Citation

Yan H, Han Y, Shan X, Li H, Liu F, Li P, Zhao J, Guo W. Biomedicines 2023; 11(9).

Copyright

(Copyright © 2023, MDPI: Multidisciplinary Digital Publications Institute)

DOI

10.3390/biomedicines11092420

PMID

37760861

PMCID

PMC10525800

Abstract

Panic disorder (PD) is a prevalent type of anxiety disorder. Previous studies have reported abnormal brain activity in the fear network of patients with PD. Nonetheless, it remains uncertain whether pharmacotherapy can effectively normalize these abnormalities. This longitudinal resting-state functional magnetic resonance imaging study aimed to investigate the spontaneous neural activity in patients with PD and its changes after pharmacotherapy, with a focus on determining whether it could predict treatment response. The study included 54 drug-naive patients with PD and 54 healthy controls (HCs). Spontaneous neural activity was measured using regional homogeneity (ReHo). Additionally, support vector regression (SVR) was employed to predict treatment response from ReHo. At baseline, PD patients had aberrant ReHo in the fear network compared to HCs. After 4 weeks of paroxetine treatment (20 mg/day), a significant increase in ReHo was observed in the left fusiform gyrus, which had shown reduced ReHo before treatment. The SVR analysis showed significantly positive correlations (p < 0.0001) between the predicted and actual reduction rates of the severity of anxiety and depressive symptoms. Here, we show patients with PD had abnormal spontaneous neural activities in the fear networks. Furthermore, these abnormal spontaneous neural activities can be partially normalized by pharmacotherapy and serve as candidate predictors of treatment response. Gaining insight into the trajectories of brain activity normalization following treatment holds the potential to provide vital insights for managing PD.

Language: en

Keywords

brain; machine learning; longitudinal studies; magnetic resonance imaging; anxiety; biomarker; fear; follow-up studies; panic disorder; paroxetine