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.

Information flow and dynamic functional connectivity during electroconvulsive therapy in patients with depression.

BACKGROUND: Electroconvulsive therapy is effectively used for treatment-resistant depression; however, its neural mechanism is largely unknown. Resting-state functional magnetic resonance imaging is promising for monitoring outcomes of electroconvulsive therapy for depression. This study aimed to explore the imaging correlates of the electroconvulsive therapy effects on depression using Granger causality analysis and dynamic functional connectivity analyses. METHODS: We performed advanced analyses of resting-state functional magnetic resonance imaging data at the beginning and intermediate stages and end of the therapeutic course to identify neural markers that reflect or predict the therapeutic effects of electroconvulsive therapy on depression. RESULTS: We demonstrated that information flow between the functional networks analyzed by Granger causality changes during electroconvulsive therapy, and this change was correlated with the therapeutic outcome. Information flow and the dwell time (an index reflecting the temporal stability of functional connectivity) before electroconvulsive therapy are correlated with depressive symptoms during and after treatment. LIMITATIONS: First, the sample size was small. A larger group is needed to confirm our findings. Second, the influence of concomitant pharmacotherapy on our results was not fully addressed, although we expected it to be minimal because only minor changes in pharmacotherapy occurred during electroconvulsive therapy. Third, different scanners were used the groups, although the acquisition parameters were the same; a direct comparison between patient and healthy participant data was not possible. Thus, we presented the data of the healthy participants separately from that of the patients as a reference. CONCLUSIONS: These results show the specific properties of functional brain connectivity.

The Japanese version of the Generalized Problematic Internet Use Scale 2 (GPIUS2): Psychometric evaluation and analysis of the theoretical model.

BACKGROUND: The Generalized Problematic Internet Use Scale 2 (GPIUS2) is a self-administered questionnaire that evaluates problematic internet use (PIU) from a multidimensional perspective. We analysed the psychometric properties and adequacy of the theoretical model of Japanese version of the GPIUS2. METHODS: This study included 291 healthy Japanese adults (median age = 25 years; interquartile range 22-43 years; 128 women) who completed the GPIUS2 and several other questionnaires evaluating the degree of PIU, self-esteem, depression, and impulsivity. RESULTS: Exploratory factor analysis (EFA) revealed a similar factor structure between the original and Japanese versions of the GPIUS2, with only minor differences in item composition. Higher-order confirmatory factor analyses revealed a good overall fit for the factorial model suggested by EFA, indicating adequate construct validity. The model showed acceptable internal consistency. Partial correlation analyses between GPIUS2 and other measures, with age as a control variable, revealed good convergent validity. Finally, structural equation modelling showed a good fit to the data, supporting the cognitive-behavioural model of Caplan (2010). CONCLUSIONS: The Japanese version of the GPIUS2 has good psychometric properties and the theoretical model of the original GPIUS2 is applicable to Japanese adults.

Additive value of "otosclerosis-weighted" images for the CT diagnosis of fenestral otosclerosis.

Background Otosclerotic foci are usually seen as minute low-density lesions and this may be attributed to relatively low sensitivity on visual assessment using computed tomography (CT). Otosclerotic foci can be detected by using the accurate region of interest (ROI) setting, while small ROI settings by less-experienced radiologists may result in false negative findings. Purpose To evaluate the diagnostic ability of our proposed method ("otosclerosis-weighted" imaging [OWI]), which is based on reversing the density, compared with conventional CT (CCT) imaging alone. Material and Methods Temporal bone CTs of consecutive patients with otosclerosis were analyzed. Gender- and age-matched control participants were also included. All CT images were obtained using a 64-detector row scanner. OWI was obtained by extracting the temporal bone region using the threshold technique and reversing the density (black to white). Four independent radiologists took part in two reading sessions. In the first session, the observers read only CCT imaging. In the second session, they read OWI along with the CCT imaging. Sensitivity was assessed for the four readers. Results Thirty temporal bones of 25 patients with otosclerosis (3 men, 22 women; mean age, 53.9 ± 9.0 years) and 30 temporal bones of 30 control participants (4 men, 26 women; mean age, 44.0 ± 16.2 years) were included. For all observers, reading with a combination of the two methods was associated with a higher sensitivity (63.3-80.0%) than with conventional CT images alone (30.0-60.0%; P < 0.05, each). Conclusion Application of our proposed method based on threshold value may help detect foci of fenestral otosclerosis.