Mapping an intrinsic MR property of gray matter in auditory cortexof living humans: A possible marker for primary cortex andhemispheric differencesiIrina S. Sigalovsky,a,b,*Bruce Fischl,c,d,eand Jennifer R. Melchera,b,faEaton-Peabody Laboratory, Massachusetts Eye and Ear Infirmary, Boston, 243 Charles Street, MA 02114, USAbHarvard – MIT Division of Health Sciences and Technology, Speech and Hearing Bioscience and Technology Program, Cambridge, MA 02139, USAcAthinoula A. Martinos Center, Massachusetts General Hospital, Charlestown, MA 02129, USAdCSAIL, Massachusetts Institute of Technology, Cambridge, MA 02139, USAeDepartment of Radiology, Harvard Medical School, Boston, MA 02114, USAfDepartment of Otology and Laryngology, Harvard Medical School, Boston, MA 02114, USAReceived 21 October 2005; revised 3 May 2006; accepted 8 May 2006Available online 27 June 2006Recently, magnetic resonance properties of cerebral gray matter havebeen spatially mapped – in vivo – over the cortical surface. In one ofthe first neuroscientific applications of this approach, this studyexplores what can be learned about auditory cortex in living humansby mapping longitudinal relaxation rate (R1), a property related tomyelin content. Gray matter R 1 (and thickness) showed repeatabletrends, including the following:(1) Regions of high R1 were always found overlapping poster-omedial Heschl’s gyrus. They also sometimes occurred inplanum temporale and never in other parts of the superiortemporal lobe. We hypothesize that the high R1 overlappingHeschl’s gyrus (wh ich likely indicates dense gray mattermyelination) reflects auditory koniocortex (i.e., primary cortex),a heavily myelinated area that shows comparable overlap withthe gyrus. High R1 overlapping Heschl’s gyrus was identified inevery instance suggesting that R1 may ultimately provide amarker for koniocortex in individuals. Such a marker would besignificant for auditory neuroimaging, which has no standardmeans (anatomic or physiologic) for localizing cortical areas inindividual subjects.(2) Inter-hemispheric comparisons revealed greater R 1 on the lefton Heschl’s gyrus, planum temporale, superior temporal gyrusand superior temporal sulcus. This asymmetry suggests greatergray matter myelination in left auditory cortex, which may be asubstrate for the left hemisphere’s specialized processing ofspeech, language, and rapid acoustic changes.These results indicate that in vivo R1 mapping can provide newinsights into the structure of human cortical gray matter and itsrelation to function.D 2006 Elsevier Inc. All rights reserved.Keywords: MRI; T1; Cortical thickness; Primary auditory cortex; Brod-mann areas; Brain asymmetry; Temporal processing; DyslexiaIntroductionSpatially mapping gray matter structure over the cortical surfaceof living humans has led to new insights into numerous centralnervous system processes, including learning, cognitive aspects ofaging, and the progression of neurologic or psychiatric disease(Rosas et al., 2002; Sailer et al., 2003; Sowell et al., 2003;Draganski et al., 2004; Salat et al., 2004; Thompson et al., 2004;Narr et al., 2005). Most work along these lines has involvedmapping gray matter thickness (Fischl and Dale, 2000; MacDonaldet al., 2000). However, another potentially illuminating but almostcompletely unapplied approach involves mapping the intrinsicmagnetic resonance (MR) properties of gray matter tissue (e.g., thelongitudinal and transverse relaxation rates; Fischl et al., 2004).These properties are sensitive to various aspects of the cellulararchitecture of the gray matter (e.g., myelin and iron content),including aspects that change during development, with disease, orspatially between cortical architectonic areas (Besson et al., 1989;Vymazal et al., 1999; Steen et al., 2000; Yoshiura et al., 2000;Gelman et al., 2001). One possibility is that mappings of intrinsicMR properties may lead to new understanding of the neuronaldifferences between cortical regions. Another possibility is thatthey may provide a way to resolve – in individual, livingsubjects – cortical areas (e.g., Brodmann areas) usually only1053-8119/$ - see front matter D 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.neuroimage.2006.05.023iPortions of this work were presented at the 26th Annual Meeting of theAssociation for Research in Otolaryngology (2003) and the 11th AnnualMeeting of the Organization for Human Brain Mapping (2005).* Corresponding author. Eaton-Peabody Laboratory, Massachusetts Eyeand Ear Infirmary, Boston, 243 Charles St., Boston, MA 02114, USA. Fax:+1 617 720 4408.E-mail address: [email protected] (I.S. Sigalovsky).Available online on ScienceDirect (www.sciencedirect.com).www.elsevier.com/locate/ynimgNeuroImage 32 (2006) 1524 – 1537defined in classical histology. If the latter possibility were borneout, it would mean that functional neuroimaging data could beprecisely related to cortical architectonic areas identified in thesame subjects. This would obviate the nee d for the usualapproximate methods for co-registering function and cortical areas(i.e., mapping functionally imaged brains into normalized coor-dinates, into surface-based coordinates or onto probabilistic atlases;Talairach and Tournoux, 1988; Fischl et al., 1999; Rademacher etal., 2001). The present study begins to examine what might belearned by mapping intrinsic properties of the gray matter innormal, living humans. Our investigations focus on the superiortemporal lobe, the region of the cerebrum housing auditory cortex.Our experiments mapped a particular intrinsic MR tissueproperty, R1 (i.e., the reciprocal of T1). R1, the longitudinalrelaxation rate for protons excited in the imaging process, candepend on various microscopic tissue properties (e.g., see Gore andKennan, 1999). However, a predominant factor influencing R1 istissue myelin content (see Discussion), a point illustrated by thestriking contrast between gray and white matter in R1-weighted(i.e., T1-weighted) images. Our approach involved estimating R1for each voxel in the brain, segmenting the gray matter, andaveraging R1 across the depth of the gray matter at finely spacedpoints covering the cortical surface (Fischl et al., 2004). Theresulting mappings do not show detailed variations in R1 across thegray matter laminae, but rather show the spatial distribution inoverall R1 over a 2D ‘‘sheet’’ covering the superior temporal lobe.In addition to mapping R1, we also mapped gray