality check, the images affected by important inhomogeneity or movement artefacts were discarded. The images were then subjected to a non-parametric non-uniform intensity normalization (N3) using Freesurfer (http://surfer.nmr.mgh.harvard.edu) [21]. This preliminary step was specifically added to remove non-uniformity artifacts and homogenize the image intensities across sites. The bias-corrected images were exported to Matlab R2014A (The Mathworks, Inc®) for the following analyses. The image pre-processing was carried out using Statistical Parametric Mapping (SPM), version 12 (http://www.fil.ion.ucl.ac.uk/spm/software/spm12/), and its VBM8 toolbox (http://www.neuro.uni-jena.de/vbm/download/). The images were segmented using SPM12 segmentation, which performs bias regularization and classifies the images into gray matter (GM), white matter (WM), cerebrospinal fluid, bone, soft tissue and air/background. In this procedure, the GM and WM images of the subjects were also rigidly aligned via an affine transformation. The Dartel (Diffeomorphic Anatomical Registration Through Exponentiated Lie algebra) tools were then used to determine the nonlinear deformations for registering the GM and WM images of all subjects. Finally, the registered images were normalized to MNI space and smoothed with a 6 mm Full Width Half Maximum (FWHM) Gaussian kernel. A final quality check on the pre-processed images based on 1) visual inspection and 2) covariance homogeneity was performed using VBM8 modules: the images with an overall covariance below two standard deviations were discarded from the analysis. The VBM and ROI analyses (2.3.2 and 2.3.3 sections) were applied to both datasets. The results on Dataset2 are described in a Supplementary section, S1 Results.

2.3.2. VBM comparison across diagnostic groups.

The VBM analysis was performed on the pre-processed GM images using SPM12. The VBM analysis of each dataset was performed using a full-factorial General Linear Model (GLM) design with two factors, the diagnosis factor with 3 levels (SCZ, BD and HC) and the scanning site factor with four levels (Jena, Milan, Santander, Verona). While the diagnosis represented the factor of interest, the center factor was included with the purpose to discard its contribution. The measurements were assumed to be independent and with unequal variance between levels. Age and gender were included as covariates, both interacting with the factor diagnosis. The volumetric differences among subjects were considered by proportional scaling for the total intracranial volume (ICV). A GM mask with optimal threshold for GMV was created using the masking toolbox (http://www0.cs.ucl.ac.uk/staff/g.ridgway/masking/) and used in the GLM analysis. Inference on the GMV differences between groups was made using double-sided t-tests (p<0.05, cFWE corrected). A conjunction analysis was performed to extract the GMV reductions common to SCZ and BD. A direct SCZ-BD comparison was also conducted.

2.3.3. ROI comparison across diagnostic groups.

The ROI analysis was performed on the anatomical regions containing the significant VBM clusters (p<0.05, cFWE corrected), with the main objective to verify whether the local GMV differences emerged from the VBM comparison were still significant at the global level in the corresponding anatomical regions. The anatomical location of the VBM clusters resulting from the pairwise t-contrasts (paragraph 2.3.2) was identified using the Automated Anatomical Labeling (AAL) atlas [22]. The volume of the selected AAL regions was estimated by summing up the GMV of the voxels within the entire anatomical region. In the GLM analyses, performed using in-house Matlab scripts, the AAL regional volume was modeled in terms of diagnosis, scanning site, age, gender and ICV. The differences between groups were investigated using double-sided t-tests. To limit the false positive rates, a very conservative multiple comparison correction was applied by using the Bonferroni method with N = 116, corresponding to the number of regions within the AAL atlas. The significance threshold was set to p = 0.05, Bonferroni corrected.

2.3.4. Clinical-neuroanatomical correlations.

In the patient groups, we investigated the possible effects of clinical variables on GMV. All the clinical variables were analyzed in separate designs, including case by case the subsets of patients having the information of interest. More details on scores and availability of clinical information can be found in Table 1.

thumbnail
Table 1. Demographic and clinical details in the two datasets.

In Dataset1, from the analysis of variance (ANOVA), significant age differences across groups emerged (p<0.01). Sex ratio (Males/Females) was 1.67 in SCZ, 0.64 in BD and 1.04 in HC (χ2 = 23.02, p<0.0001).

https://doi.org/10.1371/journal.pone.0188000.t001

In the BD sample, we analyzed the contribution of HRSD, BD type, psychotic features and therapy based on benzodiazepines, antipsychotics and mood stabilizers. In the SCZ sample, we investigated the contribution of BPRS, PANSS, SAPS, SANS, duration of untreated illness and therapy with benzodiazepines. Since SAPS/SANS scores were not uniformly distributed between Jena and Santander, their effects were studied in one center at a time. The possible impact of age of onset on GMV was separately investigated in the two groups of patients.

The contribution of each clinical parameter on GMV was evaluated first at the voxel level, using VBM, and then at the regional level, considering the significant AAL regions/clinical variables from the VBM study. The ROI study was performed to verify the significance of the clinical-neuroanatomical relation at the regional level.

In the VBM GLM analyses, the GMV of each voxel was modeled in terms of the clinical variable of interest, age and gender. Since some variables were available across different centers, in the corresponding models we included the site factor to remove its contribution. As in the main VBM analysis (section 2.3.2), we accounted for the effect of head volume by proportional scaling with ICV and we masked the results using the optimal threshold GM mask. Inference on the contribution of the variable of interest was performed using double-sided t-tests (p<0.001, >100 voxels).

In the ROI analyses, using in-house Matlab scripts, we conducted Pearson linear partial correlation analyses between regional GMV (normalized for ICV) and the clinical variable of interest, while accounting for the contribution of age and gender and, when appropriate, scanning site. The significance threshold was set to p = 0.01.

3. Results

3.1. Demographic and clinical information

The demographic and clinical details of the two datasets are reported in Table 1.

Dataset1 was composed of 383 HC (195 males, range 18–62 years, 30.4 ± 9.2 years), 243 patients with SCZ (152 males, range 17–61 years, 33.24 ± 9.41 years) and 176 patients with BD (69 males, range 20–76 years, 44.7 ± 12.08 years). Both psychotic and non-psychotic BD patients (PBD, NPBD) as well as BD type I (BD-I) and BD type II (BD-I) patients took part to the study: PBD patients were in minority, being 42 out of 176, whereas BD-I patients were in majority, being 111 out of 176. Since no significant GMV differences emerged between them (more details are in paragraph 3.3), in the main analysis BD patients were considered in a unique group. From the analysis of variance (ANOVA), significant age differences across BD, SCZ and HC emerged (p<0.01). Sex ratio (M/F) was 1.67 in SCZ, 0.64 in BD and 1.04 in HC (χ2 = 23.02, p<0.0001). The fact that SCZ is diagnosed 1.4 times more frequently in males than in females [23] supports the SCZ gender ratio. On the other hand, the women prevalence in the BD group may be explained by the fact that BD patients were mostly recruited as depressed inpatients, as women with BD may have more depressive and mixed episodes than do men with the illness.

