Cardiac-induced physiological noise in 3D gradient echo brain imaging: effect of k-space sampling scheme

Area of Science:

• Medical Imaging
• Biophysics
• Signal Processing

Background:

• Physiological noise is a significant challenge in 3D image acquisition, affecting image quality and diagnostic accuracy.
• Understanding the impact of different sampling schemes on physiological noise is crucial for optimizing imaging protocols.

Purpose of the Study:

• To investigate how various k-space sampling schemes influence physiological noise in 3D Echo Planar Imaging (EPI) for functional MRI (fMRI).
• To model and analyze the propagation of physiological noise in reconstructed images.

Main Methods:

• Comparison of five sampling schemes: Linear, Centric, Segmented, Random, and Tuned acquisition.
• Modeling physiological noise as periodic temporal oscillations and analyzing its in-phase and orthogonal components in reconstructed images.
• Numerical simulations, analytical derivations, and experimental measurements (3D noise and high temporal resolution 2D) under breath-hold conditions.

Main Results:

• Physiological noise in 3D imaging is strongly dependent on the sampling scheme used.
• Noise propagation is primarily along the slow k-space direction in 3D EPI.
• Tuned and Linear acquisition schemes exhibit better performance in terms of time-course stability and reduced noise propagation compared to Centric, Segmented, and Random schemes.

Conclusions:

• The choice of sampling scheme critically affects the level and propagation of physiological noise in fMRI.
• Tuned and Linear sampling strategies offer improved robustness against physiological noise, enhancing the reliability of fMRI data.