Cerebellar volumes in early-onset bipolar disorder: a pilot study of a stereological measurement technique

ABSTRACT OBJECTIVE Recent data from the literature have recognized the importance of cerebellum in bipolar disorder. Brain imaging studies focusing on cerebellar volumetric changes in bipolar disorder demonstrated controversial data. The aim of this study is to investigate whether there is any difference between early-onset bipolar cases and healthy controls regarding cerebellar volumetric measurements. METHODS Patients with bipolar I disorder were compared to healthy controls in terms of total cerebellar volume, volumes of the right and left cerebellar hemispheres, and cerebellar volumetric asymmetry. All the sociodemographic, clinical data, and magnetic resonance image scans were collected retrospectively. Cerebellar volumes were evaluated using the stereological method. Asymmetry indices were calculated subsequently. RESULTS We recruited 10 bipolar I cases and 10 healthy controls. There were no statistically significant differences between the bipolar and the control groups for total cerebellar volumes, volumes of right and left cerebellar hemispheres, and cerebellar asymmetry indices. CONCLUSION Future studies focusing on cerebellar changes in early-onset bipolar disorder should include large case and control series and designed as follow-up studies for being able to determine the chronic effects of the illness on cerebellar volumes.


Introduction
Bipolar disorder is a debilitating psychiatric illness characterized by severe changes in the mood [1,2]. Early-onset bipolar disorder is represented by slow response to treatment, persistent mood fluctuations, elevated risk for suicidal attempts, and severe psychosocial impairment [2]. Although the pathophysiology underlying the mood-state transitions in bipolar disorder remains largely unknown, many studies of the disorder using imaging techniques have implicated various brain structures such as the amygdala, anterior paralimbic cortices which are central to emotional processing [1,3].
Accumulating evidence have recognized the importance of cerebellum in emotional regulation and cognitive activities involving executive functioning, episodic memory, and sensorimotor processing in addition to its well-known motor control functions [3][4][5]. This feature provides a strong neuroanatomical basis for the involvement of the cerebellum in emotional regulation.
Functional magnetic resonance imaging studies found resting-state abnormalities of cerebellar functional connectivity in patients with bipolar disorder [6,7]. Cerebrocerebellar hypoconnectivity in specific brain networks may associate with abnormal timing, coordination, integration and signal processing observed as the cognitive, emotional, and behavioral symptoms of bipolar disorder [6].
Structural brain imaging studies focusing on volumetric changes of cerebellum did not find significantly smaller cerebellar hemispheres compared with healthy subjects in adult bipolar patients [8][9][10][11]. However, there is very limited data on the early-onset bipolar disorder. Adler et al. reported larger bilateral cerebellar grey matter volume in a mixed sample of adolescents and adults (age range 13-41 y) with bipolar I disorder in first-episode compared to healthy subjects [12]. Controversially, in a longitudinal study, patients with bipolar disorder showed a decrease in cerebellar grey matter density over 4 years, and this decline was more rapid than that observed in healthy control subjects [13].
James et al. [14] found loss of grey matter density in the cerebellum of adolescents with bipolar disorder. Studies investigating the importance of cerebellum in neuropsychiatric disorders have focused not only on cerebellar volumetric measurements but also on cerebellar asymmetry [15,16]. It has been indicated that there is a pattern of cerebellar asymmetry in all healthy humans. And the asymmetry index is positive as the right hemisphere volume is larger than the left [15,16]. There are some studies suggesting that cerebellar asymmetry is reversed in schizophrenia compared to healthy controls [15].
To our knowledge, there is no study exploring cerebellar volumetric asymmetry in early-onset bipolar disorder.
The aim of this study is to determine whether there are any differences between early-onset bipolar cases and healthy controls in terms of total cerebellar volume, volumes of the right and left cerebellar hemispheres, and cerebellar volumetric asymmetry.

Features of the inpatient unit
The Child and Adolescent Pscyhiatry inpatient unit of Dokuz Eylül University has been founded in 2005 with a capacity of fifteen patients. The inpatient unit is a member of the Quality Network for Inpatient CAMHS (QNIC). QNIC is an initiative of the Royal College of Psychiatrists in the United Kingdom.

