Physiological noise reduction using volumetric functional magnetic resonance inverse imaging

Area of Science:

• Neuroimaging
• Biophysics

Background:

• Physiological noise from cardiac and respiratory fluctuations complicates BOLD-contrast fMRI.
• Traditional EPI methods struggle with whole-head coverage and aliasing of physiological noise.
• External monitoring is often required for effective noise suppression.

Purpose of the Study:

• To demonstrate the suppression of physiological noise in fMRI using magnetic resonance inverse imaging (InI) and digital filtering.
• To achieve whole-head spatial coverage without auxiliary monitoring.
• To improve the estimation of hemodynamic responses.

Main Methods:

• Utilized high-speed magnetic resonance inverse imaging (InI).
• Applied digital filtering techniques, specifically moving average (MA) filters.
• Systematically studied the effects of different MA filter window sizes.

Main Results:

• Successfully suppressed cardiac and respiratory noise without external monitoring.
• Achieved whole-head spatial coverage with reasonable resolution.
• A 2-second MA filter improved peak z-statistic values by 57%-58% compared to unfiltered InI or smoothed EPI data.

Conclusions:

• High temporal sampling rates with InI enable significant physiological noise reduction.
• Standard temporal filtering techniques are effective when combined with InI.
• This approach leads to substantial improvements in hemodynamic response estimation for fMRI.