At the time of the MRI study, 23 BD patients were euthymic, 1 was in hypomanic state, 1 had mixed episodes, 27 were in manic phase and 120 were in depressed phase. This information was missing in 4 patients. The assessment of euthymic, manic and depressed states was supported by HRSD and YMRS scores. Indeed, in manic patients the mean YMRS score was 20.8, in depressed patients the mean HDRS score was 22.32, whereas in euthymic ones both HDRS and YMRS were below 8. Overall, the prevalence of patients in depressed phase is confirmed by the mean HRSD score, which was above 17 (as reported in Table 1).

In the SCZ sample, the mean BPRS and positive and negative symptoms rates from PANSS or SANS/SAPS suggest that patients from all centers were from moderately to markedly ill [24, 25].

Concerning medication, antipsychotics were assumed by most SCZ patients (95.88%) and by less than half of BD patients (32.39%). Among SCZ patients, 3 were treated with amisulpride, 49 with aripiprazole, 42 with clozapine, 13 with haloperidol, 14 with olanzapine, 6 with paliperidone, 25 with quetiapine, 32 with risperidone, 17 with ziprasidone; this information was missing for 30 patients. Among BD patients taking neuroleptics, 2 were treated with haloperidol, 4 with amisulpride, 2 with aripiprazole, 1 with chlorpromazine, 1 with clozapine, 11 with haloperidol, 5 with olanzapine, 1 with paliperidone, 15 with quetiapine and 7 with risperidone; this information was missing for 7 patients. From the available information on benzodiazepines, we know that BD patients taking benzodiazepines were in majority (at least the 53%). This drug was used also in the therapy of around one half of SCZ patients (45% or more). Mood stabilizers were included in the therapy of approximately half of the BD sample (at least the 48%) but were almost never used in the treatment of SCZ patients (≈2%).

A reduced dataset (Dataset2), more homogeneous in terms of age and diagnosis, was manually extracted from the original one for a further analysis, with the main objective to investigate the neuroanatomical differences among SCZ, BD and HC in a subset of Dataset1 without significant age differences across groups. Dataset2 included 107 HC (52 males, 20–62 years, 39.02 ± 10.26 years), 85 BD patients (34 males, 20–60 years, 39.13 ± 9.9 years) and 109 SCZ patients (67 males, 20–61 years, 39.1 ± 8.78 years). Since each group of patients was reduced by more than half, from the analysis of Dataset2 we did not expect to reproduce the findings from the main dataset, but we wanted to extract the most relevant features of the disorders. Compared to Dataset1, the balance between psychotic vs. non-psychotic BD patients and BD-I vs. BD-II patients was preserved; again, the depressed patients outnumbered the euthymic and manic ones. Concerning the SCZ sample, the mean BPRS, SAPS/SANS and PANSS scores still indicated moderate to severe schizophrenic symptoms.

Concerning medication, antipsychotics were taken by 96.33% of the SCZ sample and 38.82% of the BD sample. Among SCZ patients, 1 used amisulpride, 26 aripiprazole, 15 clozapine, 7 haloperidol, 5 olanzapine, 4 paliperidone, 9 quetiapine, 15 risperidone and 9 ziprasidone; the antipsychotic type was not known in 14 patients. Among BD patients, 1 used aripiprazole, 1 chlorpromazine, 9 haloperidol, 4 olanzapine, 10 quetiapine and 4 risperidone, whereas the neuroleptic type was not known in 4 patients. Benzodiazepines were used by almost half of both BD and SCZ patients (≈45%). Mood stabilizers were included in the therapy of around half of BD patients (48% or more) but were almost unused in SCZ patients (less than 5%).

3.2. Comparison among diagnostic groups

The VBM and ROI results of Dataset1 are separately reported in sections 3.2.1 and 3.2.2, respectively. The VBM and ROI results of Dataset2 are described in S1 Results.

3.2.1. VBM analysis.

The VBM results of Dataset1 are reported in Fig 1, where the clusters with significant GMV difference between the three couples of groups are shown. The anatomical location and statistics of the SPM peaks (the first of each cluster) and the corresponding cluster extension are listed in Table 2. Compared to HC, BD patients showed reduced GMV (p<0.05, cFWE corrected) in three areas, mainly in the right hemisphere. These regions were located in 1) anterior and mesial portions of the cingulate cortex, 2) superior temporal cortex and temporal pole, 3) calcarine cortex, cuneus and lingual gyrus. The peak voxel was in the right subgenual cingulate cortex.

thumbnail
Fig 1. Results of the VBM analysis of Dataset1: regions with significant GMV difference between SCZ, BD and HC (p<0.05, cFWE corrected).

VBM: voxel based morphometry. GMV: gray matter volume. SCZ: schizophrenia. BD: bipolar disorder. HC: healthy controls. cFWE: cluster family wise error.

https://doi.org/10.1371/journal.pone.0188000.g001

thumbnail
Table 2. Statistics of the significant ROIs (p<0.05, cFWE corrected) of the VBM analysis of Dataset1.

Degrees of freedom: [1 785]. t-stat threshold: 3.1.

https://doi.org/10.1371/journal.pone.0188000.t002

Compared to HC, SCZ patients showed widespread GMV reductions throughout the cortex (p<0.05, cFWE corrected). The GMV deficits were detected in the frontal (anterior and mid cingulate cortex, rectus, frontal superior medial and mid frontal cortex), parietal (inferior parietal cortex, supramarginal and postcentral gyri), occipital (cuneus and calcarine cortex) and temporal cortices (mid and superior temporal cortex and pole, rolandic operculum, Heschl’s gyrus, insula) of both hemispheres. An extensive GMV reduction also spanned bilaterally the thalamus, hippocampus, parahippocampal gyrus and amygdala, as well as portions of the cerebellum. The most significant voxel was in the left superior temporal cortex.

The conjunction analysis revealed shared GMV deficits in SCZ and BD patients compared to HC (p<0.05, cFWE corrected) in three clusters located in the superior temporal cortex, anterior and mid cingulate cortex and occipital cortex of the right hemisphere. Although the areas of GMV deficit of SCZ patients mostly included the BD ones, the subgenual portion of the right cingulate cortex emerged only in BD patients.

The comparison between BD and SCZ revealed significant GMV differences in the bilateral thalami and in the right insula, which showed GMV deficits in SCZ patients compared to BD patients (p<0.05, cFWE corrected).

3.2.2. ROI analysis.

The AAL regions with global GMV difference between the different groups (SCZ vs. HC, BD vs. HC, SCZ vs. BD, SCZ and BD vs. HC) are reported in Table 3 and shown in Fig 2 (p<0.05, Bonferroni corrected). Overall, at such conservative significance threshold, the local GMV abnormalities were confirmed at the regional level in SCZ patients but not in BD ones.

thumbnail
Fig 2. Results of the ROI analysis of Dataset1.

