Reverse Engineering for Robust MDD Biomarkers

Project Overview

This project applies a reverse engineering approach to identify robust and reliable biomarkers for Major Depressive Disorder (MDD). By integrating multi-modal data (neuroimaging, clinical, behavioral) and employing advanced computational techniques, we aim to uncover underlying biological signatures that can improve diagnosis, predict treatment response, and stratify patient subgroups.

Key Research Questions

• Can data-driven reverse engineering identify novel, replicable biomarkers for MDD?
• How do these biomarkers relate to specific symptom dimensions or clinical subtypes of depression?
• Can these biomarkers predict individual response to different antidepressant treatments?
• What are the underlying neurobiological pathways associated with these robust biomarkers?

Methodology

We utilize large-scale datasets containing neuroimaging (fMRI, sMRI, DTI), clinical symptom scores, cognitive assessments, and genetic information. Our methods include advanced machine learning (e.g., deep learning, dimensionality reduction), causal inference techniques, and network analysis to model complex interactions and identify stable biomarkers.

Current Status

We are currently developing and validating computational pipelines for biomarker discovery. Initial applications have identified promising candidate biomarkers related to specific neural circuits (e.g., default mode network, salience network) and cognitive functions often impaired in MDD.