Habenula volume change in Parkinson ’ s disease: A 7T MRI study

Objective: Parkinson ’ s disease (PD) is a progressive neurodegenerative disease characterized by motor and early non-motor symptoms. The habenula is implicated in the pathophysiology of depression. This study investigates habenular volume in PD patients without clinical depression to show the changes in PD unrelated to depression. Methods: The study used high-resolution 7 Tesla MRI data from the TRACK-PD study involving 104 PD patients and 44 healthy controls (HCs). The habenula was manually segmented, and volumes were measured, considering demographic data and depression scores via the Beck Depression Inventory (BDI). Results: No significant correlation was found between habenular volume and BDI scores in PD patients or HCs. However, the PD group exhibited a significantly larger mean and right habenular volume than HCs. Although PD patients showed higher BDI scores, indicating more subthreshold depression, these did not correlate with the habenular volume. Conclusion: The results suggest that while the habenula may be involved in the symptoms of PD, its role in depression within this cohort is unclear. The changes might be related to the role of the habenula in motor symptoms. This study provides a new perspective on the role of the habenula in PD, but future research could lead to a greater understanding of the neuroanatomical features of the habenula in PD.


Introduction
Parkinson's disease (PD) is a progressive neurodegenerative disease frequently associated with motor and non-motor symptoms such as sleep disorders, hyposmia, pain, constipation, and mood disorders (Cummings and Masterman, 1999;Kano et al., 2011;Antonelli et al., 2010;Dissanayaka et al., 2010).Depression is a prevalent and incapacitating non-motor symptom of PD, with reported prevalence rates varying widely across studies (2.7 %-90 %) and can be observed even in the premotor phase of the disease (Reijnders et al., 2008).The estimated 50 % prevalence of depression among PD patients reflects a significant proportion of the PD population potentially affected by this symptom.
The habenula has been identified as one of the critical brain regions in the pathophysiology of mood disorders such as depression.Mood disorders often involve dysregulation of the habenula-related reward pathway (Li et al., 2011;Proulx et al., 2014), leading to increased habenular activity and metabolism, which correlates with depression symptoms (Morris et al., 1999;Caldecott-Hazard et al., 1988).Notably, ablation of the habenula has been found to reduce depression-like behavior in rodent depression models (Caldecott-Hazard et al., 1988;Yang et al., 2008).Volumetric changes of the habenula were also reported in human studies of depression.Several studies showed that patients who have major depressive disorder (MDD) and treatment-resistant depression (TRD) had smaller habenula volumes than controls, while the volume was greater in the first episode of depression than that of healthy controls (HCs) and with chronic MDD (Savitz et al., 2011;Johnston et al., 2015;Cho et al., 2021;Carceller-Sindreu et al., 2015).Deep brain stimulation (DBS) of the habenula has been demonstrated to alleviate depressive symptoms in a small number of individuals with TRD (Sartorius et al., 2010).
Habenula is also known to be directly or indirectly related to some brain regions, such as the subthalamic nucleus and globus pallidus internus, which are essential in PD pathology (Baker et al., 2016;Benarroch, 2015).However, there is minimal data available on habenular function in PD patients with or without depression.One study found alterations in habenular functional connectivity among PD patients with punding (Markovic et al., 2017).It is unknown whether habenular changes are present among PD patients without clinically relevant depression and potentially predispose PD patients to develop depressive symptoms.
In this study, we investigated habenular volumes in patients with a recent PD diagnosis and without clinical depression using 7 Tesla MRI data from the TRACK-PD study (Wolters et al., 2020).We hypothesized that habenula volume would be different between PD patients and HCs since the connections of habenula with PD-related areas.
The TRACK-PD study was conducted according to the principles of the Declaration of Helsinki and in accordance with the Medical Research Involving Human Subjects Act.It was approved by the Institutional Review Board (IRB) of the Maastricht University Medical Centre (METC AZM/UM 18-027), and written informed consent was obtained from all participants before inclusion.
Participants were included in the TRACK-PD study if they met the following criteria: 1) For the PD group, has been diagnosed with PD according to diagnostic criteria (Postuma et al., 2015) by a neurologist within the last three years before inclusion; 2) Absence of dementia, and a score of ≥24 on the Montreal Cognitive Assessment (MoCA) at baseline; 3) Able to read and understand Dutch; 4) 18 years of age or older; 5) Accepting to be a participant in this study by written informed consent.Participants who had any contraindications for a 7 T MRI scan were excluded.Participants with a previous or active diagnosis of clinical depression at the time of TRACK-PD study enrollment were also explicitly excluded from this study.

