Background

Bipolar disorder (BD) is a common psychiatric illness characterized by disruptive mood swings presented as episodes of mania/hypomania with or without depressive episodes, with normal mood in-between episodes [1]. It is characterized by affective instability and cognitive deficits [2].

The exact pathophysiology of this mental illness is still not fully understood. It has been proved that lesions in some specific neuroanatomical areas of the brain, e.g., stroke or brain tumors, may lead to the development of secondary mental illness. With great development in structural neuroimaging, recent neurological studies are increasingly concerned with studying the relationship between the anatomy and pathophysiology of mental diseases [3].

The most widely used clinical criteria for diagnosis of mania is from the American Psychiatric Association’s Diagnostic and Statistical Manual of mental disorders (DSM-IV), and the most commonly used criteria for grading of the severity of manic episode is the Young Mania Rating Scale (YMRS) [4].

Recent advances in multimodal neuroimaging contributed much to the understanding of the relation between structural brain abnormalities and behavioral disturbances [5]. Neuroimaging studies in bipolar disorder have documented structural and functional alterations in the brain, including gray matter morphology, white matter integrity, and local/global neuronal connectivity [6].

Brain morphometry is the study of size and shape of the brain and its structures, based on non-invasive imaging data obtained from magnetic resonance imaging (MRI). Voxel-based morphometry (VBM) [7, 8] is an unbiased automated computer-based technique that enables voxel-wise measurement of brain structures. The concept of VBM is based on three basic steps which include tissue classification, spatial normalization, and spatial smoothing, and the final step in a VBM analysis is to perform voxel-wise statistical tests.

Aim of the work

The aim of this study was to determine structural anatomical changes that could be associated with mood disorder bipolar I mania through detection of volumetric changes in different parts of the brain using a fully automated voxel-based morphometric analysis. The study also tried to detect any correlation between volumetric findings in patients with BD and different clinical aspects of the disease.

Methods

Ethics statement

The protocol, design, and procedures of this prospective study were approved by the institutional Ethical Committee Board. Informed written consents were obtained from the parents or the responsible patients’ legal guardians. Informed written consents were also obtained from the healthy control subjects who were included in the study. Confidentiality was maintained throughout the whole study.

Subjects

Patients

Twenty-four patients diagnosed with mood disorder bipolar “I” manic episodes were recruited. The patients were assessed by structured clinical interview for DSM-IV criteria. The severity of the manic episode was assessed using YMRS.

Controls

Carefully selected 16 healthy subjects matched for age, gender, socioeconomic status, and education were included as controls in the study. None of these subjects had neurological illness, and they were provisionally evaluated by experienced psychiatrists to exclude any psychiatric or neurologic disorder.

Methods

MRI & VBM evaluation

All MRI examinations were performed on a 1.5 Tesla MRI scanner (Achieva, Philips Medical Systems, Best, The Netherlands), using a 16-channel phased-array head coil. Following the multiplanar localizer sequence, the MR scanning protocol included routinely: axial FLAIR, axial diffusion-weighted, and axial 3D-T1W-FFE. The FLAIR and diffusion-weighted images were carefully inspected to exclude any gross anatomical abnormality or pathological changes in the brain. The high-resolution l-mm-thick 3D-T1W-FFE images were used for the voxel-based morphometry study. Demonstrative 1-mm 3D-TFE images of a patient and a control subject are shown in Figs. 1a and 2a, respectively.

Fig. 1
figure 1

a High-resolution 1-mm 3D-TFE images of an 18-year-old male patient, used for the voxel-based morphometric evaluation. b Volbrain electronically generated volumetric data in an 18-year-old male patient. c Volbrain-generated colored graphical maps of the different intracranial structures of an 18-year-old male patient

Full size image
Fig. 2
figure 2

a High-resolution 1-mm 3D-TFE images of a 42-year-old male control. b Volbrain electronically generated volumetric data in a 42-year-old male control. c Volbrain-generated colored graphical maps of the different intracranial structures of a 42-year-old male control

Full size image

For morphometric evaluation, the 3D-T1W-FFE images were converted from the standard digital imaging and communication (DICOM) format (.dcm) to the Neuroimaging Informatics Technology Initiative (NIfTI) format (.nii) suitable for analysis and post-processing. We used MRIcron software (Chris Rodren, 2014, Neuropsychology Lab, University of South Carolina, USA). MRIcron is a cross-platform NIfTI format image viewer which provides “dcm2nii tool” for converting DICOM images to NIfTI format. During the step of format conversion, the output .nii data were chosen to be generated in a compressed form.

