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

Abstract

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

“…Because of this pattern, any segmentation that is based primarily on anatomical image contrast may underestimate the voxels located at the ventral medial portion of the Hb. This observation is likely to account for the lower in-vivo Hb volumes reported here and in previous studies ( 46 , 47 , 50 – 53 ) compared to ex-vivo estimates ( 31 , 46 , 49 ). The volumetric measurements obtained here did not distinguish between the lateral and medial subdivision of the Hb because such distinction was beyond the resolution of our imaging approach.…”

“…This is the largest neuroimaging investigation of the in vivo Hb morphology in BD and SCZ implemented using strict quality control criteria and an objective semi-automated segmentation method shown to yield consistent Hb boundary definitions ( 47 ). The results suggest that volumetric case-control differences are small but the variability in Hb volume may be linked to the severity of suicidality both in SCZ and BD.…”

“…Information regarding the development and validation of this approach has been described in previous publications ( 46 , 47 ). Specifically, (i) we have corroborated the method used here against Hb boundary definitions derived from in vivo 7T myelin-sensitive images (T1w over T2w ratio images), 7T ex vivo proton density-weighted images and images from the Allen Brain Atlas ( http://www.brain-map.org/ ); (ii) an expert neuroanatomist (Dr. Thomas Naidich) reviewed and approved the segmented Hb boundaries in previous methodological studies ( 46 , 47 ); (iii) we have tested inter-site reliability by obtaining 3T scans from the same healthy individuals ( n = 12) at 3 different sites; (iv) we have affirmed inter-site reliability by obtaining 3T scans from 27 healthy individuals scanned twice at the Icahn School of Medicine at Mount Sinai with an interval of 2 weeks; and (v) we have shown better reliability of our method compared to the geometric method ( 48 ). Briefly, this scheme consists of five steps: (i) Hb region-of-interest initialization, (ii) intensity thresholding, (iii) region growing segmentation, (iv) lateral and inferior geometric constraint, and (v) partial volume estimation ( 46 ).…”

“…The primary aim of the current study was to test for Hb volume reductions in patients with either BD or SCZ compared to healthy individuals using a recently developed objective semi-automated Hb segmentation method with reliable and reproducible boundary definitions ( 46 , 47 ). A secondary goal was to explore the relationship between the Hb volume with the severity of blunted affect, depression, emotional withdrawal and suicidality in patients with SCZ or BD given the association of the Hb with negative affective states.…”

Reproducibility and Clinical Validation of Automated Habenula Segmentation via deep Learning in Major Depressive Disorder with 7 Tesla MRI

“…In recent years, a couple of studies on automated segmentation of the Hb have been reported. The first study performed semi-and fully-automated segmentation in 3T MRI of healthy young adults [14], and the second study performed fully-automated segmentation in children, adolescents, and adults with bipolar disorder and schizophrenia [15]. In the first study, the DSC for binary segmentation reached 0.71 for semi-automated segmentation and 0.69 for fully-automated segmentation, and the DSC of the probability map reached 0.74 for both semi-and fully-automated segmentation [14].…”