INTRODUCTIONSUBJECTS AND METHODSFigure 1.Figure 2.RESULTSFigure 3.TABLE IDISCUSSIONACKNOWLEDGMENTSREFERENCESImaging Subcortical Auditory Activity in HumansA.R. Guimaraes,1,2* J.R. Melcher,3T.M. Talavage,2,3J.R. Baker,1,2P. Ledden,1B.R. Rosen,1,2N.Y.S. Kiang,2–4B.C. Fullerton,2,3and R.M. Weisskoff1,21MGH-NMR Center, Department of Radiology, Massachusetts General Hospital,Charlestown, Massachusetts 021292Harvard-Massachusetts Institute of Technology Division of Health, Sciences and Technology,Cambridge, Massachusetts 021393Eaton-Peabody Laboratory, Massachusetts Eye and Ear Infirmary, Boston, Massachusetts4Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology,Cambridge, Massachusetts 02139r rAbstract: There is a lack of physiological data pertaining to how listening humans process auditoryinformation. Functional magnetic resonance imaging (fMRI) has provided some data for the auditorycortex in awake humans, but thereisstillapaucityofcomparabledataforsubcorticalauditoryareas wherethe early stages of processing take place, as amply demonstrated by single-unit studies in animals. It isunclear why fMRI has been unsuccessful in imaging auditory brain-stem activity, but one problem may becardiac-related, pulsatile brain-stem motion. To examine this, a method eliminating such motion (usingcardiac gating) was applied to map sound-related activity in the auditory cortices and inferior colliculi inthe brain stem. Activation in both the colliculi and cortex became more discernible when gating was used.In contrast with the cortex, the improvement in the colliculi resulted from a reduction in signal variability,rather than from an increase in percent signal change. This reduction is consistent with the hypothesis thatmotion or pulsatile flow is a major factor in brain-stem imaging. The way now seems clear to studyingactivity throughout the human auditory pathway in listening humans. Hum. Brain Mapping 6:33–41,1998.r1998 Wiley-Liss, Inc.r rINTRODUCTIONMuch of the detailed information about physiologi-cal activity in the auditory nervous system is derivedfrom animal studies using invasive techniques [Irvine,1992; Phillips et al., 1991]. Direct neurophysiologicaldata from humans are considerably less detailed [Lau-ter et al., 1995; Pantev et al., 1988; Picton et al., 1974;Romani et al., 1982], although the psychophysicalcapabilities for hearing are probably better docu-mented for humans than for any other species [Long,1994; Moore, 1989]. Recently, blood-oxygenation level-dependent functional magnetic resonance imaging(fMRI) has emerged as a noninvasive method forspatially mapping activity in the brain [Bandettini etal., 1992; Kwong et al., 1992]. A number of imagingstudies on humans have described sound-evoked cor-tical activity [Binder et al., 1994; Talavage et al., 1996;Wessinger et al., 1995], but no studies have reportedactivity for the brain-stem auditory regions wheremost of the auditory neurophysiological data in anes-thetized or restrained animals have been gathered.Other noninvasive methods such as evoked poten-tialmeasurement,magnetoencephalography,and posi-tron emission tomography each have their own limita-tions in assaying brain-stem function. Auditory-Contractgrantsponsor:NIH;Contractgrantnumbers:5PO1DA09467,PO1DC00119, T32DC00038-04.*Correspondence to: A.R. Guimaraes, Ph.D., MGH-NMR Center,Department of Radiology, Massachusetts General Hospital, 149Thirteenth St., Charlestown, MA02129.Received for publication 10 September 1997; accepted 29 September1997rHuman Brain Mapping 6:33–41(1998)rr1998 Wiley-Liss, Inc.evoked potentials can provide information aboutparticular brain-stem cell populations [Melcher andKiang, 1996]; magnetoencephalographic signals frombrain-stem structures approach thelimits of detectabil-ity [Erne´ and Hoke, 1990]; images of specific subcorti-cal auditory structures have not thus far been demon-strated with positron emission tomography. If brain-stem auditory activity could be measured with fMRI, anew way to study subcortical auditory processing inbehaving humans would be available, and humanpsychophysical data could be related to animal neuro-physiological data more readily.It is not clear why brain-stem activity (demonstrablein electrophysiological recordings [Hashimoto et al.,1981; Møller and Jannetta, 1983; Starr and Hamilton,1976]) has not been readily imaged with fMRI. Thedifficulties may be due to unfavorable anatomicalcharacteristics of the vascular system, the nature of theneuronal activity, or the fact that the brain stem moveswith each arterial pulsation, as is often seen when thebrain stem is surgically exposed [Britt and Rossi, 1982;Poncelet et al., 1992].Here we demonstrate a novel variation on standardfMRI technique that eliminates any confounding ef-fects of pulsatile brain-stem motion. In a standardfMRI paradigm, magnetic resonance (MR) images areacquired while stimuli are repeatedly turned on andoff. The MR signal-changes that are temporally corre-lated with the stimulus presentations are considered‘‘activity’’ [Bandettini et al., 1992; Kwong et al., 1992].Using such standard paradigms, we can demonstrateauditory activity routinely in the cortex, but onlyrarely in the brain stem. Two modifications weretherefore made: 1) image acquisitions were synchro-nized to a particular time in the subject’s cardiac cycle(‘‘cardiac gating’’ [Vlaardingerbroek and den Boer,1996]), and 2) a postacquisition correction was appliedto adjust for interimage variations in signal intensitycaused by fluctuations in heart rate. Here, we demon-strate that imagesof auditoryactivity inthe brain stemare improved using this approach.SUBJECTS AND METHODSData were obtained from 8 volunteers (4 male and 4female) using a1.5 T scanner(General Electric) retrofit-ted for echo-planar imaging (by Advanced NMRSystems, Inc.). The volunteers gave informed consentfor participation in this study. They were then placedsupine in the scanner and imaged using a head coil.The subject’s head was immobilized by a custom-molded bite-bar mounted on the head coil. For eachsubject, 1) Contiguous sagittal images of the wholehead were acquired, and used to select the slice forfunctional imaging. The imaging plane was chosen tocut through the transverse temporal gyri of the cortexand the inferior colliculi in the brain stem, so that allfour auditory stations could be monitored simulta-neously (Fig. 1). This plane sometimes also intersectsparts