MRI Acquisition and Volumetric Analysis
Participants were scanned using a 7 T MRI scanner (Magnetom, Siemens, Erlangen, Germany) with a Nova Medical 32-channel head coil.Dielectric pads were applied to improve the signal in the temporal brain regions (Teeuwisse et al., 2012).Habenula volumes were determined on 7 T MRI MP2RAGE (Magnetization Prepared 2 Rapid Acquisition Gradient Echoes) images with Field of View (FoV) 208 ×208 mm, 0.65×0.65×0.65 mm 3 resolution, Repetition Time (TR) 5000 ms, Echo Time (TE) 2.51 ms, flip angle (FA) 5 and 3 • , and 240 sagittal slices (Wolters et al., 2020).The habenula was manually segmented using ITK-SNAP version 3.8.0(Yushkevich et al., 2006).Manual segmentation was selected since there is no standardized and common automated segmentation method for habenula yet.The habenula was delineated on coronal slices where this structure was seen bulging into the third ventricle along the ventromedial aspect of the thalamus or lying ventromedial to the stria medullaris of the thalamus (Mai et al., 2015).The medial boundary was formed by the cerebrospinal fluid of the third ventricle, and the white matter of the posterior commissure formed the ventral boundary.The dorsal and lateral borders were defined by the white matter of the stria medullaris of the thalamus in anterior planes or the mediodorsal thalamic nucleus, limitans nucleus, or pretectal area in posterior planes (Fig. 1).During the training period of the raters, the habenula segmentation was executed across all three anatomical planes.Initially, segmentation was conducted in the coronal planes, which was accepted as a reference.Subsequently, the accuracy of this segmentation was verified in the axial and sagittal planes.Adjustments were made by adding or removing voxels in these planes to ensure anatomical consistency with the coronal plane reference.Thus, the habenular volumes of each hemisphere were determined, and the mean habenular volume per participant was calculated.After the segmentation training period, the left and the right habenula were segmented for 20 participants by two independent raters (BS, SM) blinded for diagnosis.The Dice Similarity Coefficient for two raters was 0.92 for the left habenula and 0.91 for the right habenula for these 20 subjects.Then, the remaining segmentations were performed by only one rater (BS), blinded for diagnosis.The total brain volume of each participant was calculated using the standard FreeSurfer software v7.1 (http://surfer.nmr.mgh.harvard.edu/).The processing steps included non-uniform signal correction, signal and spatial normalizations, skull stripping, and brain tissue segmentation (Fischl et al., 2002).

Statistical analysis
Independent samples t-test was used to define the differences in habenula volume between the groups.MANCOVA design was applied to compare the total volumes.The total brain volume (gray matter and white matter combined) of each participant was calculated using the standard FreeSurfer processing pipeline to correct for differences in head size by including it also as a covariate in the MANCOVA designs.BDI total scores were also accepted as a covariate to find the BDIunaffected volume differences between PD and HC groups.Normality was confirmed using the Kolmogorov-Smirnov method.Independent samples t-tests or chi-square tests were used to compare the groups on clinical and demographic variables.Pearson or Spearman correlation tests and regression models were used to show if there was an association between habenula volumes and clinical and demographical features, as well as BDI scores.
Before each analysis, outliers were defined as data points that lay more than 3 standard deviations away from the mean of their respective group.This criterion is based on a widely accepted statistical practice aimed at minimizing the potential impact of extreme values on the overall analysis.The identification process was conducted using SPSS 28.0, which enabled a meticulous examination of each subject's data across the variables of interest.This rigorous approach ensured that only data points significantly divergent from the group distribution, and thus likely to skew the results, were considered outliers.Six subjects were identified as outliers based on these criteria and subsequently excluded from the analysis.
A p-value < 0.05 was considered statistically significant.SPSS 28.0 (IBM, SPSS Inc., New York, United States) package was used for the statistical analysis.GraphPad Prism 9 (GraphPad Software, Massachusetts, United States) was used for graphics.

