Abstract
IntroductionDeveloping a machine learning-based approach which could provide quantitative identification of major depressive disorder (MDD) is essential for the diagnosis and intervention of this disorder. However, the performances of traditional algorithms using static functional connectivity (SFC) measures were unsatisfactory. In the present work, we exploit the hidden information embedded in dynamic functional connectivity (DFC) and developed an accurate and objective image-based diagnosis system for MDD.MethodsMRI images were collected from 99 participants including 56 healthy controls and 43 MDD patients. DFC was calculated using a sliding-window algorithm. A non-linear support vector machine (SVM) approach was then used with the DFC matrices as features to distinguish MDD patients from healthy controls. The spatiotemporal characteristics of the most discriminative features were then investigated.ResultsThe area under the curve (AUC) of the SVM classifier with DFC measures reached 0.9913, while this value is only 0.8685 for the algorithm using SFC measures. Spatially, the most discriminative 28 connections distributed in the visual network (VN), somatomotor network (SMN), dorsal attention network (DAN), ventral attention network (VAN), limbic network (LN), frontoparietal network (FPN), and default mode network (DMN), etc. Notably, a large portion of these connections were associated with the FPN, DMN, and VN. Temporally, the most discriminative connections transited from the cortex to deeper regions.ConclusionThe results clearly suggested that DFC is superior to SFC and provide a reliable quantitative identification method for MDD. Our findings may furnish a better understanding of the neural mechanisms of MDD as well as improve accurate diagnosis and early intervention of this disorder.
Highlights
Developing a machine learning-based approach which could provide quantitative identification of major depressive disorder (MDD) is essential for the diagnosis and intervention of this disorder
Neuroimaging studies based on functional magnetic resonance imaging have provided rich evidence of abnormalities in neural activity and functional connectivity of multiple brain regions and networks of patients with Major depressive disorder (MDD), including cingulate cortex, precuneus, and medial prefrontal cortex of the default mode network (DMN), dorsolateral prefrontal cortex of the central executive network (CEN), insula of the salience network and the amygdala, hippocampus, etc. (Hamilton et al, 2015; Mulders et al, 2015; Otte et al, 2016; Ambrosi et al, 2017)
We compared the performances of the system using features obtained from dynamic functional connectivity (DFC) matrices with that of the system using features extracted from traditional static functional connectivity (SFC) matrices
Summary
Developing a machine learning-based approach which could provide quantitative identification of major depressive disorder (MDD) is essential for the diagnosis and intervention of this disorder. Developing a machine learning-based approach, which could possibly realize the quantitative characterization of the brain imaging data and achieve an objective prediction of the brain disorders (Gong and He, 2015), deserves more attention. (Hamilton et al, 2015; Mulders et al, 2015; Otte et al, 2016; Ambrosi et al, 2017) These findings collectively point toward the fact that aberrant functional connectivity can be used as an imaging metric to provide new opportunities for accurate diagnosis of MDD. After leave-one-out cross validation (LOOCV), it achieved an accuracy over 70% by support vector machine (SVM) or partial least squares (PLSs) classifiers (Cao et al, 2014; Bhaumik et al, 2016; Yoshida et al, 2017)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
