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

• Published in: Biological psychiatry global open science Vol. 4; no. 4; p. 100314
• Main Authors: Kyuragi, Yusuke, Oishi, Naoya, Hatakoshi, Momoko, Hirano, Jinichi, Noda, Takamasa, Yoshihara, Yujiro, Ito, Yuri, Igarashi, Hiroyuki, Miyata, Jun, Takahashi, Kento, Kamiya, Kei, Matsumoto, Junya, Okada, Tomohisa, Fushimi, Yasutaka, Nakagome, Kazuyuki, Mimura, Masaru, Murai, Toshiya, Suwa, Taro
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.07.2024; Elsevier
• Subjects: Archival Report; Deep learning; Depression; Habenula; Image analysis; Psychiatric/Mental Health; Sex differences; Structural MRI; Deep learning; Structural MRI; Depression; Sex differences; Image analysis; Habenula; TIV; MPRAGE; magnetic resonance imaging; MRI; inversion time; total intracranial volume; ANCOVA; analysis of variance; ANOVA; FDR; MP2RAGE; echo time; HDRS; false discovery rate; Sex Differences; repetition time; magnetization prepared rapid acquisition with a gradient echo; Image Analysis; Deep Learning; TE; TI; magnetization prepared 2 rapid acquisition gradient echo; analysis of covariance; Hamilton 17-item depression rating scale; TR
• ISSN: 2667-1743; 2667-1743
• DOI: 10.1016/j.bpsgos.2024.100314

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. 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. 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). Habenula volume could be a pathophysiologically relevant factor and diagnostic and therapeutic marker for depression, particularly in women. Accurate segmentation of the habenula, a brain region implicated in depression, is challenging. In this study, we developed an automated human habenula segmentation model using deep learning techniques. The model was confirmed to be reproducible and generalizable at various spatial resolutions. Application of this model to a multicenter dataset confirmed that habenula volume decreased with age in healthy volunteers, an association that was more pronounced in individuals with depression. In addition, habenula volume decreased with the severity of depression in women. This novel model for habenula segmentation enables further study of the role of the habenula in depression.