Correlated noise in brain magnetic resonance elastography

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Neuroscience

Background:

• Magnetic Resonance Elastography (MRE) visualizes brain mechanical properties using MRI.
• MRE is sensitive to noise from physiological and mechanical sources.
• The impact of correlated noise on brain MRE is not well understood.

Purpose of the Study:

• To characterize signal noise in brain MRE.
• To quantify physiological and vibration noise sources.
• To assess the effect of correlated noise on mechanical property estimation.

Main Methods:

• Conducted three experiments to quantify noise sources.
• Measured physiological noise from cardiac-induced movement.
• Used simulations and a correlation length metric to analyze noise effects.

Main Results:

• Physiological and vibration noise exceeded image noise and were spatially correlated.
• Simulated data with added noise showed higher error in property maps compared to Gaussian noise.
• Correlated noise significantly impacts mechanical property estimates.

Conclusions:

• Established a foundation for understanding brain MRE data quality.
• Identified key noise contributors affecting MRE accuracy.
• Provided recommendations for future noise correction strategies in MRE.