The paper introduces MRx, a groundbreaking multiplexed magnetic resonance imaging technique designed to simultaneously map 22 different structural, physiological, and molecular biomarkers in the brain. Unlike traditional MRI, which primarily offers qualitative visual scans of water molecules, MRx utilizes wideband excitation and sparse sampling to capture high-resolution data on a vast array of metabolites and tissue properties. By integrating physics-based machine learning and neural networks, the system efficiently reconstructs and quantifies markers for neuronal integrity, metabolism, and oxygenation within a single 14-minute scan. Clinical studies demonstrate that MRx can create a comprehensive tissue state index, significantly improving the ability of clinicians to subtype brain tumors and characterize multiple sclerosis lesions. This innovation moves neuroimaging toward precision medicine by providing objective, quantitative tools for disease diagnosis and therapeutic monitoring. The technology represents a shift from simple anatomical observation to deep, multi-parametric tissue characterization at various spatial scales.
References:
Li Y, Guo R, Zhao Y, et al. Multiplexed magnetic resonance imaging[J]. Nature, 2026: 1-7.