AAL regions with significant GMV differences between SCZ, BD and HC (p<0.05, Bonferroni corrected). ROI: region of interest. AAL: Automated Anatomical Labeling. GMV: gray matter volume. SCZ: schizophrenia. BD: bipolar disorder. HC: healthy controls.

https://doi.org/10.1371/journal.pone.0188000.g002

thumbnail
Table 3. Significant AAL regions of the ROI analysis of Dataset1 (p<0.05, Bonferroni corrected).

https://doi.org/10.1371/journal.pone.0188000.t003

Indeed, the local GMV deficits in SCZ compared to HC (VBM t-contrast: SCZ<HC, p<0.05, cFWE corrected) were confirmed at the regional level with a high level of significance (p<0.05, Bonferroni corrected) in the majority of AAL regions, 57 out of 88, with the most significant deficits in the bilateral rolandic operculum, insula, Heschl’s gyrus and superior temporal cortex, and in the left supramarginal gyrus and triangular inferior frontal cortex.

With regard to the GMV deficits in BD compared to HC (VBM t-contrast: BD<HC, p<0.05, cFWE corrected), the ROI analysis did not confirm the VBM results at the regional level in any of the regions. Accordingly, the shared GMV deficits of BD and SCZ compared to HC were not confirmed at the regional level.

Among the clusters showing lower GMV in SCZ compared to BD (VBM t-contrast: SCZ<BD, p<0.05, cFWE corrected), the ROI analysis confirmed the VBM results in the right rolandic operculum, right insula and left thalamus (p<0.05, Bonferroni corrected).

3.3. Clinical-neuroanatomical correlations

The results of the neuroanatomical-clinical correlation analyses are reported in Table 4 (p<0.001, >100 voxels).

thumbnail
Table 4. Statistics of the VBM analysis of clinical variables.

Degrees of freedom: [1 785]. t-stat threshold: 3.1.

https://doi.org/10.1371/journal.pone.0188000.t004

In BD patients, the contributions of HRSD scores, BD type, psychotic features and antipsychotic therapy on GMV were not significant. However, we found a positive association between age of onset and GMV in two clusters belonging to right superior frontal cortex and left mid temporal cortex. We also found neurotrophic effects of mood stabilizers, which were associated with higher GMV in thalamus and cerebellum in the left hemisphere. On the contrary, BD patients treated with benzodiazepines showed reduced GMV in bilateral frontal cortex (medial portion in the left hemisphere, inferior opercular portion in the right one), right calcarine cortex and left supplementary motor cortex. The ROI analysis confirmed this association at the regional level in right inferior frontal cortex (opercular portion) and right calcarine cortex (p<0.01).

In SCZ patients, no significant effects of BPRS scores, therapy with benzodiazepines, duration of untreated illness and age of onset emerged. In the subset of patients scored with PANSS, we did not find any GMV correlates of the scale scores. However, in SCZ patients from Jena, we found a negative association between SANS scores and GMV in left superior temporal pole. In patients from Santander, a negative relation emerged between SAPS scores and GMV in left inferior parietal cortex. These results were not confirmed at the regional level.

4. Discussion

In the present work, we integrated and quantitatively analyzed a very large sample of structural MRI data of HC and patients with SCZ and BD. Our ultimate objective was to obtain a robust large-scale identification of the structural cerebral differences associated with the two disorders, overcoming the limitations related to low sample size and statistical power that affect most of the current studies on SCZ and BD.

Our results show that the partial overlap between BD and SCZ in terms of clinical phenotype may rely on shared neuroanatomical changes, supporting the hypothesis of common brain structure endophenotypes across categorical diagnoses [26]. However, our findings also show points of difference across the two diagnoses and demonstrate the greater extent of the SCZ pathology, associated with brain structures that are more severely compromised in comparison with BD. The increased volumetric deficits in SCZ patients may be related to the more severe clinical picture and cognitive impairment currently seen in SCZ than in BD [2729]. Although clinical manifestations and cognitive deficits can be significant in BD too, in our BD sample the prevalence of non-psychotic patients may have contributed to widen the GMV differences between BD and SCZ.

4.1. Gray matter phenotypes of SCZ and BD

The analyses conducted in this multicentric study revealed overlapping volume deficits in the two groups of patients, being more severe and generalized in SCZ. The VBM results on Dataset1 from the SCZ-BD conjunction analysis showed in the two patient groups compared to HC a GM atrophy in superior temporal cortex and temporal pole, anterior and mid cingulate cortex and calcarine cortex of the right hemisphere; the shared GMV reduction in right temporal cortex, more pronounced in SCZ, was confirmed by the analyses on Dataset2. In the latter analysis, the emergence of GMV deficits in BD only at the uncorrected significance level may be the effect of the removal of many older patients from the BD sample. Overall, our findings are consistent with the ones of previous studies, showing in SCZ and BD overlapping deficits in cingulate cortex and temporal lobe [8, 14].

Previous studies using different MRI contrasts have confirmed the key role of these regions in these disorders. Reduced WM volume, fractional anisotropy and GM complexity in frontal and temporo-parietal regions emerged in both SCZ and BD [15, 3035]. The cortical thickness, cortical volume, water diffusion and blood perfusion in fronto-temporo-limbic regions have also emerged as relevant features for classification of first episode psychosis (FEP) [3638]. Functional MRI studies reported dysfunctional superior temporal gyrus and disrupted fronto-temporal connectivity in patients with FEP and in ultra-high-risk subjects [39, 40]. Furthermore, the dorsal anterior cingulate cortex represents a key node of the cingulo-opercular network, exerting top-down control over sensory and limbic regions, which emerged to be disconnected in patients with SCZ [41] and BD [42]. More in general, global structural and functional connectivity metrics have been shown to be abnormal in psychotic illnesses [43].

4.2. Specific characteristics of SCZ

In line with previous studies [8, 12, 44], our VBM and ROI findings confirm the larger extent and magnitude of GMV deficits in SCZ patients than in BD patients when compared to HC, with the regions of GMV reduction in BD largely included in the SCZ ones.

The VBM analyses of both datasets revealed widespread deficits of SCZ patients, which appeared to be highly significant and symmetrically distributed across the two hemispheres. The ROI analyses strengthened the above results, providing evidence for the consistence of such deficits at the regional level in SCZ but not in BD. While the local GMV deficits associated with BD were mainly in the right hemisphere, the SCZ ones spread to the left hemisphere. The right frontal and temporal areas showing lower GMV in SCZ compared to HC were more extended than BD ones, including the rolandic operculum, Heschl’s gyrus, insula, dorsolateral prefrontal cortex and wider portions of the cingulate cortex. Noticeably, in SCZ the left temporal deficits were more significant than the right ones: this finding, besides being supported by wide literature evidence [10, 45, 46], may be related to the key role of left temporal cortex in auditory information processing, hallucinations and thought disorders.

The VBM analyses also highlighted thalamic, hippocampal and amygdalar deficits in SCZ patients but not in BD ones. The anatomical abnormalities of these structures in SCZ patients appear to be consistent with their extensive cognitive deficits [28, 47, 48]. In line with the hypothesis of “cognitive dysmetria”, the GMV reduction in key nodes of information processing may underlay the deficits in integrating and coordinating information of SCZ [49]. The VBM and ROI results on Dataset2 confirmed the extended volumetric deficits characterizing SCZ compared to HC, involving both cortical and subcortical regions.

