Premium
Sliding‐window sensitivity encoding (SENSE) calibration for reducing noise in functional MRI (fMRI)
Author(s) -
Law Christine S.,
Liu Chunlei,
Glover Gary H.
Publication year - 2008
Publication title -
magnetic resonance in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.696
H-Index - 225
eISSN - 1522-2594
pISSN - 0740-3194
DOI - 10.1002/mrm.21701
Subject(s) - sensitivity (control systems) , functional magnetic resonance imaging , noise (video) , voxel , smoothing , image resolution , computer science , communication noise , artificial intelligence , calibration , sense (electronics) , sliding window protocol , computer vision , signal to noise ratio (imaging) , nuclear magnetic resonance , window (computing) , physics , image (mathematics) , chemistry , telecommunications , linguistics , philosophy , neuroscience , electronic engineering , engineering , biology , operating system , quantum mechanics
Functional magnetic resonance imaging (fMRI) at high magnetic field with parallel imaging (PI) has become increasingly popular for high‐resolution imaging. We present a method of self‐calibrated PI‐fMRI in which sensitivity profiles are calculated using a sliding window of fully sampled multishot imaging data. We show that by updating these sensitivity profiles in a sliding fashion, thermal noise is reduced in the reconstructed image time series. This is accomplished by retaining thermal noise in the sensitivity profiles; no spatial smoothing is performed. These noisy profiles actually provide a closer match to those required for thermal noise‐free reconstruction than conventional sensitivity map generation. Our proposed technique is especially applicable for acquiring high‐spatial‐resolution images, where thermal noise exceeds physiological noise. With conventional sensitivity calculation, PI‐fMRI sensitivity is preserved only when using a voxel size large enough such that physiological noise predominates. With small voxel size, our technique reveals activation from visual stimulation where conventional sensitivity calculation techniques falter. Our technique enhances fMRI detection, especially when higher spatial resolution is desired. Magn Reson Med 60:1090–1103, 2008. © 2008 Wiley‐Liss, Inc.