Anomalous Transverse Relaxation in 'H Spectroscopy in Human Brain at 4 Tesla Stefan Posse, Charles Andre Cuenod, Robert Risinger, Denis Le Bihan, Robert S. Balaban Longitudinal (T,) and apparent transverse relaxation times (TJ of choline-containing compounds (Cho), creatinelphospho- creatine (Cr/PCr), and N-acetyl aspartate (NAA) were mea- sured in vivo in human brain at 4 Tesla. Measurements were performed using a water suppressed stimulated echo pulse sequence with complete outside volume presaturation to im- prove volume localization at short echo times. T,-values of Cho (1.2 fi 0.1 s), Cr (1.6 f 0.3 s), and NAA (1.6 f 0.2 s) at 4 Tesla in occipital brain were only slightly larger than those reported in the literature at 1.5 Tesla. Thus, TR will not ad- versely affect the expected enhancementof signal-to-noise at 4 Tesla. Surprisingly, apparent T,-values of Cho (142 i 34 ms), Cr (140 -c 13 ms), and NAA (185 f 24 ms) at 4 Tesla were significantly smaller than those at 1.5 Tesla and further de- creased when increasing the mixing interval TM. Potential contributing factors, such as diffusion in local susceptibility related gradients, dipolar relaxation due to intracellular para- magnetic substances and motion effects are discussed. The results suggest that short echo time spectroscopy is advan- tageous to maintain signal to noise at 4 Tesla. Key words: nuclear magnetic resonancespectroscopy;relax- ation; diffusion; proton spectroscopy. INTRODUCTION Longitudinal (T,) and transverse ( T2) relaxation times are key parameters in localized proton ('H) spectroscopy, because they determine spectral sensitivity and play an important role in understanding molecular organization in vivo. They directly effect the design of spectroscopy pulse sequences and the selection of pulse sequence parameters. Differences among 'H-relaxation times of different metabolite resonances, water and mobile fatty acids can be used to enhance spectral localization and to extract spectral information of interest. For example, brain spectroscopy at clinical field strengths is frequently performed at long echo times to facilitate localization by attenuating overwhelming resonances from water and MRM 33:246-252 (1995) From the Diagnostic Radiology Department, Warren Grant Magnuson Clin- ical Center (S.P., D.L.B.), Laboratory of Diagnostic Radiology Research (C.A.C.), Section on Clinical Pharmacology. National Institute of Mental Health (R.R.), Laboratory of Cardiac Energetics, National Heart, Lung and Blood Institute (R.S.B.), The National Institutes of Health, Bethesda, Mary- land. Stefan Posse, Ph.D., Institute of Medicine, Research Center Julich GmbH, P.O. Box 1913, D-52425 Julich, Germany. Received May 2,1994; revised August 15, 1994; accepted October 4.1994. Copyright 0 1995 by Williams 8. Wilkins All rights of reproduction in any form reserved. 0740-3194195 $3.00 extracranial fat which exhibit much shorter T,-relaxation times than the dominant singlet resonances from choline (CHO),creatine (CR), and N-acetyl-aspartate (NAA). Me- tabolite resonances with short T,-relaxation times and J-coupling are attenuated as well at long echo times which facilitates spectral interpretation. However, pre- liminary experiments in human and animal brain suggest that T,-values of metabolite signals are species depen- dent (1-6) and may change with disease (7-9). This will confound spectral quantitation at long echo times unless T,-values are precisely measured. Recently, increased magnetic field strengths of up to 4 Tesla became available for human studies. Preliminary reports indicate that considerable gains in spectral reso- lution and sensitivity as compared with clinical field strengths can be achieved (10-15). However, T,-values are expected to increase with field strengths thus reduc- ing the signal-to-noise advantage gained by the higher field strength (16). From theoretical considerations T,- values are expected to be independent of field strength (16). However, preliminary results from our laboratory (17), which have been confirmed by Hetherington and coworkers (13) indicate that the apparent T,-values of metabolite resonances in human brain decrease with in- creasing field strength. In this paper we present a more complete account of our measurements of metabolite Tl- values and apparent T,-values in normal human brain at 1.5 and 4 Tesla. MATERIALS AND METHODS Relaxation measurements on normal volunteers (n = 16) were performed on an experimental 4 Tesla whole body scanner (General Electric Medical Systems, Milwaukee, WI). One study was performed at a clinical 1.5 Tesla whole body scanner (General Electric Medical Systems, Milwaukee, WI) to confirm methodology and literature values at that field strength. Both scanners were equipped with 10 mT/m actively shielded gradient coils. Informed consent was obtained from all volunteers prior to the measurements and FDA guidelines for RF power deposition were observed. On both scanners the same stimulated echo method with complete outer volume suppression (Fig. 1) was used to obtain single volume spectra or spectroscopic images at echo times as short as 7 ms. Details of the pulse sequence have been described earlier (18). Slice selective 4-ms sinc pulses were used for the stimulated echo part of the pulse sequence and 6-ms sinc pulses were used for outer volume suppression. 246