The areas of GMV reduction in SCZ highlighted in our study were consistently reported in previous voxel-based [8, 10, 15, 44, 50, 51] and region-based studies, showing GM changes in, for example, anterior cingulate cortex [52], superior temporal gyrus [53], insula [54], entorhinal cortex [55] and amygdala [47]. Changes in WM have also been reported in SCZ patients [27], mainly in frontal and temporo-occipital areas [46, 56], corpus callosum [46, 57, 58] and thalamus [59].

4.3. Specific characteristics of BD

Despite the wider deficits characterizing SCZ in comparison with HC, the VBM analysis of Dataset1 highlighted a specific deficit of BD patients compared to HC in a portion of the right anterior cingulate cortex, crossing the perigenual and subgenual areas. This specific deficit of BD was already described [8] and is in line with postmortem studies on BD, showing in this region decreased neural density and thickness of cortical layers III, V and VI [60]. Since the subgenual anterior cingulate cortex is implicated in the processing and modulation of emotions, its specific abnormality in BD patients may be interpreted on the light of the emotional dysregulation that characterizes mood disorders [61, 62]. Regional GMV deficits in these areas have been associated with a personal history of attempted suicide in BD [63] and are ameliorated by lifetime lithium treatment [64]. Although the relationship between neural and behavioral alterations in BD needs further investigation [65], the altered control of prefrontal cortex over limbic structures may trigger the emotional hyper-reactivity in BD patients [66].

The deficit in right anterior cingulate cortex resulting from the VBM analysis was not confirmed at the regional level by the ROI analysis on Dataset1, probably due to the extension of the anterior cingulate region in the AAL atlas, which did not differentiate among its subregions. It should be mentioned that diverse findings came from previous region-based studies, which showed no GMV deficits in bilateral subgenual prefrontal cortex [67] and in right anterior cingulate cortex [68] in BD compared with HC. However, these results emerged from smaller patient samples compared to the present work.

It is worth noticing that our VBM analyses showed a right lateralization of the GMV deficits in BD patients, which is in line with the findings of previous structural [11] and functional [69] studies. This result is supported by the knowledge that right prefrontal cortex plays a key role in cognitive inhibition [70] and is corroborated by evidence that transcranial magnetic stimulation of this region represents an effective add-on treatment for BD [71, 72]. Such an asymmetry may also reflect the lower WM integrity in prefrontal-limbic, limbic and callosal tracts observed in BD compared with HC [30, 7375]. However, Bellani and colleagues [76] investigated laterality effects in SCZ and BD compared with HC using a visuo-motor task, without finding any difference between BD and HC. Therefore, the issue of laterality in BD should be further investigated.

4.4. Differences between BD and SCZ

One of the main novelties of the present study concerns the direct comparison between SCZ and BD. Significant inter-diagnostic differences emerged only from the analysis of Dataset1, suggesting slighter differences between SCZ and BD patients than between each group of patients and HC. However, the VBM and ROI analyses of Dataset1 highlighted GMV deficits in SCZ compared to BD in the right insula and its operculum and in the thalami, especially in the left hemisphere.

So far, the limited number of studies that compared SCZ and BD, either directly or indirectly, usually found GMV deficits in SCZ compared to BD. Although the cortical regions interested by these deficits were rather heterogeneous across studies, our results partially confirm the previous ones.

In this respect, the volume losses in the insula and/or thalamus that we found in SCZ compared to BD have already been reported in the literature [3, 1315, 51]. In the work from Bora and colleagues [51], such a difference was found to be related to the male dominance in the SCZ group, which also characterized our and other datasets. However, the hypothesis of a specific involvement of the thalamus in the pathophysiology of SCZ has been recently emphasized in the literature [77], corroborated by postmortem evidence of reduction in volume and number of neurons of the medial dorsal thalamic nucleus, the principal source of thalamic projections to the prefrontal cortex [78]. Pharmacological effects cannot be ruled out, because antipsychotics seem to exert effects on thalamic volumes [79]. The structural differences between SCZ and BD in the thalamus may also be reflected in the lower thalamic function of SCZ compared to BD that was reported in [80].

The further hypothesis of a specific role of the insula in SCZ pathology is supported by the exaggerated thalamic-insular functional connectivity characterizing SCZ compared to HC [81]. The above deficits may also relate to the lower metabolism in frontal, parietal and temporal WM tracts that emerged in SCZ compared to BD [82].

Differently from our study, GMV reductions in SCZ compared to BD were also described in the right lingual gyrus [12], amygdala, putamen and hippocampus [2] as well as in the cerebellum [83].

4.5. Comparison with other multicentric studies

To our knowledge, this is the one of the first multicentric studies on SCZ and BD with such a large number of patients and combining ROI and VBM analyses.

A very recent multicentric study that deserves mention was conducted by Ivleva and colleagues [12], who compared GMV in a large sample of patients with SCZ, PBD and schizoaffective disorder (SAD), their first-degree relatives and HC. In their study, GMV was investigated at the voxel-level and in 5-mm radius spheres centered on the VBM peak voxels. In line with our results, they showed broader GMV deficits in SCZ than in BD when compared to HC, as well as GMV deficits in BD in the cingulate and superior temporal cortices, mainly in the right hemisphere. It is worth noticing that Ivleva et al. contrasted GMV across DSM diagnoses and psychosis biotypes, providing evidence for the meaningful classification of patients based on neurobiological constructs. This represents an interesting perspective that should be investigated in the next future, when a larger sample of PBD patients will be available.

Another study that deserves mention [84] involved 13 Centers and around 500 subjects, either SCZ patients or HC, and produced VBM and ROI results on SCZ that were overall in agreement with ours. However, compared to our study, it was limited by the absence of BD patients, besides being based on a smaller sample of subjects distributed among a larger number of centers, which made the results more susceptible to confounding center effects.

4.6. Methodological issues

Our results were obtained by applying both voxel-based and region-based analyses, whose integration represents one of the novelties of our study. Here, the ROI analysis was introduced to inspect the reproducibility of the VBM findings at the regional level. It should be remarked that the ROI analysis was VBM-driven, as the selection of the AAL regions was based on VBM results, and that VBM and ROI results were not fully independent, as the ROI volume was computed on the same tissue density maps of the VBM study. However, the two approaches provide different information and have complementary advantages and disadvantages. In region-based approaches, the results reliability depends on the choice of the regions, which can be either manual or automatic, but the results are less sensitive to local imperfections than in voxel-based approaches. Therefore, in our study, the combination of VBM and ROI analyses allowed to investigate both local and global GMV differences and to provide more robust evidence of the neuroanatomical bases of BD and SCZ.

The choice to include the analysis of Dataset2, a subset of Dataset1, was motivated by the inhomogeneous characteristics of Dataset1. Although the GLM design modeled and removed the contribution of age, gender and center, the comparison between the results of Dataset1 and Dataset2 allowed to identify the most relevant results. The findings of Dataset2 were less significant compared to the ones of Dataset1 due to the smaller sample, but less vulnerable to confounding factors. A more robust validation may come from the analysis of an independent sample, which is encouraged in the next future.

