Disrupted brain connectivity in postpartum depression: Insights from resting-state fMRI and machine learning

Area of Science:

• Neuroscience
• Psychiatry
• Medical Imaging

Background:

• Postpartum depression (PPD) is a prevalent women's health issue.
• Previous research identified regional brain abnormalities in PPD.
• Global functional topological alterations in PPD require further investigation.

Purpose of the Study:

• To investigate alterations in functional topological properties of the brain in PPD patients.
• To characterize global functional connectome changes associated with PPD.
• To differentiate PPD from healthy controls using neuroimaging.

Main Methods:

• Resting-state functional MRI (rs-fMRI) data from PPD patients, healthy pregnant women (HPW), and healthy controls (HC).
• Construction and graph theory analysis of functional brain networks using Pearson's correlation.
• Application of machine learning to functional connectome data for classification.

Main Results:

• PPD patients exhibited a more regularized functional brain network topology compared to HC and HPW.
• PPD showed altered topological properties in the salience network (SN) and subcortical regions (e.g., insula, amygdala).
• Machine learning accurately predicted PPD with 88% accuracy based on functional connectivity.

Conclusions:

• Pregnancy impacts the default mode network (DMN), while PPD disrupts the salience network (SN).
• The insula and precuneus are key regions for distinguishing PPD from healthy individuals.
• Functional connectome alterations are clinically significant for early PPD detection.