Developing an AI-empowered head-only ultra-high-performance gradient MRI system for high spatiotemporal neuroimaging

Area of Science:

• Neuroimaging
• Magnetic Resonance Imaging (MRI)
• Gradient Coil Systems

Background:

• High-resolution MRI is crucial for understanding brain structure and function.
• Advancements in gradient systems are needed for improved spatial and temporal encoding in neuroimaging.

Purpose of the Study:

• To introduce NeuroFrontier, a novel head-only gradient system designed for ultra-high performance neuroimaging.
• To demonstrate the system's capabilities in enhancing image resolution, acquisition speed, and signal-to-noise ratio.

Main Methods:

• Development of a head-only gradient system with 650 mT/m gradient strength and 600 T/m/s slew rate.
• Utilized a 7MW ultra-high power system via novel amplifier paralleling.
• Employed a force/torque balanced coil design and a high-density integrated RF system.
• Integrated AI-empowered compressed sensing for accelerated acquisition and prospective head motion correction using 3D optical tracking.

Main Results:

• Demonstrated superior imaging resolution, speed, and signal-to-noise ratio for structural MRI (sMRI), functional MRI (fMRI), and diffusion MRI (dMRI).
• Achieved submillimeter layer-specific fMRI and dMRI.
• Showcased enhanced lesion contrast and improved microstructural parameter estimation in dMRI.

Conclusions:

• The NeuroFrontier system significantly advances neuroimaging capabilities for neuroscience applications.
• The system enables ultra-fast, high-resolution brain imaging, facilitating detailed microstructural mapping and analysis.