The VBM and ROI results on both datasets showed an overall agreement, which demonstrates not only the reliability of the findings on the two datasets, but also the validity of the quantitative techniques that were applied. In both datasets, the site represented a relevant contributor: the different imaging parameters, types of MR scanner and head coils led to images with different spatial resolution, contrast and intensity, whose contribution was considered in the analysis.

In the main study design, we did not account for clinical information such as age of illness, psychopathological scores, pharmacological treatment and substance abuse, which may have an influence on the brain structural characteristics. This choice was motivated by the heterogeneity of available information and by the variety of treatments across the four centers. However, we investigated the contribution of a set of clinical variables on GMV in subsamples of patients, according to the available information. The results concerning the BD sample suggest that mood state, BD type and psychotic symptoms should have not influenced the results. However, we cannot exclude possible confounding effects due to 1) age of onset and medication in BD patients, 2) symptoms severity in SCZ patients. These limitations should be taken into account when interpreting the current results.

4.7. Future perspectives

The results of this large neuroimaging study demonstrate the importance of comparing BD and SCZ (between themselves and with HC) to gain new knowledge of the shared and specific neuroanatomical bases of the two disorders, bringing added value to the clinical management of BD and SCZ.

However, our results need to be confirmed by future investigations on large independent samples and should be integrated with the study of gender and age effects in both BD and SCZ. On the one hand, the investigation of brain sexual dimorphism in BD and SCZ can shed new light on the relationship between sexual differentiation processes and vulnerability to develop psychiatric disorders. On the other hand, the analysis of age-related neuroanatomical changes has potential to highlight any deviations from normal brain maturation shared by BD and SCZ or diagnosis-specific. These complex topics will be the object of a future dedicated investigation from the ENPACT group.

In the near future, we are interested in repeating the SCZ-BD comparison by using a larger BD sample, with a sufficient and balanced number of psychotic and non-psychotic patients, with the objective to identify any neurobiological characteristics associated with psychotic features across traditionally distinct categorical diagnoses.

5. Conclusions

Although SCZ and BD are conceptualized in a continuum of clinical phenotypes, the underlying neural mechanisms are still largely unknown. In the present study, structural MRI images of a large sample of subjects were analyzed with voxel-based and region-based techniques to identify GMV differences and similarities between SCZ and BD. The two disorders exhibited shared fronto-temporo-occipital GMV deficits in the right hemisphere, which support the hypothesis of a continuous GM endophenotype for SCZ and BD. Our study demonstrates that the two disorders are not completely dichotomous in terms of GM anatomy, but also suggest that SCZ is associated with brain structures that are more severely compromised in comparison with BD. A specific involvement of thalamus and insular cortex in SCZ is suggested, which needs further investigation. The results of our study provide a key piece of information for the comprehensive understanding of the two disorders, opening the door to advanced diagnostic and treatment strategies.

Supporting information

S1 Results. Results of the VBM analysis of Dataset2.

https://doi.org/10.1371/journal.pone.0188000.s001

(DOCX)

S1 Database. Database information.

Demographic and clinical information on Dataset1 and Dataset2.

https://doi.org/10.1371/journal.pone.0188000.s002

(XLSX)

S1 Fig.

Left: Results of the VBM analysis of Dataset2. Significant regions emerging from the SCZ<HC t-contrast (p<0.05, cFWE corrected). Right: Results of the ROI analysis of Dataset2. AAL regions with significant GMV differences between SCZ and HC (p<0.05, Bonferroni corrected). VBM: voxel based morphometry. ROI: region of interest. AAL: Automated Anatomical Labeling. GMV: gray matter volume. SCZ: schizophrenia. HC: healthy controls. cFWE: cluster family wise error.

https://doi.org/10.1371/journal.pone.0188000.s003

(TIFF)

S1 Table. Statistics of the significant ROIs (p<0.05, cFWE corrected) of the VBM analysis of Dataset2.

Degrees of freedom: [1 284]. t-stat threshold: 3.12. SCZ: schizophrenia. HC: healthy controls. FWE: family wise error. AAL: Automated Anatomical Labeling. R: right hemisphere. L: left hemisphere.

https://doi.org/10.1371/journal.pone.0188000.s004

(DOCX)

S2 Table. Significant AAL regions of the region-based analysis of Dataset2 (p<0.05, Bonferroni corrected).

SCZ: schizophrenia. HC: healthy controls. AAL: Automated Anatomical Labeling.

https://doi.org/10.1371/journal.pone.0188000.s005

(DOCX)

Acknowledgments

Thanks to Gianluca Rambaldelli for helping in managing the dataset. PB was partially funded by grants from the Ministry of Health (RF-2011-02352308). Grant support of EM was provided by the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement no. 602450 (IMAGEMEND Project). Part of the present study was conducted at the Hospital Universitario Marqués de Valdecilla, University of Cantabria (Santander, Spain), under the following grant support: Carlos III Health Institute PI020499, PI050427, PI060507, Plan Nacional de Drugs Research Grant 2005- Orden sco/3246/2004, SENY Fundació Research Grant CI 2005–0308007 and Fundación Marqués de Valdecilla API07/011. We wish to acknowledge IDIVAL Neuroimaging Unit for imaging acquirement and analysis. Part of the study was conducted at the Ospedale San Raffaele, Milano, supported by the European Union EU-FP7-HEALTH-F2-2008-222963 “MOODINFLAME” and by the Italian Ministry of Health RF-2011-02350980 projects. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

ENPACT group:

Leader: Prof. Paolo Brambilla, University of Milan (paolo.brambilla1@unimi.it).

Participants:

JUH: Igor Nenadic, Christian Gaser, Heinrich Sauer.

UHMV: Benedicto Crespo-Facorro, Roberto Roiz-Santiañez.

UVSSR: Francesco Benedetti, Sara Poletti.

VUH—UNIMI: Eleonora Maggioni, Veronica Marinelli, Marcella Bellani, Cinzia Perlini, Mirella Ruggeri, Cecilia Prunas, Gian Mario Mandolini, A. Carlo Altamura, Vaibhav A. Diwadkar, Paolo Brambilla.