Clinical and demographical results
PD and HC groups were matched by age, sex, handedness, and educational status.Thirty-three PD patients (31.7 %) were at Hoehn-Yahr stage 1, 67 (64.4 %) were at stage 2, and 4 (3.9 %) were at stage 3.The clinical and demographical data of the participants are shown in Table 1.

Clinical correlations
There was no significant difference between the H&Y stages regarding BDI total score (p = 0.320) among PD patients (Fig. 2).A significant positive correlation was detected between LEDD and BDI total score (r = 0.068, p = 0.005), while there was no relationship between disease duration and BDI score (p = 0.138).

Beck Depression Inventory
The mean BDI total score was significantly higher in the PD group compared to HCs (7.2 ± 5.05 vs. 3.5 ± 3.64; p < 0.001) (Table 1).However, the mean BDI total score for both groups remained below 9, indicating the overall absence of relevant depression.In the PD group, there were seven PD patients (6.7 %) with depression, 34 patients (32.7 %) with subD, and 63 patients (60.6 %) had no depression by BDI scores.In the HC group, there were no participants with depression, five participants (11.4 %) with subD, and 39 HCs (88.6 %) had no depression.

Differences of Habenula volumes between hemispheres and groups
The left habenula was significantly larger compared to the right habenula in both groups (36.6 ± 6.79 mm 3 for the left and 31.9 ± 5.75 mm 3 for the right side in the PD group; 34.3 ± 6.05 mm 3 for the left, and 29.1 ± 5.44 mm 3 for the right side in HC group; p < 0.001 for both groups).

Relationship between habenula volumes, BDI scores, and clinical features
There were no differences in habenula volume in both PD patients and HCs when they were divided into three depression groups (nondepression, subD, and significant depression) by BDI score (p = 0.939 for left and p = 0.391 for right for PD (Fig. 5a); p = 0.508 for left and p = 0.594 for right for HCs (Fig. 5b)).After the PD group was divided into two groups by BDI score, a group with ≥ 9 (n = 42) and a group with ≤ 8 BDI total score (n = 62), there were no significant volumetric differences (p = 0.345 for left, and p = 0.444 for right) between these groups.There were also 39 non-depressive and five subD participants in the HC group, and we found no significant volumetric differences between these groups (p = 0.508 for the left and p = 0.594 for the right).In addition, no significant correlations were found between habenula volumes and BDI total scores (r = 0.028 and p = 0.737 for left habenula volume, r = 0.051 and p = 0.544 for right habenula volume).
There were also no volumetric differences between the PD groups by H&Y stage (p = 0.243 for left and p = 0.291 for right habenula), and no correlation between the volumes with age (r = -0.008and p = 0.928 for left habenula volume, r = 0.015 and p = 0.855 for right habenula volume), LEDD (r = 0.098 and p = 0.242 for left habenula volume, r = 0.132 and p = 0.115 for right habenula volume), and disease duration (r = -0.022and p = 0.825 for left habenula volume, r = 0.001 and p = 0.990 for right habenula volume).

