New paper is out!
- 19 hours ago
- 2 min read
For years, hemodynamic correction has been considered an important—but largely technical—step in widefield calcium imaging. Most of us assume it simply removes vascular contamination and improves signal quality before the "real" analyses begin.
Our new bioRxiv preprint suggests something much more profound.
We found that hemodynamic signals don't just add noise—they introduce structured biological variance that propagates through nearly every major analytical framework used in mesoscale imaging in neuroscience. As a result, whether or not hemodynamic correction is applied can fundamentally alter how we interpret brain-wide neural activity.
Using dual-wavelength widefield imaging across multiple GCaMP mouse lines, behavioral paradigms, and a glioblastoma model, we systematically evaluated the impact of hemodynamic correction on the analyses that underpin modern widefield imaging studies.
We found that hemodynamic correction consistently reshaped:
🧠 Functional cortical organization
🧠 Functional connectivity and network architecture
🧠 Spectral organization of neural activity
🧠 Brain–behavior coupling
🧠 Low-dimensional cortical state dynamics
Perhaps most strikingly, in glioblastoma, hemodynamic correction uncovered disease-associated network abnormalities that were largely masked in conventional single-wavelength recordings, demonstrating how vascular remodeling can fundamentally influence interpretation of brain dysfunction. Since tumor-associated platelet activation and platelet–tumor interactions promote coagulation, vascular permeability, endothelial dysfunction, and angiogenesis, further reshaping local hemodynamics within the tumor microenvironment, these processes provide an additional biological basis for why vascular signals may become increasingly decoupled from underlying neuronal activity during tumor progression.
To help the field adopt these findings, we also developed HemoTCN, a deep learning framework that predicts hemodynamically corrected calcium signals directly from standard single-wavelength recordings. We hope this will make robust hemodynamic correction accessible not only for future experiments but also for the many existing datasets collected over the past decade.
This project has been several years in the making and represents an incredible team effort. I'm especially proud of Tenesha Connor, who led this work with exceptional rigor and perseverance. Many thanks to all of our co-authors and collaborators whose expertise made this interdisciplinary study possible.
Thanks to The National Institutes of Health Cleveland Clinic Research Case Comprehensive Cancer Center American Brain Tumor Association for their support!
📄 Read the preprint here: https://lnkd.in/gGFfgmGG
Comments