References

  1. 1. Arnone D, Cavanagh J, Gerber D, Lawrie SM, Ebmeier KP, McIntosh AM. Magnetic resonance imaging studies in bipolar disorder and schizophrenia: meta-analysis. Br J Psychiatry. 2009;195(3):194–201. pmid:19721106.
  2. 2. Brown GG, Lee JS, Strigo IA, Caligiuri MP, Meloy MJ, Lohr J. Voxel-based morphometry of patients with schizophrenia or bipolar I disorder: a matched control study. Psychiatry Res. 2011;194(2):149–56. pmid:21924872; PubMed Central PMCID: PMCPMC3196272.
  3. 3. Nenadic I, Maitra R, Langbein K, Dietzek M, Lorenz C, Smesny S, et al. Brain structure in schizophrenia vs. psychotic bipolar I disorder: A VBM study. Schizophr Res. 2015;165(2–3):212–9. pmid:25935815.
  4. 4. Lichtenstein P, Yip BH, Bjork C, Pawitan Y, Cannon TD, Sullivan PF, et al. Common genetic determinants of schizophrenia and bipolar disorder in Swedish families: a population-based study. Lancet. 2009;373(9659):234–9. pmid:19150704; PubMed Central PMCID: PMCPMC3879718.
  5. 5. International Schizophrenia C, Purcell SM, Wray NR, Stone JL, Visscher PM, O'Donovan MC, et al. Common polygenic variation contributes to risk of schizophrenia and bipolar disorder. Nature. 2009;460(7256):748–52. pmid:19571811; PubMed Central PMCID: PMCPMC3912837.
  6. 6. Kim MD, Seo HJ, Yun HJ, Jung YE, Park JH, Lee CI, et al. The Relationship between Cognitive Decline and Psychopathology in Patients with Schizophrenia and Bipolar Disorder. Clin Psychopharmacol Neurosci. 2015;13(1):103–8. pmid:25912543; PubMed Central PMCID: PMCPMC4423162.
  7. 7. Fischer BA, Carpenter WT Jr. Will the Kraepelinian dichotomy survive DSM-V? Neuropsychopharmacology. 2009;34(9):2081–7. pmid:19295511; PubMed Central PMCID: PMCPMC2721021.
  8. 8. Ellison-Wright I, Bullmore E. Anatomy of bipolar disorder and schizophrenia: a meta-analysis. Schizophr Res. 2010;117(1):1–12. pmid:20071149.
  9. 9. Wright IC, Rabe-Hesketh S, Woodruff PW, David AS, Murray RM, Bullmore ET. Meta-analysis of regional brain volumes in schizophrenia. Am J Psychiatry. 2000;157(1):16–25. pmid:10618008.
  10. 10. Honea R, Crow TJ, Passingham D, Mackay CE. Regional deficits in brain volume in schizophrenia: a meta-analysis of voxel-based morphometry studies. Am J Psychiatry. 2005;162(12):2233–45. pmid:16330585.
  11. 11. Selvaraj S, Arnone D, Job D, Stanfield A, Farrow TF, Nugent AC, et al. Grey matter differences in bipolar disorder: a meta-analysis of voxel-based morphometry studies. Bipolar Disord. 2012;14(2):135–45. pmid:22420589.
  12. 12. Ivleva EI, Clementz BA, Dutcher AM, Arnold SJ, Jeon-Slaughter H, Aslan S, et al. Brain Structure Biomarkers in the Psychosis Biotypes: Findings From the Bipolar-Schizophrenia Network for Intermediate Phenotypes. Biol Psychiatry. 2016. pmid:27817844.
  13. 13. Ivleva EI, Bidesi AS, Thomas BP, Meda SA, Francis A, Moates AF, et al. Brain gray matter phenotypes across the psychosis dimension. Psychiatry Res. 2012;204(1):13–24. pmid:23177922; PubMed Central PMCID: PMCPMC3589584.
  14. 14. Ivleva EI, Bidesi AS, Keshavan MS, Pearlson GD, Meda SA, Dodig D, et al. Gray matter volume as an intermediate phenotype for psychosis: Bipolar-Schizophrenia Network on Intermediate Phenotypes (B-SNIP). Am J Psychiatry. 2013;170(11):1285–96. pmid:24185241.
  15. 15. McDonald C, Bullmore E, Sham P, Chitnis X, Suckling J, MacCabe J, et al. Regional volume deviations of brain structure in schizophrenia and psychotic bipolar disorder: computational morphometry study. Br J Psychiatry. 2005;186:369–77. pmid:15863740.
  16. 16. Molina V, Galindo G, Cortes B, de Herrera AG, Ledo A, Sanz J, et al. Different gray matter patterns in chronic schizophrenia and chronic bipolar disorder patients identified using voxel-based morphometry. Eur Arch Psychiatry Clin Neurosci. 2011;261(5):313–22. pmid:21188405.
  17. 17. Maggioni E, Bellani M, Altamura AC, Brambilla P. Neuroanatomical voxel-based profile of schizophrenia and bipolar disorder. Epidemiol Psychiatr Sci. 2016;25(4):312–6. pmid:27095442.
  18. 18. First MB, Gibbon M. The Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I) and the Structured Clinical Interview for DSM-IV Axis II Disorders (SCID-II). 2004.
  19. 19. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry. 1960;23:56–62. pmid:14399272; PubMed Central PMCID: PMCPMC495331.
  20. 20. Young RC, Biggs JT, Ziegler VE, Meyer DA. A rating scale for mania: reliability, validity and sensitivity. Br J Psychiatry. 1978;133:429–35. pmid:728692.
  21. 21. Fischl B. FreeSurfer. Neuroimage. 2012;62(2):774–81. pmid:22248573; PubMed Central PMCID: PMCPMC3685476.
  22. 22. Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O, Delcroix N, et al. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage. 2002;15(1):273–89. pmid:11771995.
  23. 23. Picchioni MM, Murray RM. Schizophrenia. BMJ. 2007;335(7610):91–5. pmid:17626963; PubMed Central PMCID: PMCPMC1914490.
  24. 24. Leucht S, Kane JM, Kissling W, Hamann J, Etschel E, Engel RR. What does the PANSS mean? Schizophr Res. 2005;79(2–3):231–8. pmid:15982856.
  25. 25. Leucht S, Kane JM, Kissling W, Hamann J, Etschel E, Engel R. Clinical implications of Brief Psychiatric Rating Scale scores. Br J Psychiatry. 2005;187:366–71. pmid:16199797.
  26. 26. Thaker GK. Neurophysiological endophenotypes across bipolar and schizophrenia psychosis. Schizophr Bull. 2008;34(4):760–73. pmid:18502737; PubMed Central PMCID: PMCPMC2632457.
  27. 27. Brambilla P, Tansella M. The role of white matter for the pathophysiology of schizophrenia. Int Rev Psychiatry. 2007;19(4):459–68. pmid:17671878.
  28. 28. Brambilla P, Perlini C, Rajagopalan P, Saharan P, Rambaldelli G, Bellani M, et al. Schizophrenia severity, social functioning and hippocampal neuroanatomy: three-dimensional mapping study. Br J Psychiatry. 2013;202(1):50–5. pmid:23284150; PubMed Central PMCID: PMCPMC4113031.
  29. 29. Daban C, Martinez-Aran A, Torrent C, Tabares-Seisdedos R, Balanza-Martinez V, Salazar-Fraile J, et al. Specificity of cognitive deficits in bipolar disorder versus schizophrenia. A systematic review. Psychother Psychosom. 2006;75(2):72–84. pmid:16508342.