In all the processes, DICOM data were anonymized to protect the privacy of the participants. The anonymization process was also performed using MRIcron.

The compressed .nii data were then uploaded through the internet to the online-based website of volBrain (https://volbrain.upv.es/) for analysis. volBrain is a fully automated segmentation pipeline developed by Manjón (Universidad Politécnica de Valencia, Spain) and Coupé (Université de Bordeaux, France) [9].

The technique of data processing, as provided by the pipeline developers, consists of a cascade of steps that include image de-noising, field and/or radiofrequency inhomogeneity correction, multiple-atlas label registration, intracranial cavity extraction, tissue classification, hemispheric segmentation, and finally, subcortical structures segmentation.

After completion of the analysis process, an electronic report was generated for each uploaded case containing the volumes of the main intracranial structures including the cerebrospinal fluid (CSF), cerebral gray matter (GM), and white matter (WM), as well as volume information of macroscopic areas such as the cerebral hemispheres, the cerebellum, and brainstem. The report also included the calculated volumes of different subcortical structures. Labeled colored graphical maps of the above-mentioned anatomical structures were also provided for each case.

Statistical analysis

We used t test to quantify the differences in the volumetric data of different parts of the brain between BD patients and the controls. We also used Spearman’s rank correlation coefficient within the BD patients’ group for detection of potential correlations between the volumetric data and the clinical parameters including the patient’s age, age of onset of BD, duration of illness, number of manic attacks, duration of the last attack, and YMRS as covariates.

Results

Demographic data

The study population included two groups of subjects; 24 patients with well-characterized bipolar disorder type I, and 16 healthy control subjects. In the patients’ group, there were 23 male patients and 1 female patient; their ages ranged from 18 to 47 years (Mean ± SD = 28.00 ± 9.09). The control group included 15 male subjects and 1 female; their age range was 19–46 years (Mean ± SD = 31.56 ± 7.30; Table 1).

Table 1 Demographic data of patients and healthy controls
Full size table

Clinical data

The age of onset of BD in the patients ranged from 12 to 46 years, with mean ± SD of 23.33 ± 8.34. The duration of illness ranged from 1 to 16 years, with mean ± SD of 4.87 ± 4.34. The number of manic episodes in patients ranged from 1 to 7 with mean ± SD of 3.08 ± 1.53. The duration of the last episode ranged from 2 to 22 h, with mean ± SD of 7.29 ± 5.41. The Young Mania Rating Scale of patients ranged from 17 to 45 with mean ± SD of 32.04 ± 6.97 (Table 2). Among the 24 patients of the study, 9 had positive history of substance abuse.

Table 2 Clinical data of patients of the study
Full size table

Volumetric MRI data

Volumetric differences between BD patients and healthy controls are presented in Table 3. The mean volume ± SD of the right globus pallidus was found significantly larger in the patients group than in the healthy controls (0.99 ± 0.19 and 0.84 ± 0.30, respectively; p = 0.03; Fig. 3). Although the mean volume ± SD of the left globus pallidus was found relatively larger in the patients’ group compared to the healthy controls, the difference was not significant (0.99 ± 0.20 and 0.84 ± 0.29, respectively; p = 0.07).

Table 3 Volumetric data of different brain structures in patients and controls
Full size table
Fig. 3
figure 3

Volumetric data of the globus pallidus in healthy patients and controls

Full size image

The calculated mean volumes of other subcortical structures, namely the caudate nucleus, putamen, thalamus, amygdale, and hippocampus were all relatively larger in the patients group than in the healthy controls, but the differences were statistically insignificant as follows:

The mean volume ± SD of both right and left caudate nuclei in the patients were 3.55 ± 0.46 and 3.58 ± 0.42, respectively, and in the healthy controls were 3.52 ± 0.52 and 3.44 ± 0.53, respectively; p = 0.83 for the right caudate and 0.31 for the left caudate.

The mean volume ± SD of both right and left putamen in the patients were 4.07 ± 0.36 and 4.00 ± 0.35, respectively, and in the healthy controls were 3.93 ± 0.80 and 3.94 ± 0.75, respectively; p = 0.33 for the right putamen and 0.49 for the left putamen.

