top of page
How did we design NEUROFUSE?
Design
Design 1:
Pain as a Biomarker
-
Focused on correlation with disease
-
Descriptive, retrospective
-
❌ No targeting, no regulation
Design 2:
Simulated Metabolic Models
-
Introduced machine learning
-
Relied on simulated mitochondrial dysfunction
-
❌ Models learned simulation artifacts, not physiology
Design 3:
Imaging + ML
-
Used MRSI features
-
Treated brain regions independently
-
❌ No modeling of network propagation
Final Design:
NEUROFUSE
-
Combines metabolism + connectivity
-
Identifies drivers vs bystanders
-
Enables closed-loop adaptation
Required Breakthroughs
-
Higher-resolution MRSI
-
More precise connectome mapping
-
Improved interpretability between signatures and biology
-
Longitudinal stability validation
bottom of page