Disease-specific network pattern of perinatal depression revealed by Common Orthogonal Basis Extraction

Area of Science:

• Neuroscience
• Medical Imaging
• Psychiatry

Background:

• Clinical diagnosis of perinatal depression (PD) is challenging due to overlapping symptoms with normal pregnancy mood changes and altered brain activity.
• Traditional neuroimaging methods struggle to identify specific neural markers for PD.

Purpose of the Study:

• To develop an innovative approach combining brain network analysis and Common Orthogonal Basis Extraction (COBE) to identify a PD-Specific Network Pattern.
• To validate the efficacy of this pattern in the diagnosis and assessment of PD.

Main Methods:

• Resting-state electroencephalography (EEG) data were collected from patients with PD and healthy pregnant (HP) individuals.
• Functional brain networks were constructed and analyzed using an optimized COBE method to extract group-specific and common network patterns.
• PD-Specific Features were derived from the identified PD-Specific Network Pattern.

Main Results:

• An optimized COBE method successfully identified a PD-Specific Network Pattern.
• Machine learning models (support vector machine, multiple linear regression) trained on PD-Specific Features demonstrated efficacy in individual-level classification and assessment of PD.
• The study successfully addressed limitations of traditional neuroimaging in PD diagnosis.

Conclusions:

• The developed PD-Specific Network Pattern offers a novel approach for objective screening and dynamic monitoring of perinatal depression.
• This method provides a new avenue for improving the accuracy and efficiency of PD diagnosis.
• The findings highlight the potential of advanced brain network analysis in understanding and managing perinatal mood disorders.