Reproducibility of myelin content-based human habenula segmentation at 3 Tesla

Segmentation and Volume Estimation of the Habenula Using Deep Learning in Patients With Depression

Background

The habenula is involved in the pathophysiology of depression. However, its small structure limits the accuracy of segmentation methods, and the findings regarding its volume have been inconsistent. This study aimed to create a highly accurate habenula segmentation model using deep learning, test its generalizability to clinical magnetic resonance imaging, and examine differences between healthy participants and patients with depression.

Methods

This multicenter study included 382 participants (patients with depression: N = 234, women 47.0%; healthy participants: N = 148, women 37.8%). A 3-dimensional residual U-Net was used to create a habenula segmentation model on 3T magnetic resonance images. The reproducibility and generalizability of the predictive model were tested on various validation cohorts. Thereafter, differences between the habenula volume of healthy participants and that of patients with depression were examined.

Results

A Dice coefficient of 86.6% was achieved in the derivation cohort. The test-retest dataset showed a mean absolute percentage error of 6.66, indicating sufficiently high reproducibility. A Dice coefficient of >80% was achieved for datasets with different imaging conditions, such as magnetic field strengths, spatial resolutions, and imaging sequences, by adjusting the threshold. A significant negative correlation with age was observed in the general population, and this correlation was more pronounced in patients with depression (p < 10-7, r = -0.59). Habenula volume decreased with depression severity in women even when the effects of age and scanner were excluded (p = .019, η2 = 0.099).

Conclusions

Habenula volume could be a pathophysiologically relevant factor and diagnostic and therapeutic marker for depression, particularly in women.

Habenula Volume and Functional Connectivity Changes Following Laparoscopic Sleeve Gastrectomy for Obesity Treatment

BACKGROUND

Neuroimaging studies have revealed alterations in habenular (Hb) structure and functional connectivity (FC) in psychiatric conditions. The Hb plays a particularly critical role in regulating negative emotions, which trigger excessive food intake and obesity. However, obesity and weight loss intervention (i.e., laparoscopic sleeve gastrectomy [LSG])-associated changes in Hb structure and FC have not been studied.

METHODS

We used voxel-based morphometry analysis to measure changes in gray matter volume (GMV) in the Hb in 56 patients with obesity at pre-LSG and 12 months post-LSG and in 78 normal-weight (NW) control participants. Then, we conducted Hb seed-based resting-state FC (RSFC) to examine obesity-related and LSG-induced alterations in RSFC. Finally, we used mediation analysis to characterize the interrelationships among Hb GMV, RSFC, and behaviors.

RESULTS

Compared with NW participants, Hb GMV was smaller in patients at pre-LSG and increased at 12 months post-LSG to levels equivalent to that of NW; in addition, increases in Hb GMV were correlated with reduced body mass index (BMI). Compared with NW participants, pre-LSG patients showed greater RSFCs of the Hb-insula, Hb-precentral gyrus, and Hb-rolandic operculum and weaker RSFCs of the Hb-thalamus, Hb-hypothalamus, and Hb-caudate; LSG normalized these RSFCs. Decreased RSFC of the Hb-insula was correlated with reduced BMI, Yale Food Addiction Scale rating, and emotional eating; reduced hunger levels were correlated with increased RSFCs of the Hb-thalamus and Hb-hypothalamus; and reduced BMI and Yale Food Addiction Scale ratings were correlated with increased RSFCs of the Hb-thalamus and Hb-hypothalamus, respectively. The bidirectional relationships between Hb GMV and RSFC of the Hb-insula contributed to reduced BMI.

CONCLUSIONS

These findings indicate that LSG increased Hb GMV and that its related improvement in RSFC of the Hb-insula may mediate long-term benefits of LSG for eating behaviors and weight loss.