The mean volume ± SD of both right and left thalami in the patients were 5.34 ± 0.62 and 5.46 ± 0.69, respectively, and in the healthy controls were 3.61 ± 0.42 and 3.54 ± 0.44, respectively; p = 0.51 for the right thalamus and 0.66 for the left thalamus.

The mean volume ± SD of the right and left hippocampus in the patients were 3.62 ± 0.36 and 3.72 ± 0.41, respectively, and in the healthy controls were 5.26 ± 0.70 and 5.42 ± 0.71, respectively; p = 0.85 for the right hippocampus and 0.21 for the left hippocampus.

The mean volume ± SD of both right and left amygdala in the patients were 0.57 ± 0.16 and 0.58 ± 0.15, respectively, and in the healthy controls were 0.55 ± 0.17 and 0.54 ± 0.18, respectively; p = 0.31 for the right hippocampus and 0.39 for the left hippocampus.

There was no statistically significant difference in the mean volume ± SD of the right and left cerebral hemispheres between patients (530.52 ± 60.92 and 533.88 ± 62.54, respectively) and healthy controls (524.75 ± 75.87 and 525.42 ± 74.19, respectively); p = 0.73 for the right cerebral hemisphere and p = 0.58 for the left cerebral hemisphere.

Also, no significant difference was detected in the mean volume ± SD of the cortical gray matter of both right and left cerebral hemispheres between patients (290.79 ± 56.30 and 293.73 ± 58.00, respectively) and healthy controls (287.77 ± 43.98 and 291.66 ± 44.90, respectively); p = 0.44 for the right cerebral hemisphere and p = 0.45 for the left cerebral hemisphere.

There was no significant difference in the mean volume ± SD of the white matter of both right and left cerebral hemispheres between patients (239.72 ± 68.62 and 240.15 ± 74.27, respectively) and healthy controls (236.98 ± 41.55 and 233.76 ± 39.99, respectively); p = 0.60 for the right cerebral hemisphere and p = 0.82 for the left cerebral hemisphere.

The mean volume ± SD of the right lateral ventricle was significantly larger than in the healthy controls (8.76 ± 4.88 and 5.97 ± 4.63, respectively; p = 0.03; Fig. 4); however, although the mean volume ± SD of the left lateral ventricle was relatively larger in the patients’ group compared to healthy controls, the difference was not statistically significant (7.43 ± 3.34 and 6.44 ± 4.29, respectively; p = 0.18).

Fig. 4
figure 4

Volumetric differences in the lateral ventricles between patients and controls

Full size image

Regarding the cerebellum, there was no significant difference in the mean volume ± SD of the right and left cerebellar hemispheres between patients (65.25 ± 13.25 and 67.10 ± 12.17, respectively) and healthy controls (65.25 ± 8.81 and 65.63 ± 8.66, respectively); p = 0.76 for the right cerebellar hemisphere and p = 0.54 for the left cerebellar hemisphere.

Examples of the electronically generated reports and colored graphical maps of the volumetric data for a patient and a control subject are demonstrated in Figs. 1b and c and 2b and c, respectively.

The correlation between volumetric data and some clinical criteria in BD patients

The Spearman’s analysis of the correlation between the mean volumes of different anatomical structures in the brain yielded the following results (Table 4):

Table 4 Correlation between clinical data and volumetric data of different brain structures
Full size table

There was a significant negative correlation between the mean volume of the right caudate nucleus and the patient’s age (r = −0.42, p = 0.04), as well as the age of onset of BD (r = −0.44, p = 0.03).

There was a highly significant negative correlation between the mean volume of the right putamen and the patient’s age (r = −0.62, p = 0.001) and significant negative correlation between the mean volume of the right putamen and the age of onset of BD (r = −0.46, p = 0.02). Similar findings were found in the left putamen where there is highly significant negative correlation between the mean volume of the left putamen and the patient’s age (r = −0.60, p = 0.002) and significant negative correlation between the mean volume of the left putamen and the age of onset of BD (r = −0.43, p = 0.02).

Regarding the globus pallidus, there was a highly significant negative correlation between the mean volume of the right globus pallidus and the patient’s age (r = −0.55, p = 0.005). The mean volume of the left globus pallidus shows a significant negative correlation with the patient’s age (r = −0.43, p = 0.03), Also, it shows significant negative correlation with the age of onset of BD (r = −0.46, p = 0.02).

