Speech production real-time MRI at 0.55 T

Area of Science:

• Medical Imaging
• Biophysics
• Speech Science

Background:

• Real-time MRI (RT-MRI) is crucial for understanding speech articulation.
• Conventional MRI field strengths (1.5T, 3T) present limitations in speed and image quality for dynamic speech tasks.
• Contemporary lower field strength systems offer potential for improved RT-MRI performance.

Purpose of the Study:

• To demonstrate speech-production RT-MRI using a 0.55T MRI system.
• To evaluate performance trade-offs of different imaging parameters.
• To compare 0.55T system capabilities with conventional higher field strengths.

Main Methods:

• Healthy adult volunteers underwent RT-MRI on a 0.55T system with high-performance gradients and an 8-channel upper airway coil.
• Spiral-based balanced SSFP (bSSFP) and gradient-recalled echo (GRE) pulse sequences were employed with temporal finite-difference constrained reconstruction.
• Three spiral readout durations (8.90, 5.58, 3.48 ms) were tested to assess articulator contrast, blurring, and artifacts.

Main Results:

• Both spiral GRE and bSSFP captured tongue boundary dynamics during rapid speech.
• Spiral bSSFP with the shortest readout (3.48 ms, TR=5.30 ms) yielded the best image quality, offering a 1.54x SNR boost over GRE.
• bSSFP provided higher SNR but was susceptible to banding artifacts at TR > 10.9 ms; longer readouts increased SNR efficiency but also blurring.

Conclusions:

• High-performance 0.55T MRI systems are viable for speech-production RT-MRI.
• Spiral bSSFP can be utilized effectively for vocal tract imaging without banding artifacts.
• The 0.55T system demonstrates superior SNR efficiency and image quality compared to 1.5T or 3T systems for speech RT-MRI.