Visualizing the Habenula Using 3T High-Resolution MP2RAGE and QSM: A Preliminary Study

BACKGROUND AND PURPOSE

The habenula is a key node in the regulation of emotion-related behavior. Accurate visualization of the habenula and its reliable quantitative analysis is vital for the assessment of psychiatric disorders. To obtain high-contrast habenula images and allow them to be compatible with clinical applications, this preliminary study compared 3T MP2RAGE and quantitative susceptibility mapping with MPRAGE by evaluating the habenula segmentation performance.

MATERIALS AND METHODS

Ten healthy volunteers were scanned twice with 3T MPRAGE and MP2RAGE and once with quantitative susceptibility mapping. Image quality and visibility of habenula anatomic features were analyzed by 3 radiologists using a 5-point scale. Contrast assessments of the habenula and thalamus were also performed. The reproducibility of the habenula volume from MPRAGE and MP2RAGE was evaluated by manual segmentation and the Multiple Automatically Generated Template brain segmentation algorithm (MAGeTbrain). T1 values and susceptibility were measured in the whole habenula and habenula geometric subregion using MP2RAGE T1-mapping and quantitative susceptibility mapping.

RESULTS

The 3T MP2RAGE and quantitative susceptibility mapping demonstrated clear boundaries and anatomic features of the habenula compared with MPRAGE, with a higher SNR and contrast-to-noise ratio (all P < .05). Additionally, 3T MP2RAGE provided reliable habenula manual and MAGeTbrain segmentation volume estimates with greater reproducibility. T1-mapping derived from MP2RAGE was highly reliable, and susceptibility contrast was highly nonuniform within the habenula.

CONCLUSIONS

We identified an optimized sequence combination (3T MP2RAGE combined with quantitative susceptibility mapping) that may be useful for enhancing habenula visualization and yielding more reliable quantitative data.

Recent Advances in Habenula Imaging Technology: A Comprehensive Review

The habenula (Hb) is involved in many natural human behaviors, and the relevance of its alterations in size and neural activity to several psychiatric disorders and addictive behaviors has been presumed and investigated in recent years using magnetic resonance imaging (MRI). Although the Hb is small, an increasing number of studies have overcome the difficulties in MRI. Conventional structural-based imaging also has great defects in observing the Hb contrast with adjacent structures. In addition, more and more attention should be paid to the Hb's functional, structural, and quantitative imaging studies. Several advanced MRI methods have recently been employed in clinical studies to explore the Hb and its involvement in psychiatric diseases. This review summarizes the anatomy and function of the human Hb; moreover, it focuses on exploring the human Hb with noninvasive MRI approaches, highlighting strategies to overcome the poor contrast with adjacent structures and the need for multiparametric MRI to develop imaging markers for diagnosis and treatment follow-up. LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY STAGE: 2.

Abnormal functional connectivity in the habenula is associated with subjective hyperarousal state in chronic insomnia disorder

Background

The hyperarousal process model plays a central role in the physiology of chronic insomnia disorder (CID). Recent evidence has demonstrated that the habenula is involved in the arousal and sleep-wake cycle. However, whether the intrinsic habenular functional network contributes to the underlying mechanism of CID and its relationship to the arousal state in CID remains unclear.

Methods

This single-centered study included 34 patients with subjective CID and 22 matched good sleep control (GSC), and underwent a series of neuropsychological tests and resting-state functional magnetic resonance imaging scans. The habenular functional network was assessed using seed-based functional connectivity (FC) analysis. The subjective arousal state was evaluated with the hyperarousal scale (HAS). Alterations in the habenular FC network and their clinical significance in patients with CID were explored.

Results

Compared with the GSC group, the CID group showed decreased habenular FC in the left caudate nucleus and right inferior parietal lobule and increased FC in the right habenula, bilateral calcarine cortex, and posterior cingulate cortex. The decreased FC between the left habenula and caudate nucleus was associated with an increased arousal state in the CID group.

Conclusion

The present results provide evidence for a dysfunctional habenular network in patients with CID. These findings extend our understanding of the neuropathological mechanisms underlying the hyperarousal model in chronic insomnia.