Discussion
This is the first neuroimaging study utilizing a 7 T MRI to investigate  the habenula volume in PD within a large cohort.In this study, we found that the mean habenula volume was larger in PD patients than in HCs.In particular the right habenula showed larger volumes comparing PD and HC groups.There was no association between habenula volumes and depression scores.PD patients had considerably higher BDI scores than HCs, and 32.7 % of PD patients had a BDI total score consistent with subD (11.4 % in the HC group).SubD is characterized by clinically depressive symptoms that do not meet the criteria for MDD, and the prevalence is known to be 8 % -28.8 % in PD (Santangelo et al., 2014).However, there is an overlap between mood symptoms and motor symptoms in PD, and diagnosis of subD, a subtle presentation, could be a challenge.Therefore, more PD patients may have undiagnosed subD or depression.We found BDI scores associated with higher LEDD and not H&Y stages.Thus, it is essential to acknowledge depression as a non-motor symptom of PD in all stages of the disease.
The habenula is believed to be involved in the regulation of emotions, particularly negative ones like fear, stress, anxiety, and depression.The lateral subdivision of the habenula, which regulates the activity of serotonin and dopamine neurons, has been linked to the pathophysiology of depression in PD.Within the framework of the monoamine deficiency hypothesis of depression, the lateral habenula couples the dopaminergic and serotonergic systems, and altered basal ganglia activity leads to lateral habenula hyperactivity, which downregulates the serotonergic system, causing depressive symptoms in patients with PD (Sourani et al., 2012).
Neuroimaging research has recently focused on structural alterations of the habenula in mood disorders.Savitz et al. (2011).conducted the   first MRI volume examination on patients with MDD and bipolar disorder (BD).Currently, depressed women with MDD had smaller habenula volumes than female controls.The study concluded that decreased volume might have functional repercussions that raise the risk of having an affective disorder.Several investigations have demonstrated that MDD or TRD patients had smaller habenula volumes than HCs (Johnston et al., 2015;Cho et al., 2021).Carceller-Sindreu et al. (2015).investigated whether habenular volume altered between HCs and patients with varying phases of unipolar MDD.The habenular volume of first-episode depressive women was greater than that of HCs and women with chronic MDD.The authors reported that changes in habenular volume in unipolar MDD could be absent at any stage of the illness, including in patients with severe chronic depression.However, they also claimed that the increased white matter volume observed in women with a first episode indicates that the habenula and its projections play a role in the early phases of recovery and the progression of MDD.Additionally, Liu et al. (2017) discovered increased habenula volume in patients with MDD, which was associated with anhedonic symptoms.Schafer et al. (2018) examined the differences between BD and HCs and found no longitudinal volumetric difference between the two groups.The variable results concerning the relationship between habenular volume and mood disorders may be influenced by methodological factors, such as the relatively small number of participants and the segmentation method.In this study, we only had a minimal number of PD patients and HC with clinical manifest depression.This may be the reason why we did not find a relationship between depression scores and volume.
Only one clinical study reported on habenular imaging in PD with behavioral impairment.Markovic et al. (2017).obtained structural and resting state-fMRI in 22 PD patients with punding, 30 PD patients without impulse control disorders who were matched for disease stage and duration, motor impairment, and cognitive status, and 30 HCs.A seed-based technique was used to examine the resting state-functional connectivity (FC) of the habenula and amygdala, and the volumes and cortical thickness of both structures were determined.They discovered that PD patients with punding had an increased FC of the habenula and amygdala with the bilateral thalamus and striatum and a decreased FC of the bilateral habenula with the left frontal and precentral cortices.PD patients with punding had a lower FC between the right amygdala and the hippocampus than PD patients without punding.The habenula and amygdala did not differ in size between groups.The authors suggested that a disruption in connectivity between critical nodes of the reward circuit (i.e., habenula, amygdala, basal ganglia (BG), and frontal cortex) may contribute to punding in PD.Borgonovo et al. (2017) also published a review of the evidence of early disruption in brain circuitry associated with depression in PD, focusing on the pre-clinical, premotor, and early motor stages of the disease.They asserted that the limbic loop of the BG and lateral habenula play crucial roles in the early development of depression in PD and that alterations in the neural circuitry associated with emotional control may serve as sensitive markers of the ongoing neurodegenerative process.According to the BDI scores, there was no Fig. 5. Differences in habenula volumes (mm3) between PD patients with no depression, subthreshold depression, and significant depression (a), and HCs with no depression and subthreshold depression (b) according to BDI scores.There were no significant differences between the groups.
relationship between habenula volumes and BDI scores in both groups in our cohort.With these results, since there are no clinically diagnosed depression patients in our cohort, it can be thought that these changes in the neural circuits involved in the development of depression and changes in habenula volume have not yet occurred in Parkinson's patients during the subclinical depression period.Additionally, we might have yet to find a correlation because we only conducted volumetric analyses.Pathological changes could be detected earlier in connectivity analyses.Further neuroimaging investigations in longitudinal datasets with depressed and non-depressed patients are required to elucidate the link between depression in PD and the habenula.
The human habenula shows asymmetry in size due to cytoarchitectonic organization, neurochemistry, and functional connectivity (Bianco and Wilson, 2009).Postmortem volumetric analysis (Ahumada- Galleguillos et al., 2017) and volumetric MRI scans revealed that HCs had increased left habenular volume (Savitz et al., 2011).In this study, we report, for the first time, a differential volume between the left and right reining nuclei in patients with PD, a finding that echoes previous observations of asymmetry in non-PD subjects but underscores a novel aspect of neuroanatomical variation in PD.This phenomenon dovetails with the work of Scherfler et al. (2012), who highlighted a marked reduction in dopamine transporter binding in the left posterior putamen compared to the right, suggesting a left hemispheric predominance of nigrostriatal dysfunction in PD.Such lateralized neurodegenerative patterns offer a compelling explanation for the asymmetry we observed, proposing that the uneven dopaminergic dysfunction may contribute to differential volumetric changes in the habenula.Theories also suggest an inherently lower level of dopamine in the left versus right nigrostriatal pathway in right-handed patients or a selective vulnerability of left nigrostriatal projections to the pathological mechanisms that characterize PD (Djaldetti et al., 2006;Melamed and Poewe, 2012).These hypotheses align with our findings and suggest that the structural brain changes observed in PD patients-specifically, the volumetric differences in the habenula-may be a direct consequence of the disease's asymmetric impact on dopaminergic innervation.We also found that the right habenula volume was significantly larger in PD patients than in HCs.In the PD group, we demonstrated significantly higher BDI scores than the HCs in our cohort, and more PD patients had subD.It might be hypothesized that larger habenula volumes accompany the first symptoms of depression as a compensatory mechanism and control of neurotransmitter systems.As compensatory mechanisms fail and MDD or TRD occurs, volumes of the habenula may subsequently decrease.However, in our study, we only had a minimal number of patients with clinical depression, and there was no correlation between the volumes and BDI scores.Additionally, it is known that the right habenula has more connections with the substantia nigra and ventral tegmental area (VTA) than the left (Hetu et al., 2016).Dopamine neurons in the VTA are thought to play essential roles in motivation, reward, various cognitive and emotional processes, and also in negative affect.The nigrostriatal dopaminergic system, in which substantia nigra pars compacta, plays essential roles in sensorimotor functions, including motor execution and habit formation, and also has affective and motivational roles (Ilango et al., 2014).Hence, this increased volume could be related to motor symptoms in people with PD.The habenula might be overactivated to compensate for dopaminergic disturbances that cause motor and mood symptoms in PD.For future research, it would be prudent to study whether there is an association between the severity of motor and other non-motor symptoms in PD patients and habenula volumes.
A limitation of the current study is the lack of clinically depressed participants and PD patients.A second limitation is that the scores of the patients' detailed motor scales are not included.We also have a smaller HC group than PD.Future studies with depressed PD patients and larger HC groups would be meaningful to understand the relationship between PD and habenula.However, the main strength of the study is that this is the first large-scale examination of the habenula volume in PD using a 7 T MRI.Acquiring in vivo habenula imaging data is challenging because of its central location, small size, and close proximity to the third ventricle.The latter causes partial volume effects, especially in clinical MRI scanners.The availability of high-resolution MRI is thought to have aided in detailing habenular morphology, connectivity, functional activation, and reduced magnetic susceptibility.