  30. 30. Brambilla P, Bellani M, Yeh PH, Soares JC, Tansella M. White matter connectivity in bipolar disorder. Int Rev Psychiatry. 2009;21(4):380–6. pmid:20374151.
  31. 31. Bruno SD, Barker GJ, Cercignani M, Symms M, Ron MA. A study of bipolar disorder using magnetization transfer imaging and voxel-based morphometry. Brain. 2004;127(Pt 11):2433–40. pmid:15469950.
  32. 32. Bruno S, Cercignani M, Ron MA. White matter abnormalities in bipolar disorder: a voxel-based diffusion tensor imaging study. Bipolar Disord. 2008;10(4):460–8. pmid:18452442.
  33. 33. Haznedar MM, Roversi F, Pallanti S, Baldini-Rossi N, Schnur DB, Licalzi EM, et al. Fronto-thalamo-striatal gray and white matter volumes and anisotropy of their connections in bipolar spectrum illnesses. Biol Psychiatry. 2005;57(7):733–42. pmid:15820230.
  34. 34. Hao Y, Liu Z, Jiang T, Gong G, Liu H, Tan L, et al. White matter integrity of the whole brain is disrupted in first-episode schizophrenia. Neuroreport. 2006;17(1):23–6. pmid:16361944.
  35. 35. Squarcina L, De Luca A, Bellani M, Brambilla P, Turkheimer FE, Bertoldo A. Fractal analysis of MRI data for the characterization of patients with schizophrenia and bipolar disorder. Phys Med Biol. 2015;60(4):1697–716. pmid:25633275.
  36. 36. Squarcina L, Perlini C, Peruzzo D, Castellani U, Marinelli V, Bellani M, et al. The use of dynamic susceptibility contrast (DSC) MRI to automatically classify patients with first episode psychosis. Schizophrenia research. 2015;165(1):38–44. pmid:25888338
  37. 37. Squarcina L, Castellani U, Bellani M, Perlini C, Lasalvia A, Dusi N, et al. Classification of first-episode psychosis in a large cohort of patients using support vector machine and multiple kernel learning techniques. NeuroImage. 2015.
  38. 38. Peruzzo D, Castellani U, Perlini C, Bellani M, Marinelli V, Rambaldelli G, et al. Classification of first-episode psychosis: a multi-modal multi-feature approach integrating structural and diffusion imaging. Journal of Neural Transmission. 2015;122(6):897–905. pmid:25344845
  39. 39. Crossley NA, Mechelli A, Fusar‐Poli P, Broome MR, Matthiasson P, Johns LC, et al. Superior temporal lobe dysfunction and frontotemporal dysconnectivity in subjects at risk of psychosis and in first‐episode psychosis. Human brain mapping. 2009;30(12):4129–37. pmid:19530219
  40. 40. Yoon YB, Yun J-Y, Jung WH, Cho KIK, Kim SN, Lee TY, et al. Altered fronto-temporal functional connectivity in individuals at ultra-high-risk of developing psychosis. PloS one. 2015;10(8):e0135347. pmid:26267069
  41. 41. Tu P-C, Hsieh J-C, Li C-T, Bai Y-M, Su T-P. Cortico-striatal disconnection within the cingulo-opercular network in schizophrenia revealed by intrinsic functional connectivity analysis: a resting fMRI study. Neuroimage. 2012;59(1):238–47. pmid:21840407
  42. 42. Mamah D, Barch DM, Repovš G. Resting state functional connectivity of five neural networks in bipolar disorder and schizophrenia. Journal of affective disorders. 2013;150(2):601–9. pmid:23489402
  43. 43. O'Donoghue S, Cannon D, Perlini C, Brambilla P, McDonald C. Applying neuroimaging to detect neuroanatomical dysconnectivity in psychosis. Epidemiology and psychiatric sciences. 2015;24(04):298–302.
  44. 44. Yu K, Cheung C, Leung M, Li Q, Chua S, McAlonan G. Are bipolar disorder and schizophrenia neuroanatomically distinct? An anatomical likelihood meta-analysis. Frontiers in human neuroscience. 2010;4:189. pmid:21103008
  45. 45. Perlini C, Bellani M, Brambilla P. Structural imaging techniques in schizophrenia. Acta Psychiatrica Scandinavica. 2012;126(4):235–42. pmid:22533735
  46. 46. Bellani M, Marzi CA, Brambilla P. Interhemispheric communication in schizophrenia. Epidemiologia e Psichiatria Sociale. 2009;18(01):19–22.
  47. 47. Tomasino B, Bellani M, Perlini C, Rambaldelli G, Cerini R, Isola M, et al. Altered microstructure integrity of the amygdala in schizophrenia: a bimodal MRI and DWI study. Psychological medicine. 2011;41(02):301–11.
  48. 48. Andreasen NC, Arndt S, Swayze V, Cizadlo T, Flaum M, O’Leary D, et al. Thalamic abnormalities in schizophrenia visualized through magnetic resonance image averaging. SCIENCE-NEW YORK THEN WASHINGTON-. 1994:294-.
  49. 49. Andreasen NC, Paradiso S, O'leary DS. “Cognitive dysmetria” as an integrative theory of schizophrenia: a dysfunction in cortical-subcortical-cerebellar circuitry? Schizophrenia Bulletin. 1998;24(2):203–18. pmid:9613621
  50. 50. Tomelleri L, Jogia J, Perlini C, Bellani M, Ferro A, Rambaldelli G, et al. Brain structural changes associated with chronicity and antipsychotic treatment in schizophrenia. European Neuropsychopharmacology. 2009;19(12):835–40. pmid:19717283
  51. 51. Bora E, Fornito A, Yücel M, Pantelis C. The effects of gender on grey matter abnormalities in major psychoses: a comparative voxelwise meta-analysis of schizophrenia and bipolar disorder. Psychological medicine. 2012;42(2):295. pmid:21835091
  52. 52. Baiano M, David A, Versace A, Churchill R, Balestrieri M, Brambilla P. Anterior cingulate volumes in schizophrenia: a systematic review and a meta-analysis of MRI studies. Schizophrenia research. 2007;93(1):1–12.
  53. 53. Tosato S, Bellani M, Bonetto C, Ruggeri M, Perlini C, Lasalvia A, et al. Is neuregulin 1 involved in determining cerebral volumes in schizophrenia? Preliminary results showing a decrease in superior temporal gyrus volume. Neuropsychobiology. 2012;65(3):119–25. pmid:22378022
  54. 54. Crespo-Facorro B, Kim J-J, Andreasen NC, O'Leary DS, Bockholt HJ, Magnotta V. Insular cortex abnormalities in schizophrenia: a structural magnetic resonance imaging study of first-episode patients. Schizophrenia research. 2000;46(1):35–43. pmid:11099884
  55. 55. Baiano M, Perlini C, Rambaldelli G, Cerini R, Dusi N, Bellani M, et al. Decreased entorhinal cortex volumes in schizophrenia. Schizophrenia research. 2008;102(1):171–80.
  56. 56. Andreone N, Tansella M, Cerini R, Versace A, Rambaldelli G, Perlini C, et al. Cortical white-matter microstructure in schizophrenia. The British Journal of Psychiatry. 2007;191(2):113–9.