High-resolution Structural Magnetic Resonance Imaging and Quantitative Susceptibility Mapping

High-resolution 7-T imaging and quantitative susceptibility mapping produce greater anatomic detail compared with conventional strengths because of improvements in signal/noise ratio and contrast. The exquisite anatomic details of deep structures, including delineation of microscopic architecture using advanced techniques such as quantitative susceptibility mapping, allows improved detection of abnormal findings thought to be imperceptible on clinical strengths. This article reviews caveats and techniques for translating sequences commonly used on 1.5 or 3 T to high-resolution 7-T imaging. It discusses for several broad disease categories how high-resolution 7-T imaging can advance the understanding of various diseases, improve diagnosis, and guide management.

Smoking status links habenular volume to glycated hemoglobin: Findings from the Human Connectome Project-Young Adult

BACKGROUND

The habenula-pancreas axis regulates the stimulatory effects of nicotine on blood glucose levels and may participate in the emergence of type 2 diabetes in human tobacco smokers. This secondary analysis of young adults from the Human Connectome Project (HCP-YA) evaluated whether smoking status links the relationship between habenular volume and glycated hemoglobin (HbA1c), a marker of long-term glycemic control.

METHODS

Habenula segmentation was performed using a fully-automated myelin content-based approach in HCP-YA participants and the results were inspected visually (n = 693; aged 22-37 years). A linear regression analysis was used with habenular volume as the dependent variable, the smoking-by-HbA1c interaction as the independent variable of interest, and age, gender, race, ethnicity, education, income, employment status, body mass index, and total gray matter volume as covariates.

RESULTS

Habenula volume and HbA1c were similar in smokers and nonsmokers. There was a significant interaction effect (F_{(1, 673)}= 5.03, p = 0.025) indicating that habenular volume was related to HbA1c in a manner that depended on smoking status. Among participants who were smokers (n = 120), higher HbA1c was associated with apparently larger habenular volume (β = 6.74, standard error=2.36, p = 0.005). No such association between habenular volume and HbA1c was noted among participants who were nonsmokers (n = 573).

DISCUSSION

Blood glucose levels over an extended time period, reflected by HbA1c, were correlated with habenular volume in smokers, consistent with a relationship between the habenula and blood glucose homeostasis in smokers. Future studies are needed to evaluate how habenular function relates to glycemic control in smokers and nonsmokers.

Multimodal Investigations of Reward Circuitry and Anhedonia in Adolescent Depression

Depression is a highly prevalent condition with devastating personal and public health consequences that often first manifests during adolescence. Though extensively studied, the pathogenesis of depression remains poorly understood, and efforts to stratify risks and identify optimal interventions have proceeded slowly. A major impediment has been the reliance on an all-or-nothing categorical diagnostic scheme based solely on whether a patient endorses an arbitrary number of common symptoms for a sufficiently long period. This approach masks the well-documented heterogeneity of depression, a disorder that is highly variable in presentation, severity, and course between individuals and is frequently comorbid with other psychiatric conditions. In this targeted review, we outline the limitations of traditional diagnosis-based research and instead advocate an alternative approach centered around symptoms as unique dimensions of clinical dysfunction that span across disorders and more closely reflect underlying neurobiological abnormalities. In particular, we highlight anhedonia-the reduced ability to anticipate and experience pleasure-as a specific, quantifiable index of reward dysfunction and an ideal candidate for dimensional investigation. Anhedonia is a core symptom of depression but also a salient feature of numerous other conditions, and its severity varies widely within clinical and even healthy populations. Similarly, reward dysfunction is a hallmark of depression but is evident across many psychiatric conditions. Reward function is especially relevant in adolescence, a period characterized by exaggerated reward-seeking behaviors and rapid maturation of neural reward circuitry. We detail extensive work by our research group and others to investigate the neural and systemic factors contributing to reward dysfunction in youth, including our cumulative findings using multiple neuroimaging and immunological measures to study depressed adolescents but also trans-diagnostic cohorts with diverse psychiatric symptoms. We describe convergent evidence that reward dysfunction: (a) predicts worse clinical outcomes, (b) is associated with functional and chemical abnormalities within and beyond the neural reward circuitry, (c) is linked to elevated peripheral levels of inflammatory biomarkers, and (d) manifests early in the course of illness. Emphasis is placed on high-resolution neuroimaging techniques, comprehensive immunological assays, and data-driven analyses to fully capture and characterize the complex, interconnected nature of these systems and their contributions to adolescent reward dysfunction.

