Optimized EPI for fMRI using a slice-dependent template-based gradient compensation method to recover local susceptibility-induced signal loss

• Published in: Magma (New York, N.Y.) Vol. 23; no. 3; pp. 165 - 176
• Main Authors: Rick, Jochen, Speck, Oliver, Maier, Simon, Tüscher, Oliver, Dössel, Olaf, Hennig, Jürgen, Zaitsev, Maxim
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer-Verlag 01.06.2010; Springer Nature B.V
• Subjects: Algorithms; Amygdala - blood supply; Biomedical Engineering and Bioengineering; Brain; Brain Mapping; Calibration; Cerebral Cortex - blood supply; Compensation; Computer Appl. in Life Sciences; Echo-Planar Imaging - methods; Health Informatics; Humans; Imaging; Imaging, Three-Dimensional; Magnetic Resonance Imaging - methods; Medical imaging; Medicine; Medicine & Public Health; NMR; Nuclear magnetic resonance; Oxygen; Oxygen - blood; Oxygen Consumption - physiology; Oxygenation; Radiology; Research Article; Signal Processing, Computer-Assisted; Solid State Physics; Gradient compensation; fMRI; Susceptibility; Field map; Signal loss; BOLD sensitivity; Template-based
• ISSN: 0968-5243; 1352-8661; 1352-8661
• DOI: 10.1007/s10334-010-0215-x

Object Most functional magnetic resonance imaging (fMRI) experiments use gradient-echo echo planar imaging (GE EPI) to detect the blood oxygenation level-dependent (BOLD) effect. This technique may fail in the presence of anatomy-related susceptibility-induced field gradients in the human head. In this work, we present a novel 3D compensation method in combination with a template-based correction that can be optimized over particular regions of interest to recover susceptibility-induced signal loss without acquisition time penalty. Materials and methods Based on an evaluation of B 0 field maps of eight subjects, slice-dependent gradient compensation moments are derived for maximal BOLD sensitivity in two compromised regions: the orbitofrontal cortex and the amygdala areas. A modified EPI sequence uses these additional gradient moments in all three imaging directions. The method is compared to non-compensated, template-based and subject-specific correction gradients and also in a breath-holding experiment. Results The slice-dependent gradient compensation method significantly improves signal intensity/BOLD sensitivity by about 35/43% in the orbitofrontal cortex and by 17/30% in the amygdala areas compared to a conventional acquisition. Template-based correction and subject-specific correction perform equally well. The BOLD sensitivity in the breath hold experiment is effectively increased in compensated regions. Conclusion The new method addresses the problem of susceptibility-induced signal loss, without compromising temporal resolution. It can be used for event-related functional experiments without requiring additional subject-specific calibration or calculation time.