  57. 57. Brambilla P, Cerini R, Gasparini A, Versace A, Andreone N, Vittorini E, et al. Investigation of corpus callosum in schizophrenia with diffusion imaging. Schizophrenia Research. 2005;79(2):201–10.
  58. 58. Diwadkar VA, DeBellis MD, Sweeney JA, Pettegrew JW, Keshavan MS. Abnormalities in MRI-measured signal intensity in the corpus callosum in schizophrenia. Schizophrenia research. 2004;67(2):277–82.
  59. 59. Agarwal N, Rambaldelli G, PsychD CP, Dusi N, Kitis O, Bellani M, et al. Microstructural thalamic changes in schizophrenia: a combined anatomic and diffusion weighted magnetic resonance imaging study. Journal of psychiatry & neuroscience: JPN. 2008;33(5):440.
  60. 60. Bouras C, Kövari E, Hof PR, Riederer BM, Giannakopoulos P. Anterior cingulate cortex pathology in schizophrenia and bipolar disorder. Acta neuropathologica. 2001;102(4):373–9. pmid:11603813
  61. 61. Delvecchio G, Sugranyes G, Frangou S. Evidence of diagnostic specificity in the neural correlates of facial affect processing in bipolar disorder and schizophrenia: a meta-analysis of functional imaging studies. Psychological Medicine. 2013;43(03):553–69.
  62. 62. Drevets WC, Savitz J, Trimble M. The subgenual anterior cingulate cortex in mood disorders. CNS spectrums. 2008;13(8):663. pmid:18704022
  63. 63. Benedetti F, Radaelli D, Poletti S, Locatelli C, Falini A, Colombo C, et al. Opposite effects of suicidality and lithium on gray matter volumes in bipolar depression. Journal of affective disorders. 2011;135(1):139–47.
  64. 64. Benedetti F, Bollettini I, Barberi I, Radaelli D, Poletti S, Locatelli C, et al. Lithium and GSK3-β promoter gene variants influence white matter microstructure in bipolar disorder. Neuropsychopharmacology. 2013;38(2):313–27. pmid:22990942
  65. 65. Wessa M, Perlini C, Brambilla P. Neuropsychological underpinnings of the dynamics of bipolar disorder. Epidemiology and psychiatric sciences. 2015;24(06):479–83.
  66. 66. Houenou J, Perlini C, Brambilla P. Epidemiological and clinical aspects will guide the neuroimaging research in bipolar disorder. Epidemiology and psychiatric sciences. 2015;24(02):117–20.
  67. 67. Brambilla P, Nicoletti MA, Harenski K, Sassi RB, Mallinger AG, Frank E, et al. Anatomical MRI study of subgenual prefrontal cortex in bipolar and unipolar subjects. Neuropsychopharmacology. 2002;27(5):792–9. pmid:12431853
  68. 68. Kaur S, Sassi RB, Axelson D, Nicoletti M, Brambilla P, Monkul ES, et al. Cingulate cortex anatomical abnormalities in children and adolescents with bipolar disorder. American Journal of Psychiatry. 2005;162(9):1637–43. pmid:16135622
  69. 69. Caligiuri MP, Brown GG, Meloy M, Eyler LT, Kindermann SS, Eberson S, et al. A functional magnetic resonance imaging study of cortical asymmetry in bipolar disorder. Bipolar Disorders. 2004;6(3):183–96. pmid:15117397
  70. 70. Aron AR, Robbins TW, Poldrack RA. Inhibition and the right inferior frontal cortex. Trends in cognitive sciences. 2004;8(4):170–7. pmid:15050513
  71. 71. Michael N, Erfurth A. Treatment of bipolar mania with right prefrontal rapid transcranial magnetic stimulation. Journal of affective disorders. 2004;78(3):253–7. pmid:15013251
  72. 72. Praharaj SK, Ram D, Arora M. Efficacy of high frequency (rapid) suprathreshold repetitive transcranial magnetic stimulation of right prefrontal cortex in bipolar mania: a randomized sham controlled study. Journal of affective disorders. 2009;117(3):146–50. pmid:19178948
  73. 73. Bellani M, Yeh P-H, Tansella M, Balestrieri M, Soares JC, Brambilla P. DTI studies of corpus callosum in bipolar disorder. Portland Press Limited; 2009.
  74. 74. Benedetti F, Yeh P-H, Bellani M, Radaelli D, Nicoletti MA, Poletti S, et al. Disruption of white matter integrity in bipolar depression as a possible structural marker of illness. Biological psychiatry. 2011;69(4):309–17. pmid:20926068
  75. 75. Sarrazin S, Poupon C, Linke J, Wessa M, Phillips M, Delavest M, et al. A multicenter tractography study of deep white matter tracts in bipolar I disorder: psychotic features and interhemispheric disconnectivity. JAMA psychiatry. 2014;71(4):388–96. pmid:24522197
  76. 76. Bellani M, Marzi CA, Savazzi S, Perlini C, Cerruti S, Ferro A, et al. Laterality effects in schizophrenia and bipolar disorder. Experimental brain research. 2010;201(2):339–44. pmid:19795112
  77. 77. Csernansky JG, Schindler MK, Splinter NR, Wang L, Gado M, Selemon LD, et al. Abnormalities of thalamic volume and shape in schizophrenia. American Journal of Psychiatry. 2004;161(5):896–902. pmid:15121656
  78. 78. Leung C. Neuropsychopharmacology: The fifth generation of progress. Hong Kong Journal of Psychiatry. 2003;13(3):31–3.
  79. 79. Clinton SM, Meador-Woodruff JH. Thalamic dysfunction in schizophrenia: neurochemical, neuropathological, and in vivo imaging abnormalities. Schizophrenia research. 2004;69(2):237–53.
  80. 80. Lui S, Yao L, Xiao Y, Keedy S, Reilly J, Keefe R, et al. Resting-state brain function in schizophrenia and psychotic bipolar probands and their first-degree relatives. Psychological medicine. 2015;45(01):97–108.
  81. 81. Corradi‐Dell'Acqua C, Tomelleri L, Bellani M, Rambaldelli G, Cerini R, Pozzi‐Mucelli R, et al. Thalamic‐insular dysconnectivity in schizophrenia: Evidence from structural equation modeling. Human brain mapping. 2012;33(3):740–52. pmid:21484952
  82. 82. Altamura AC, Bertoldo A, Marotta G, Paoli RA, Caletti E, Dragogna F, et al. White matter metabolism differentiates schizophrenia and bipolar disorder: a preliminary PET study. Psychiatry Research: Neuroimaging. 2013;214(3):410–4. pmid:24144506
  83. 83. Amann B, Canales‐Rodríguez E, Madre M, Radua J, Monte G, Alonso‐Lana S, et al. Brain structural changes in schizoaffective disorder compared to schizophrenia and bipolar disorder. Acta Psychiatrica Scandinavica. 2016;133(1):23–33. pmid:25968549
  84. 84. Segall JM, Turner JA, van Erp TG, White T, Bockholt HJ, Gollub RL, et al. Voxel-based morphometric multisite collaborative study on schizophrenia. Schizophrenia bulletin. 2009;35(1):82–95. pmid:18997157