Reproducibility of automated habenula segmentation via deep learning in major depressive disorder and normal controls with 7 Tesla MRI

The habenula is one of the most important brain regions for investigating the etiology of psychiatric diseases such as major depressive disorder (MDD). However, the habenula is challenging to delineate with the naked human eye in brain imaging due to its low contrast and tiny size, and the manual segmentation results vary greatly depending on the observer. Therefore, there is a great need for automatic quantitative analytic methods of the habenula for psychiatric research purposes. Here we propose an automated segmentation and volume estimation method for the habenula in 7 Tesla magnetic resonance imaging based on a deep learning-based semantic segmentation network. The proposed method, using the data of 69 participants (33 patients with MDD and 36 normal controls), achieved an average precision, recall, and dice similarity coefficient of 0.869, 0.865, and 0.852, respectively, in the automated segmentation task. Moreover, the intra-class correlation coefficient reached 0.870 in the volume estimation task. This study demonstrates that this deep learning-based method can provide accurate and quantitative analytic results of the habenula. By providing rapid and quantitative information on the habenula, we expect our proposed method will aid future psychiatric disease studies.

Abnormal functional connectivity of habenula in untreated patients with first-episode major depressive disorder

Major depressive disorder (MDD) is associated with abnormalities in emotional/cognitive processing and low reward sensitivity. The habenula has a pivotal role in these processes that may contribute to depression. However, there has been little research on the abnormal connectivity between the habenula and whole brain of first-onset MDD. We aimed to explore the differences of functional connectivity between patients and healthy controls using functional magnetic resonance imaging. We used seed-based resting-state fMRI to examine functional connectivity between the habenula and whole-brain in 49 first-episode depressive patients and 25 healthy controls. Compared to controls, patients with MDD demonstrated significant increases in functional connectivity between the habenula and the dorsolateral prefrontal cortex (dlPFC). Furthermore, the receiver operating characteristic (ROC) curve proved that connectivity between the habenula and dlPFC was highly predictive. Additionally, there was a positive correlation between Hamilton Rating Scale for Depression (HAMD) score and functional connectivity between the habenula and right dlPFC. We found that the aberrant functional connectivity to the habenula and dlPFC can distinguish MDD patients from the normal.

Detailed mapping of human habenula resting-state functional connectivity

The habenula (Hb) inhibits dopaminergic reward signaling in response to negative outcomes and has been linked to numerous functional domains relevant to mental health, including reward prediction, motivation, and aversion processing. Despite its important neuroscientific and clinical implications, however, the human Hb remains poorly understood due to its small size and the associated technical hurdles to in vivo functional magnetic resonance imaging (fMRI) investigation. Using high-resolution 3 T fMRI data from 68 healthy young adults acquired through the Human Connectome Project, we developed a rigorous approach for mapping the whole-brain resting-state functional connectivity of the human Hb. Our study combined an optimized strategy for defining subject-level connectivity seeds to maximize Hb blood-oxygen-level-dependent (BOLD) signal sensitivity with high-quality surface-based alignment for robust functional localization and cortical sensitivity. We identified significant positive Hb connectivity with: (i) conserved brainstem targets, including the dopaminergic ventral tegmental area, serotonergic raphe nuclei, and periaqueductal gray; (ii) subcortical structures related to reward and motor function, including the nucleus accumbens, dorsal striatum, pallidum, thalamus, and cerebellum; and (iii) cortical areas associated with the Salience Network and early sensory processing, including the dorsal anterior cingulate, anterior insula, and primary visual and auditory cortices. Hb connectivity was strongly biased towards task-positive brain regions, with weak or negative connectivity observed throughout the task-negative Default Mode Network. Our study provides a detailed characterization of Hb resting-state functional connectivity in healthy young adults, demonstrating both the feasibility and clinical potential of studying the human Hb using high-resolution 3 T fMRI.

