Spooky action at a distance of the mouse brain. Our new preprint shows that brain metabolism is organized as a coherent network. Local intervention in one region can normalize metabolism across distant but metabolically similar regions — even in an Alzheimer’s model — without direct gene or plaque changes there. We don’t yet know the mechanism, but we tested it with two different challenges and saw clear network-level metabolic normalization. How we got there (MALDI imaging → conclusion): Mapped the spatial metabolome across 12 major mouse brain divisions using MALDI imaging + Allen Brain Atlas registration. Defined metabolic coherence as an optimal-transport-derived inter-regional metabolite similarity metric. Hippocampus emerged as a hub with high similarity to cortex, thalamus, hypothalamus, etc. In the 5xFAD amyloid model, individual metabolites and lipids changed dramatically (mitochondrial dysfunction signatures), yet the overall inter-regional coherence structure stayed largely preserved. Metabolites shifted in a coordinated way that maintained network relationships. Targeted the left hippocampus with two distinct local interventions: lentiviral shHIF1α knockdown and neuronal AAV-AOX expression. Both challenges normalized metabolites at the injection site. More strikingly, normalization extended to distal regions sharing high metabolic coherence — even though gene modulation and amyloid plaque reduction remained strictly local. Network-level metabolic rescue was accompanied by improved social memory. Coherence itself was preserved. Bottom line: Metabolic coherence functions as essential “spooky action at a distance” — a network property that allows coordinated metabolic responses across anatomically distant brain regions. This opens new ways of thinking about how local perturbations (or therapies) can influence distributed brain networks in health and disease. biorxiv.org/content/10.64898…

🧠🏃 New preprint exercise neuroscience from the lab! We built a multi-organ spatial metabolomic atlas of long-term exercise in mice. Six organs. 224 metabolic features. Coordinated inter-organ remodeling. The biggest surprise? The brain reshaped its metabolism the most, region by region. We took the same approach into PS19 tau mice. Exercise cut over 70% of observable tau pathology in the hippocampus and restored the mitochondrial metabolome. Integrated proteomics + spatial metabolomics pointed to one convergent node: NADH dehydrogenase Complex I. So we tested it. We expressed yeast Ndi1 in PS19 neurons. No exercise. This alone raised brain antioxidants, restored shuttle metabolites, and reduced tau pathology. Boosting Complex I activity reproduces the core anti-tau effects of exercise. A molecular handle on why movement may protect the aging brain. biorxiv.org/content/10.64898…

I wanted to do a threadorial on our new study: "Hyperglycosylation is a metabolic driver of Alzheimer’s disease" highlighting new tech, mouse models, spatial isotopic tracing, mechanistic insights and the connection to AI/EHR. nature.com/articles/s42255-0…

1/6 we optimized a MALDI-MSI workflow to perform spatial metabolomics + lipidomics + glycomics on the same fresh-frozen human post-mortem brain section. Key finding: Dramatic hyperglycosylation (excess N-glycans) in AD grey matter — a clear, progressive, and notable metabolic phenotype of the disease.