There is a significant negative correlation between the mean volume of the right thalamus and YMRS of the patient (r = −0.40, p = 0.04). The mean volume of the left thalamus shows highly significant negative correlation with the YMRS of the patient (r = −0.58, p = 0.003).

The mean volume of the left hippocampus shows significant positive correlation with the YMRS of the patient (r = 0.46, p = 0.02).

The mean volume of the right amygdala shows significant negative correlation with the number of manic attacks (r = −0.47, p = 0.02).

The mean total volume of each cerebral hemisphere shows significant negative correlation with the patient’s age as well as the YMRS of the patient. For the right cerebral hemisphere, the values were r = −0.43, p = 0.03 and r = −0.47, p = 0.02, respectively; and for the left cerebral hemisphere, the values were r = −0.45, p = 0.02 and r = −0.48, p = 0.01, respectively.

The mean volume of the gray matter of each cerebral hemisphere shows highly significant negative correlation with the patient’s age. For the right cerebral hemisphere, the values were r = −0.63, p = 0.001 and for the left cerebral hemisphere, the values were r = 0.62, p = 0.001. Also, the volume of the gray matter of each cerebral hemisphere shows significant negative correlation with the number of attacks as well as the duration of illness. The values for the right cerebral hemisphere were r = −0.49, p = 0.01 and r = −0.47, p = 0.02, respectively, and for the left cerebral hemisphere, were r = −0.46, p = 0.02 and r = −0.49, p = 0.01, respectively.

The mean volume of the white matter of each cerebral hemisphere shows highly significant negative correlation with the patient’s YMRS. For the right cerebral hemisphere, the values were r = −0.68, p = 0.001 and for the left cerebral hemisphere, the values were r = 0.65, p = 0.001. Also, the volume of the white matter of the left cerebral hemisphere shows significant negative correlation with the duration of the last attack of mania (r = −0.42, p = 0.04).

The mean volume of the right lateral ventricle shows significant negative correlation with the duration of illness (r = −0.48, p = 0.01).

Discussion

Bipolar disorder is a psychiatric illness characterized by severe mood-state transitions from euthymic to depressed and manic states. The underlying pathophysiology of the disease represents a major research target for recent neuroimaging studies. There is a growing body of evidence that BD arises from abnormalities within brain systems that modulate emotional behavior implicating the frontal and limbic regions of the brain, which are central to emotional processing [10, 11].

Although the specific control of emotional function in humans is not completely understood, two networks appear to modulate emotional behavior. Both these networks map to specific striatal, pallidal, and thalamic brain areas to form feedback loops that process information and modulate the amygdala and other limbic areas. These two networks are likely central in the functional neuroanatomy of bipolar disorder [12,13,14,15,16]. Moreover, there is accumulated evidence suggesting that cortical-subcortical imbalance could represent a cause of bipolar disorder [17].

The prefrontal cortical areas that modulate human emotional function seem to do so within relatively independent prefrontal-striatal-pallidal-thalamic iterative circuits, the so-called cortico-striato-pallido-thalamo-cortical (CSPTC) circuits, which project back to the frontal lobe [18]. The globus pallidus is thought to be positioned at the heart of this CSPTC circuitry [19, 20].

Structural and functional neuroimaging studies on BD assume that abnormal prefrontal-subcortical-limbic connectivity cause cortical-limbic dysregulation reducing the inhibition of the ventral-limbic network and enhancing emotional responses [20,21,22]. Subcortical gray matter regions have been implicated in mood disorders, including BD [23]. Volumetric analysis of structural imaging data of the brain has been a mainstay investigation in psychiatry recently [24].

This study investigated structural neuroanatomical changes in patients with BD in comparison to healthy controls using VBM. VBM is continually being adapted as a research tool as it has advantages over conventional brain volume measurement in that it can detect small changes as it examines across the whole brain [25].

The primary focus was on the volumetric differences of different brain structures. Seven subcortical structures were explored: the caudate, putamen, globus pallidus, hippocampus, thalamus, and amygdala. The differences in the volumes of both cerebral and cerebellar hemispheres, the gray and white matter of both cerebral hemispheres, as well as the volume of the lateral ventricles were also assessed.

The findings of our study show that patients with BD had significantly larger mean volume of the right globus pallidus (p = 0.05) compared to healthy controls. BD patients also had significantly larger right lateral ventricles (p = 0.05) than healthy controls. None of the other evaluated subcortical structures showed significant volumetric differences between BD patients and control subjects. Also, no significant difference was detected between patients and controls concerning the mean total volume of both cerebral hemispheres, the mean gray matter and white matter volumes of both cerebral hemispheres, or the mean volume of both cerebellar hemispheres.