Reproducibility of myelin content-based human habenula segmentation at 3 Tesla

In vivo morphological study of the human habenula, a pair of small epithalamic nuclei adjacent to the dorsomedial thalamus, has recently gained significant interest for its role in reward and aversion processing. However, segmenting the habenula from in vivo magnetic resonance imaging (MRI) is challenging due to the habenula's small size and low anatomical contrast. Although manual and semi-automated habenula segmentation methods have been reported, the test-retest reproducibility of the segmented habenula volume and the consistency of the boundaries of habenula segmentation have not been investigated. In this study, we evaluated the intra- and inter-site reproducibility of in vivo human habenula segmentation from 3T MRI (0.7-0.8 mm isotropic resolution) using our previously proposed semi-automated myelin contrast-based method and its fully-automated version, as well as a previously published manual geometry-based method. The habenula segmentation using our semi-automated method showed consistent boundary definition (high Dice coefficient, low mean distance, and moderate Hausdorff distance) and reproducible volume measurement (low coefficient of variation). Furthermore, the habenula boundary in our semi-automated segmentation from 3T MRI agreed well with that in the manual segmentation from 7T MRI (0.5 mm isotropic resolution) of the same subjects. Overall, our proposed semi-automated habenula segmentation showed reliable and reproducible habenula localization, while its fully-automated version offers an efficient way for large sample analysis.

Imaging Habenula Volume in Schizophrenia and Bipolar Disorder

The habenula (Hb), a bilateral nucleus located next to the dorsomedial thalamus, is of particular relevance to psychiatric disorders based on preclinical evidence linking the Hb to depressive and amotivational states. However, studies in clinical samples are scant because segmentation of the Hb in neuroimaging data is challenging due to its small size and low contrast from the surrounding tissues. Negative affective states dominate the clinical course of schizophrenia and bipolar disorder and represent a major cause of disability. Diagnosis-related alterations in the volume of Hb in these disorders have therefore been hypothesized but remain largely untested. To probe this question, we used a recently developed objective and reliable semi-automated Hb segmentation method based on myelin-sensitive magnetic resonance imaging (MRI) data. We ascertained case-control differences in Hb volume from high resolution structural MRI data obtained from patients with schizophrenia (n = 95), bipolar disorder (n = 44) and demographically matched healthy individuals (n = 52). Following strict quality control of the MRI data, the final sample comprised 68 patients with schizophrenia, 32 with bipolar disorder and 40 healthy individuals. Regardless of diagnosis, age, sex, and IQ were not correlated with Hb volume. This was also the case for age of illness onset and medication (i.e., antipsychotic dose and lithium-treatment status). Case-control differences in Hb volume did not reach statistical significance; their effect size (Cohen's d) was negligible on the left (schizophrenia: 0.14; bipolar disorder: -0.03) and small on the right (schizophrenia: 0.34; bipolar disorder: 0.26). Nevertheless, variability in the volume of the right Hb was associated with suicidality in the entire patient sample (ρ = 0.29, p = 0.004) as well as in each patient group (bipolar disorder: ρ = 0.34, p = 0.04; schizophrenia: ρ = 0.25, p = 0.04). These findings warrant replication in larger samples and longitudinal designs and encourage more comprehensive characterization of Hb connectivity and function in clinical populations.