A Reconfigurable Platform for Magnetic Resonance Data Acquisition and Processing.

Area of Science:

• Medical Imaging
• Hardware Engineering

Background:

• Magnetic Resonance Imaging (MRI) systems increasingly use more array coils and sensors, leading to significant data volume challenges.
• Existing MRI data acquisition methods face limitations in throughput and handling large datasets due to increasing radio-frequency (RF) channels and duty cycles.
• Diverse sensor data in MRI requires distinct, low-latency processing for real-time scanner operation adjustments.

Purpose of the Study:

• To propose a novel approach using reconfigurable hardware and a system-on-chip (SoC) for efficient MRI data acquisition and processing.
• To address the growing data handling challenges in advanced MRI by implementing early compression and preprocessing.
• To demonstrate a versatile, integrated hardware solution for real-time processing of diverse sensor data in MRI.

Main Methods:

• Development of an integrated system-on-chip (SoC) solution with tailored interfaces and real-time processing resources.
• Integration of fiber-optical modules for versatile interfacing in modular, in-field MRI systems.
• Demonstration of concurrent imaging and field sensing with hardware-based coil compression and trajectory extraction.

Main Results:

• The proposed SoC platform provides sufficient throughput and hardware-based parallel processing for demanding MRI applications.
• Fiber-optical modules enabled modular systems with effective in-field operation.
• Concurrent imaging and field sensing were successfully demonstrated, showcasing hardware-based coil compression and trajectory extraction.

Conclusions:

• The integrated SoC solution with reconfigurable hardware effectively addresses throughput and data handling limitations in modern MRI.
• The platform facilitates real-time processing of diverse sensor data, enabling advanced imaging techniques like expanded encoding model-based reconstruction.
• This approach enhances MRI capabilities for time-series and anatomical imaging by enabling efficient preprocessing and reconstruction.