Acquisition Technology for Functional and Quantitative MRI at the Mesoscopic Scale

Massachusetts General Hospital, Boston MA

Investigators

Linked publications, trials & patents

Abstract

MRI is today the predominant tool for noninvasive imaging of brain function and structure, which has revolutionized our understanding of the human brain and has played an integral role in large-scale human neuroscience studies. However, conventional MRI, limited to macroscopic spatial resolution, cannot resolve brain activity and structure at finer levels such as those across cortical columns and layers, leading to the gap between human MRI with the microscopic measurements obtained in animal models or ex vivo. Achieving mesoscopic functional and structural MRI in humans is crucial for bridging this gap for neuroscience, offering the potential to link our understanding of the brain systems with circuit-level insights. This would open new avenues for investigating cortical architecture across thin layers and columns, advancing our ability to decipher the brain's hierarchical organization. Despite these promises, significant technical barriers persist for MRI to achieve mesoscopic resolution in vivo due to intrinsic limitations of current MRI acquisitions, including limited SNR, system imperfections, and susceptibility to subject motion and physiological noise. In this project, we will develop next-generation acquisition and reconstruction technologies to achieve functional (fMRI) and quantitative MRI (qMRI) at the mesoscopic scale. A series of synergistic acquisition, reconstruction, and hardware technologies, leveraging new ways of data encoding and image formation, will be developed to achieve ultra-high spatiotemporal resolution. These technologies will be further integrated with techniques to address challenges that impede true spatiotemporal accuracy, such as motion and physiology-induced field variations. Additionally, we will address the prominent image artifacts caused by system imperfections at the acquisition level to ensure high-quality data. Finally, contrast encoding or source separation algorithms will be deployed to enable the extraction of functional and anatomical information with high spatiotemporal and neurobiological specificity. To maximize the impact, we will disseminate our acquisition, reconstruction, and processing software accessible on widely available clinical scanners to the broad neuroimaging community.