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.

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.

Amyloid-ß prediction machine learning model using source-based morphometry across neurocognitive disorders.

Previous studies have developed and explored magnetic resonance imaging (MRI)-based machine learning models for predicting Alzheimer's disease (AD). However, limited research has focused on models incorporating diverse patient populations. This study aimed to build a clinically useful prediction model for amyloid-beta (Aß) deposition using source-based morphometry, using a data-driven algorithm based on independent component analyses. Additionally, we assessed how the predictive accuracies varied with the feature combinations. Data from 118 participants clinically diagnosed with various conditions such as AD, mild cognitive impairment, frontotemporal lobar degeneration, corticobasal syndrome, progressive supranuclear palsy, and psychiatric disorders, as well as healthy controls were used for the development of the model. We used structural MR images, cognitive test results, and apolipoprotein E status for feature selection. Three-dimensional T1-weighted images were preprocessed into voxel-based gray matter images and then subjected to source-based morphometry. We used a support vector machine as a classifier. We applied SHapley Additive exPlanations, a game-theoretical approach, to ensure model accountability. The final model that was based on MR-images, cognitive test results, and apolipoprotein E status yielded 89.8% accuracy and a receiver operating characteristic curve of 0.888. The model based on MR-images alone showed 84.7% accuracy. Aß-positivity was correctly detected in non-AD patients. One of the seven independent components derived from source-based morphometry was considered to represent an AD-related gray matter volume pattern and showed the strongest impact on the model output. Aß-positivity across neurological and psychiatric disorders was predicted with moderate-to-high accuracy and was associated with a probable AD-related gray matter volume pattern. An MRI-based data-driven machine learning approach can be beneficial as a diagnostic aid.

Intergenerational concordance of brain structure between depressed mothers and their never-depressed daughters.

AIM: Parents have significant genetic and environmental influences, which are known as intergenerational effects, on the cognition, behavior, and brain of their offspring. These intergenerational effects are observed in patients with mood disorders, with a particularly strong association of depression between mothers and daughters. The main purpose of our study was to investigate female-specific intergenerational transmission patterns in the human brain among patients with depression and their never-depressed offspring. METHODS: We recruited 78 participants from 34 families, which included remitted parents with a history of depression and their never-depressed biological offspring. We used source-based and surface-based morphometry analyses of magnetic resonance imaging data to examine the degree of associations in brain structure between four types of parent-offspring dyads (i.e. mother-daughter, mother-son, father-daughter, and father-son). RESULTS: Using independent component analysis, we found a significant positive correlation of gray matter structure between exclusively the mother-daughter dyads within brain regions located in the default mode and central executive networks, such as the bilateral anterior cingulate cortex, posterior cingulate cortex, precuneus, middle frontal gyrus, middle temporal gyrus, superior parietal lobule, and left angular gyrus. These similar observations were not identified in other three parent-offspring dyads. CONCLUSIONS: The current study provides biological evidence for greater vulnerability of daughters, but not sons, in developing depression whose mothers have a history of depression. Our findings extend our knowledge on the pathophysiology of major psychiatric conditions that show sex biases and may contribute to the development of novel interventions targeting high-risk individuals.

Regional Gray Matter Volume Identifies High Risk of Unsafe Driving in Healthy Older People.

In developed countries, the number of traffic accidents caused by older drivers is increasing. Approximately half of the older drivers who cause fatal accidents are cognitively normal. Thus, it is important to identify older drivers who are cognitively normal but at high risk of causing fatal traffic accidents. However, no standardized method for assessing the driving ability of older drivers has been established. We aimed to establish an objective assessment of driving ability and to clarify the neural basis of unsafe driving in healthy older people. We enrolled 32 healthy older individuals aged over 65 years and classified unsafe drivers using an on-road driving test. We then utilized a machine learning approach to distinguish unsafe drivers from safe drivers based on clinical features and gray matter volume data. Twenty-one participants were classified as safe drivers and 11 participants as unsafe drivers. A linear support vector machine classifier successfully distinguished unsafe drivers from safe drivers with 87.5% accuracy (sensitivity of 63.6% and specificity of 100%). Five parameters (age and gray matter volume in four cortical regions, including the left superior part of the precentral sulcus, the left sulcus intermedius primus [of Jensen], the right orbital part of the inferior frontal gyrus, and the right superior frontal sulcus), were consistently selected as features for the final classification model. Our findings indicate that the cortical regions implicated in voluntary orienting of attention, decision making, and working memory may constitute the essential neural basis of driving behavior.

Fluorescence-labeled peptide pI markers for capillary isoelectric focusing.

Nineteen fluorescent pH standards or pI markers ranging pH 3.64-10.12 were developed for use in capillary isoelectric focusing using laser-induced fluorescence detection. Tetra- to tridecapeptides containing one cysteine residue were designed to focus sharply at their respective isoelectric points by including amino acids that contain charged side chains, the pKa values of which are close to the corresponding pI values. An iodoacetylated derivative of tetramethylrhodamine was coupled to the thiol group of cysteine to yield fluorescent pI markers. The pI values of the labeled peptides were precisely determined after isoelectric focusing on polyacrylamide gel slabs by direct measurement of the pH of the focused bands. The markers were subjected to capillary isoelectric focusing for 10-15 min in coated capillaries under conditions of low electroosmosis and were detected by means of a scanning laser-induced fluorescence detector down to a level of subpicomolar range. The markers permitted the calibration of a wide-range pH gradient formed in a capillary by fluorescence detection for the first time and should facilitate the development of highly sensitive analytical methods based on a combination of capillary isoelectric focusing and laser-induced fluorescence detection.

Immunoassay of serum alpha(1)-antitrypsin by affinity-probe capillary isoelectric focusing using a fluorescence-labeled recombinant antibody fragment.

An immunoassay for human alpha(1)-antitrypsin based on affinity-probe capillary isoelectric focusing (AP-CIEF) is described. The method is based on the separation of free and bound labeled antibody fragments by CIEF with laser-induced fluorescence detection. A recombinant Fab' fragment of mouse immunoglobulin G1 (IgG1) against human alpha(1)-antitrypsin was labeled with tetramethylrhodamine on the single cysteine residue at the hinge region. A single pI isomer of the labeled Fab' was purified by IEF in a slab of agarose gel and was then used as the affinity probe for alpha(1)-antitrypsin. The use of recombinant Fab' considerably simplified the labeling process. Although there was some difficulty in the quantification of native alpha(1)-antitrypsin with the affinity probe, carbamylation of the antigen sample by heat treatment with urea made the complex peaks appear reproducibly and more distinct, thus facilitating the identification and quantification of the complex. The system provided an almost linear response to a pure sample of alpha(1)-antitrypsin over a concentration range of 5-1000 ng/mL and the detection limit extended down to around 2 ng/mL. Alpha(1)-antitrypsin in a serum sample was determined using this system to be 1.2 mg/mL which is comparable to the reported value for the protein.