The effect of small rotations on R 2 * measured with echo planar imaging Elisabeth C. Caparelli, a, * Dardo Tomasi, a and Thomas Ernst b a Medical Department, Brookhaven National Laboratory, Building 490, Upton, NY 11973, USA b University of Hawaii, Queen’s University Tower, 7th floor, 1356 Lusitana Street, Honolulu, HI 96813, USA Received 3 June 2004; revised 4 November 2004; accepted 8 November 2004 Available online 5 January 2005 Several modern MRI techniques, such as functional MRI (fMRI), rely on the detection of microscopic changes in magnetic susceptibility. However, differences in magnetic susceptibility between brain tissue, bone, and air also produce local magnetic field gradients that may interfere with the contrast of interest, particularly at high field strengths. Since the magnetic field distribution depends on the orientation of the human head in the MRI scanner, head rotations can change the effective transverse relaxation rate (R 2 *) and confound fMRI studies. The size of the R 2 * changes produced by small head rotations was estimated from a brain-shaped gel-phantom at 4 T, by measuring the signal decay at 96 different echo times. Similar measurements were carried out in a human study. Rotations larger than 28 changed R 2 * more than 1.5 Hz in the phantom, and indicate that even small rotations may compromise fMRI results. D 2004 Elsevier Inc. All rights reserved. Keywords: Effective Transverse Relaxation Rate; Brain rotation; Suscepti- bility effects Introduction Functional magnetic resonance imaging (fMRI) with blood oxygenation level dependent (BOLD) contrast relies on the detection of microscopic magnetic susceptibility changes during brain activation, as a result of regional increases in blood flow, volume, and oxygenation (Ogawa et al., 1990). However, other mechanisms, such as head motion, may also induce magnetic field changes and therefore confound activation signals. Specifically, the magnetic field homogeneity is distorted as a result of differences in magnetic permeability among biological tissues and surrounding air. Therefore, shimming is used to reduce these field inhomogeneities (Li et al., 1995). However, head rotations typically reorient the magnetic permeability distribution relative to the main magnetic field and therefore induce changes in the magnetic field as well as the effective transverse relaxation rate, R 2 *. The echo planar imaging (EPI) (Mansfield, 1977) technique is commonly used in functional MRI (fMRI) studies due to its high sensitivity to microscopic permeability changes resulting from the BOLD effect. Because EPI sequences are very sensitive to in- homogeneities in the magnetic field distribution, susceptibility effects adjacent to the air/tissue interface may cause signal losses and image distortions; these artifacts are most prominent in the orbitofrontal cortex and the medial and inferior temporal lobes (Finsterbusch and Frahm, 1999; Glover and Law, 2001; Lipschutz et al., 2001; Ojemann et al., 1997; Stenger et al., 2000; Weiger et al., 2002; Wilson and Jezzard, 2003; Yang et al., 2002). Several groups have investigated and proposed solutions to reduce signal loss (Deichmann et al., 2002, 2003; Lipschutz et al., 2001; Schmitt et al., 1998) and geometrical distortions (Andersson et al., 2001; Bowtell et al., 1994; Cusack et al., 2003; Hutton et al., 2002; Jezzard and Balaban, 1995; Sutton et al., 2004; Ward et al., 2002; Weisskoff and Davis, 1992) in EPI. However, even if geometrical distortions and signal losses in EPI sequences are corrected perfectly, these techniques cannot compensate for R 2 * changes due to head rotations. Nevertheless, the effects of susceptibility changes on R 2 * have not been well explored. One early study at 1.5 T demonstrated qualitatively that head rotations may cause erroneous functional signals within the brain (Wu et al., 1997); however, a quantitative assessment of R 2 * changes due to head motion is still pending. Therefore, we measured the effect of small rotations (V108) on R 2 * in a brain- shaped gel-phantom and in vivo using EPI. Method Phantom construction The phantom consisted of a plastic skull that closely resembles a real skull in shape and size. The brain cavity of the skull was filled with an aqueous solution of ox gelatin (56 g/l of unflavored 1053-8119/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2004.11.011 * Corresponding author. Fax: +1 631 344 7671. E-mail address: caparelli@bnl.gov (E.C. Caparelli). Available online on ScienceDirect (www.sciencedirect.com). www.elsevier.com/locate/ynimg NeuroImage 24 (2005) 1164– 1169