Sample and procedure
The study is retrospective in design. All of the children and adolescents who have been diagnosed as earlyonset bipolar I disorder based on DSM-IV-TR [17] and received treatment between the years of 2005 and 2018 in our inpatient unit with magnetic resonance image (MRI) scans obtained during hospitalization composed the cases of our study. From the total of fourteen adolescents three male and one left-handed female case were excluded from the study since handedness and gender have been inquired for cerebellar volumes, asymmetry, connectivity, and activation [18,19].
The healthy control group composed of 10 age-and gender-matched right-handed children and adolescents who had admitted to the child neurology department of the same hospital for headaches. Those who had a normal neurological examination and normal MRI scan results, did not have a history of any neurological disorder, head trauma, any chronic disorder including any psychiatric disorder, and had no history of admission to psychiatric services for any reason are recruited to the study.
Sociodemographic, clinical, and MRI image acquisition data were collected retrospectively from the hospital records of both cases and controls.
This study was approved by the Medical Sciences Research Ethics Committee of Dokuz Eylul University with decision number 2017/25-18.

MRI acquisition and stereological analysis
MRI images were obtained via an MRI machine (1.5 Tesla Gyroscan Achieva, release 8:1; Philips Medical Systems, Best, The Netherlands) from bipolar cases and healthy controls. Volumetric measurements of cerebellum were performed by using a stereological method described by Cavalieri (Cavalieri principle combined with point counting) [20,21]. This stereological method is known as efficient, simple, unbiased, and reliable for volumetric evaluations of anatomical structures [21].
T1-weighted sagittal sections (3 mm thick) were obtained with a 1.5-T magnetic resonance (MR) machine. The stereological measurements of the total and hemi cerebellar volumes were performed by two blinded observers. An optimal plane was obtained using the smallest diameter of each anisotropic structure that can be measured in two-dimensional analyses on sagittal MRI sections. This value was multiplied by the cross-sectional thickness and the volume parameter. A uniform point grid with a point-associated area of 0.625 cm 2 was randomly superimposed on each MR image using the grid. Points that hit the cerebellum were manually counted for volume estimation. The cerebellar volumes were calculated according to formula as described below: In this formula, "t" defines the section thickness, "SU" the scale unit (the real length of the scale marked on the MRIs), "d" the distance between two points in the point grid, "SL" the scale length (the actual measure of the scale on MRI), and "P" the number of points counted.
Also, coefficient of error (CE) values, which provide information about the accuracy of volumetric sampling, were calculated [21].
To evaluate the interhemispheric asymmetry between the hemispheres, middle sections were identified by clear visualization of the cerebral aqueduct. The number of points of the middle section was divided by two and the results were added to the total point counts for each hemisphere separately. Thereby, the volumes of each cerebellar hemisphere were calculated. The stereological point counting technique representing the point grid superimposed on cerebellum is described in Figure 1.

Statistical analysis
Shapiro-Wilk test was used to assess the normality of distributions of the variables, and Levene's test was used to assess the homogeneity of variances in different groups. Independent sample t-test was used for comparison of the independent groups and paired sample t-test was used for dependent groups. The results of tests were expressed as the number of observations (n), the mean ± standard deviation (mean ± SD). Data analyses were performed with SPSS 24.0 software. A P-value less than .05 was considered as statistically significant.

Results
There were 10 bipolar I cases and 10 healthy controls. All the cases and the controls were right-handed girls. The mean (±SD) age was 184.40 (±21.73) months (15.30 ± 1.88 years) for bipolar cases and 183.30 (±22.54) months (15.30 ± 1.88 years) for healthy controls. Comparison of the ages of both groups did not display any significant differences (P = .978). Additional clinical data such as medication used and the results of clinical scales applied during treatment of the cases are provided in Table 1.