Conclusion
This study is the first to demonstrate larger mean and right habenula volumes in PD patients compared to HCs.These results might be significant for understanding the role of habenula in PD pathophysiology.Using high-resolution 7 T-MRI in large groups of patients and longitudinal studies could lead to a greater understanding of the neuroanatomical features of the habenula in PD.

Fig. 1 .
Fig. 1.Coronal MRI sections showing the habenula and the local anatomical landmarks for manual segmentation.

Fig. 4 .
Fig. 4. Right and left habenula of a PD patient (a, b, c) and a HC (d, e, f).The right habenula volume in the PD patient is larger than the HC (red-framed area), and the left habenula volume is larger than the right in both this PD and HC patient.

Table 1
The clinical and demographical data and Beck Depression Inventory scores of the participants.
Noted are the mean ± standard deviation unless otherwise stated.PD: Parkinson's disease, HC: Healthy control, M: Male, F: Female, R: Right, L: Left, LEDD: Levodopa equivalent daily dose.

Table 2
The habenula volumes of the participants.

group (n ¼ 104) HC group (n ¼ 44) p Left habenula volume (mm 3 )
Noted are the mean ± standard deviation unless otherwise stated.PD: Parkinson's disease, HC: Healthy control.Beck depression inventory scores were used as covariates in the analysis.