The findings of this study converge with multiple previous meta-analytical studies [26,27,28] which have shown enlarged lateral ventricles and globus pallidus in BD patients compared to healthy controls. The findings of this study also agree with previous research that described larger mean volume of the globus pallidus in BD patients compared to healthy controls [29, 30].

Concerning the volume of the lateral ventricle, the findings of this study are also concordant with those of a mega-analysis performed by Hallahan et al. [31] who reported enlargement of the right lateral ventricle volume in BD patients compared to healthy controls. Lateral ventricular enlargement was detected both in adolescent and adult BD patients, especially those with repeated manic episodes. This could suggest an underlying developmental mechanism, or might reflect the presence of a degenerative mechanism or progressive periventricular volume losses resulting from mood episodes [32, 33].

In BD patients, relative right hemisphere dysfunction has been postulated supported by reports of relatively greater impairment in visuospatial functioning and lateralization abnormalities [34]. The definite role of the globus pallidus in emotion and cognition is not clearly understood yet. Hence, potential explanations for the volume differences observed in the current study and similar research are not apparent. However, the association of lesions in the globus pallidus with depression and apathy, as well as the diminished reward effects of drugs of abuse, suggest roles in reward and motivation [35].

One of the major points of uncertainty in BD has centered on potential volumetric changes in the hippocampus. In this study, no significant volumetric difference was detected in the hippocampus between BD patients and healthy controls. Two large single-site studies reported a significant hippocampal reduction [36, 37], while other multisite studies reported no significant differences [26,27,28, 31].

This study did not find any significant change in the mean volume of the thalamus between BD patients and healthy controls. This is concordant with the results of multiple meta-analyses [26,27,28, 31], whose findings describe absence of any significant change in the size of the thalamus between BD patients and healthy controls. On the other hand, some studies like Rimol et al. [36] described reduction in the thalamus volume in BD patients.

This study did not find any significant volumetric change in the amygdale between BD patients and the healthy controls. This is consistent with previous literature-based meta-analysis which could not detect a case-control volumetric difference in the amygdala [37]. However, individual studies in adults with BD had reported either increase or no change in the amygdala volume [38, 39]. In contrast, reduced amygdale volume has been reported in adolescents with BD [40] and has been attributed to developmental abnormalities in this region in adolescence.

In this study, there was no significant difference in the volume of the cortical gray matter of both cerebral hemispheres in BD patients compared to healthy controls. These findings are consistent with those of Yüksel et al. [41] and Rimol et al. [42] who described absence of any volumetric changes in the cortex of both cerebral hemispheres.

In this study, there was no significant change in the volume of the cerebral white matter of both cerebral hemispheres. Our findings are concordant with multiple meta-analysis studies [26,27,28] which show that the total cerebral white matter volume in BD is preserved without a change.

It should be considered when interpreting the results of neuroantomical studies that abnormalities of structure do not necessarily mean abnormalities of function. In the absence of clinico-anatomical correlations, it is not possible to directly demonstrate that the structure is involved in the expression of BD. Actually, it cannot be determined whether the volumetric abnormalities preceded or, alternatively, developed during the course of disease [43].

To the best of the knowledge of the authors of this study, this work is the first to try to detect correlations between structural volumetric changes and different clinical aspects of bipolar I mania, including the age of onset of the disease, the duration of illness, the number of manic attacks, and the YMRS of the patient. No similar studies were found despite extensive research in available published scientific database.

The present study has some potential limitations. One of the limitations is that the global cortical gray matter of both cerebral hemispheres was studied without specifically examining the cortical regions implicated in emotion and cognitive processing and emotion regulation, including the orbitofrontal, anterior cingulate, and anterior temporal cortices. Another limitation is the relatively small sample size of patients and healthy controls, which might affect the sensitivity and specificity of the results; a larger independent sample is needed to confirm the reproducibility of findings.

Conclusions

VBM could address specific structural findings in the brain in patients with mood disorder bipolar I mania that may contribute to the pathophysiology of the disease. The findings of this study also show that there is significant correlation of these structural changes with different clinical aspects of the disease. These findings encourage similar studies in different neuropsychological disorders.