NASA’s new dark energy space telescope can also detect killer asteroids
technologyreview.com·3h ago
AI Research
arXiv CS.AI·Jun 12·5 min read
Fusion Learning from Dynamic Functional Connectivity: Combining the Amplitude and Phase of fMRI Signals to Identify Brain Disorders
TL;DR
Existing methods for identifying brain disorders using fMRI signals often overlook critical information. A multi-scale fusion learning framework (MSFL) was developed to integrate both amplitude and phase data from dynamic functional connectivity (dFC).
✦ Why It Matters
Engineers and researchers can leverage MSFL to enhance diagnostic accuracy for brain disorders using fMRI data.
Key Takeaways
- Utilize MSFL for enhanced brain disorder classification.
- Incorporate both amplitude and phase data in analyses.
- Leverage SHAP for model explanation and feature importance.
- Test MSFL on diverse datasets for broader applicability.
How It Works
MSFL combines two types of dynamic functional connectivity features: amplitude correlations from sliding window correlation (SWC) and phase coherence from phase synchronization (PS). This dual approach allows for a more comprehensive analysis of brain activity, capturing both the strength and timing of neural signals.