Abnormal degree centrality values as a potential imaging biomarker for major depressive disorder: A resting-state functional magnetic resonance imaging study and support vector machine analysis

Abstract

Objective: Previous studies have revealed abnormal degree centrality (DC) in the structural and functional networks in the brains of patients with major depressive disorder (MDD). There are no existing reports on the DC analysis method combined with the support vector machine (SVM) to distinguish patients with MDD from healthy controls (HCs). Here, the researchers elucidated the variations in DC values in brain regions of MDD patients and provided imaging bases for clinical diagnosis.

Methods: Patients with MDD (N = 198) and HCs (n = 234) were scanned using resting-state functional magnetic resonance imaging (rs-fMRI). DC and SVM were applied to analyze imaging data.

Results: Compared with HCs, MDD patients displayed elevated DC values in the vermis, left anterior cerebellar lobe, hippocampus, and caudate, and depreciated DC values in the left posterior cerebellar lobe, left insula, and right caudate. As per the results of the SVM analysis, DC values in the left anterior cerebellar lobe and right caudate could distinguish MDD from HCs with accuracy, sensitivity, and specificity of 87.71% (353/432), 84.85% (168/198), and 79.06% (185/234), respectively. Our analysis did not reveal any significant correlation among the DC value and the disease duration or symptom severity in patients with MDD.

Conclusion: Our study demonstrated abnormal DC patterns in patients with MDD. Aberrant DC values in the left anterior cerebellar lobe and right caudate could be presented as potential imaging biomarkers for the diagnosis of MDD.

Keywords: biomarker; degree centrality; major depressive disorder; rest state fMRI; support vector machine.

FIGURE 1

Statistical map depicts higher and lower degree centrality (DC) of major depressive disorder (MDD) patients compared with healthy controls (HCs). The threshold was set at p < 0.01. Blue denotes lower DC, and red denotes higher DC. The color bar indicates t-values from the two-sample t-test.

FIGURE 2

Visualization of classifications through support vector machine (SVM) using the combination of the degree centrality (DC) values in the left cerebellum anterior and right caudate. Left: SVM parameter result of 3D view. Right: Dimension 1 and dimension 2 represent the DC values in the left cerebellum anterior and right caudate, respectively. Red crosses represent healthy controls (HCs), and green crosses represent patients with major depressive disorder (MDD).