Comparison of bipolar cases and controls for cerebellar volumes and cerebellar volumetric asymmetry indices
There were no statistically significant differences between the bipolar and the control group for total cerebellar volumes, volumes of right and left cerebellar hemispheres, and cerebellar asymmetry indices. Results are presented in Table 2.
In the bipolar group, mean CE values for right and left hemispheres were 0.029 ± 0.010 and 0.025 ± 0.008, respectively. In the control group, mean CE values for right and left hemispheres were 0.032 ± 0.009 and 0.051 ± 0.008, respectively. All calculated CE values were in an acceptable range reflecting the reliability of the stereological measurements [23,24].

Discussion
We did not find any statistically significant differences in any of cerebellar volumetric measures including asymmetry indices between bipolar patients and healthy controls.
In the literature, there are controversial data about cerebellar volumetric changes in patients with bipolar disorder. In some studies, conducted with adult patients with bipolar disorder, it was shown that cerebellar volumes did not differ according to healthy controls [8][9][10][11]. Mills et al. [10] found midline vermal changes in bipolar adult patients with a history of multiple mood episodes. Monkul et al. did not find any statistically significant differences between bipolar patients and healthy controls in any of the cerebellum and vermis regions measured [25]. However, they reported non-significant tendency for smaller vermis V2 (the superior posterior lobe) area in patients. They also found a significant inverse correlation between the number of prior affective episodes and V2 area in male bipolar patients [25]. James et al. [14] reported the loss of grey matter density in the cerebellum of paediatric bipolar disorders.
In their study, Brambilla et al. [9] compared bipolar patients who had a first-degree relative with a history of mood disorders with the bipolar patients who had not a first-degree relative with mood disorders in terms of cerebellar volumes. As an interesting result, they found that patients with a family history of    mood disorders had smaller cerebellar hemispheres and vermis volumes and larger left lateral ventricle volumes [9]. Although data from the literature showed cerebellar volumetric changes in the patients with bipolar disorder compared to healthy controls, our study did not display any difference between early-onset bipolar patients with an average age of 15 and matched controls. An explanation for this discrepancy might be related to the age of our cases. A search of the literature reveals data regarding the cerebellar volumes in bipolar patients come from adult patients with multiple mood episodes, therefore with a chronic course. Whereas our adolescent patients might not have been reflecting the chronic effects of multiple episodes of the disorder. Therefore, follow-up of our patients into adulthood might yield to different results.
In addition to cerebellar volumetric measurements, we also investigated the presence of cerebellar asymmetry in our study. Although there are some studies investigating cerebellar asymmetry in adult and early-onset schizophrenia, to our knowledge, there is no study exploring cerebellar volumetric asymmetry in both adult and early-onset bipolar disorder [15,16,22]. In their study, Szesko et al. [15] submitted that although patients and healthy comparison subjects did not differ in any of cerebellar volumes, cerebellar asymmetry is reversed in patients with schizophrenia. On the other hand, a study conducted by Özbek et al. [22] did not reveal any cerebellar asymmetry in earlyonset schizophrenia cases in comparison to controls. Similar to Özbek et al.'s finding in early-onset schizophrenia patients, we could not find a significant difference in cerebellar asymmetry indices between earlyonset bipolar cases and healthy controls. Similar to the other cerebellar volumetric changes, age of onset and level of chronicity might also account for this finding. Additionally, this novel issue needs to be investigated in early-onset bipolar disorder with a larger sample size with various ages of onset.

Strengths and limitations
The retrospective design of this study provides some advantages such as enabling the collection of data for a rare disorder from the hospital records and being cost-effective. But there are some limitations as well. An important limitation of this study is the low number of cases. Besides, none of our patients were medication free. Therefore, we cannot eliminate the effects of psychotropic medications. As another limitation, we used 3 mm slices to measure cerebellar volumes. Although 3 mm slice thickness is suggested to be sufficient for analysing brain MRIs, thinner slices might have provided different results [26]. In spite of several limitations, the findings of our study may have value in guiding future studies evaluating the role of cerebellum in bipolar disorder.

Disclosure statement
No potential conflict of interest was reported by the authors.