High-resolution 3D flow reconstruction of cerebrospinal fluid microcirculation using physics-informed neural network: Conceptualization and application to a large animal model
Computer Methods and Programs in Biomedicine · 2026-08 · 283 · 109405
DOI: 10.1016/j.cmpb.2026.109405 ↗Kurz gefasst
Fachartikel in der Zeitschrift Computer Methods and Programs in Biomedicine, 2026-08. Verfasst von Mark Epshtein und 6 weiteren Autor:innen. Beteiligte Einrichtung: University of Massachusetts Chan Medical School, Department of Radiology, New England Center for Stroke Research, Worcester, MA 01655, USA. Hauptthemen laut MEDLINE: Cerebrospinal Fluid, Imaging, Three-Dimensional, Neural Networks, Computer, Microcirculation. Die Arbeit nennt 37 Literaturangaben.
Abstract (Originalsprache)
In this study, we present a novel Physics-Informed Neural Network (PINN) framework that reconstructs 3D flows using planar velocity projections from arbitrarily oriented planes. The method is designed for the reconstruction of low-Reynolds-number flows typical of cerebrospinal fluid (CSF) in the subarachnoid space (SAS). The method utilizes a projection loss function combined with gradient smoothing regularization during network training. We show that plane orientation with perturbations of 0.05 (relative to the main flow axis) or greater is sufficient axial data for accurate reconstruction. Additionally, gradient exponential moving average smoothing with amplification improves convergence and stability, particularly for near-parallel planes of acquisition. The method was compared against computational fluid dynamics (CFD) data and applied to flow in a realistic canine SAS geometry derived from intravascular optical coherence tomography (OCT), demonstrating the framework's potential for in-vivo CSF flow imaging.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- University of Massachusetts Chan Medical School, Department of Radiology, New England Center for Stroke Research, Worcester, MA 01655, USA.
- Technion Israel Institute of Technology, Department of Biomedical Engineering, Haifa 32000, Israel.
- University of Massachusetts Chan Medical School, Department of Radiology, New England Center for Stroke Research, Worcester, MA 01655, USA. Electronic address: matthew.gounis@umassmed.edu.
Zitieren
APA
Epshtein, M., Mekler, T., Shazeeb, M. S., Lindsay, C., Gounis, M. J., Korin, N., & Anagnostakou, V. (2026). High-resolution 3D flow reconstruction of cerebrospinal fluid microcirculation using physics-informed neural network: Conceptualization and application to a large animal model. Computer Methods and Programs in Biomedicine, 283, 109405. https://doi.org/10.1016/j.cmpb.2026.109405
Vancouver
Epshtein M, Mekler T, Shazeeb MS, Lindsay C, Gounis MJ, Korin N, et al. High-resolution 3D flow reconstruction of cerebrospinal fluid microcirculation using physics-informed neural network: Conceptualization and application to a large animal model. Computer Methods and Programs in Biomedicine. 2026;283:109405. doi:10.1016/j.cmpb.2026.109405
BibTeX
@article{Epshtein2026,
author = {Epshtein, Mark and Mekler, Tirosh and Shazeeb, Mohammed Salman and Lindsay, Clifford and Gounis, Matthew J and Korin, Netanel and Anagnostakou, Vania},
title = {High-resolution 3D flow reconstruction of cerebrospinal fluid microcirculation using physics-informed neural network: Conceptualization and application to a large animal model},
journal = {Computer Methods and Programs in Biomedicine},
year = {2026},
volume = {283},
pages = {109405},
doi = {10.1016/j.cmpb.2026.109405}
}
EUVIMED