AUTHOR=Li Ying , Qian Linze , Li Gang , Zhang Zhe TITLE=Frequency specificity of aberrant triple networks in major depressive disorder: a resting-state effective connectivity study JOURNAL=Frontiers in Neuroscience VOLUME=17 YEAR=2023 URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2023.1200029 DOI=10.3389/fnins.2023.1200029 ISSN=1662-453X ABSTRACT=

Major depressive disorder (MDD) has been associated with aberrant effective connectivity (EC) among the default mode network (DMN), salience network (SN), and central executive network (CEN)—collectively referred to as triple networks. However, prior research has predominantly concentrated on broad frequency bands (0.01–0.08 Hz or 0.01–0.15 Hz), ignoring the influence of distinct rhythms on triple network causal dynamics. In the present study, we aim to investigate EC alterations within the triple networks across various frequency bands in patients with MDD. Utilizing a data-driven frequency decomposition approach and a multivariate Granger causality analysis, we characterized frequency-specific EC patterns of triple networks in 49 MDD patients and 54 healthy controls. A support vector machine classifier was subsequently employed to assess the discriminative capacity of the frequency-specific EC features. Our findings revealed that, compared to controls, patients exhibited not only enhanced mean EC within the CEN in the conventional frequency band (0.01–0.08 Hz), but also decreased mean EC from the SN to the DMN in a higher frequency band (0.12–0.18 Hz), and increased mean EC from the CEN to the SN in a sub-frequency band (0.04–0.08 Hz); the latter was significantly correlated with disease severity. Moreover, optimal classification performance for distinguishing patients from controls was attained by combining EC features across all three frequency bands, with the area under the curve (AUC) value of 0.8831 and the corresponding accuracy, sensitivity, and specificity of 89.97%, 92.63%, and 87.32%, respectively. These insights into EC changes within the triple networks across multiple frequency bands offer valuable perspectives on the neurobiological basis of MDD and could aid in developing frequency-specific EC features as potential biomarkers for disease diagnosis.