Identify clinical-aware functional connectivity dimensions through contrastive PCA and sparse CCA

Zhu, H., Tong, X., Carlisle, N. B., Xie, H., Keller, C. J., Oathes, D. J., ... & Zhang, Y. (2024). Contrastive functional connectivity defines neurophysiology-informed symptom dimensions in major depression. bioRxiv.

Dataset used in study

MDD patients

• EMBARC dataset

Healthy controls

• EMBARC dataset
• UCLA-CNP dataset
• AOMIC dataset
• LEMON dataset

Usage

• Step 1: prepare healthy control connectivity data, patient connectivity data, and patient clinical assessment data (Example format of clinical assessment data is provided).
• Step 2: Train the model through cpca+scca2-exp.py. Example: python cpca+scca2-exp.py --alpha 2.5 --penaltyu 0.8 --penaltyv 0.8