cortex xxx (2011) 1e11 Available online at www.sciencedirect.com Journal homepage: www.elsevier.com/locate/cortex Research report Hemispheric asymmetries of cortical volume in the human 5 brain Elkhonon Goldberg a,*, Donovan Roediger a, N. Erkut Kucukboyaci a,b, Chad Carlson a, Orrin Devinsky a, Ruben Kuzniecky a, Eric Halgren b and Thomas Thesen a a New York University School of Medicine, New York, NY, USA b Multimodal Imaging Laboratory, University of California, San Diego, CA, USA article info abstract Article history: Hemispheric asymmetry represents a cardinal feature of cerebral organization, but the Received 19 June 2011 nature of structural and functional differences between the hemispheres is far from fully Reviewed 2 September 2011 understood. Using Magnetic Resonance Imaging morphometry, we identified several Revised 27 September 2011 volumetric differences between the two hemispheres of the human brain. Heteromodal Accepted 28 October 2011 inferoparietal and lateral prefrontal cortices are more extensive in the right than left Action editor Alan Beaton hemisphere, as is visual cortex. Heteromodal mesial and orbital prefrontal and cingulate Published online xxx cortices are more extensive in the left than right hemisphere, as are somatosensory, parts of motor, and auditory cortices. Thus, heteromodal association cortices are more exten- Keywords: sively represented on the lateral aspect of the right than in the left hemisphere, and MRI morphometry modality-specific cortices are more extensively represented on the lateral aspect of the left Cortical asymmetry than in the right hemisphere. On the mesial aspect heteromodal association cortices are Hemispheric specialization more extensively represented in the left than right hemisphere. Prefrontal cortex ª 2011 Elsevier Srl. All rights reserved. Parietal cortex 1. Introduction between brain biology and function may be expressed on many levels other than that of gross morphology (cytoarchi- Hemispheric specialization is among the central features of tectonic, biochemical, etc.). Thus any attempt to infer regional functional cortical organization in humans. Recognition of the brain function from regional brain morphology, however functional differences between the hemispheres often trig- tempting, requires great caution and any assertion of a “bigger gers interest in their morphological differences and vice versa. is better” structureefunction relationship must be tempered Indeed, gross morphological differences between the by this caveat. Such concerns notwithstanding, evidence is hemispheres are particularly interesting if they can be related growing that a reasonably direct “bigger is better” relationship to functional differences. The degree to which such relation- often does exist between functional proficiency and gross ships can be drawn remains uncertain, since the relationship morphometric cortical characteristics of the underlying 5 Authors’ Note: The study was approved by the Institutional Review Board of New York University. Written informed consent was obtained from all participants involved in the study. We thank Dmitri Bougakov, Barry Cohen, Michal Harciarek, Dolores Malaspina, Ralph Nixon, and Kenneth Podell for their comments. * Corresponding author. NYU School of Medicine, 145 East 32nd Street, 5th Floor, New York, NY 10016, USA. E-mail addresses: [email protected], [email protected] (E. Goldberg). 0010-9452/$ e see front matter ª 2011 Elsevier Srl. All rights reserved. doi:10.1016/j.cortex.2011.11.002 Please cite this article in press as: Goldberg E, et al., Hemispheric asymmetries of cortical volume in the human brain, Cortex (2011), doi:10.1016/j.cortex.2011.11.002 2 cortex xxx (2011) 1e11 substrate, such as regional volume or surface area size (Goldberg and Costa, 1981). If this were to be the case, the (Blackmon et al., 2010; Draganski et al., 2004; Fleming et al., functional implications of such cortical space allocation 2010; Maguire et al., 2000; Schneider et al., 2002). differences could be intriguing and would merit further Early efforts to identify morphological hemispheric asym- examination. However, this assertion was based on old find- metries were to a large degree motivated by the desire to ings and was limited to cortical convexity; therefore its val- identify the biological bases of the asymmetric cortical idity must be re-examined with up-to-date methods which language representation. A number of morphological asym- would target both lateral and mesial aspects of the hemi- metries have been described, notably involving planum tem- spheres. Here, we report hemispheric differences in regional porale and pars opercularis, and their relationship to left human brain volume across multiple cortical regions, both hemispheric dominance for language asserted, but some of lateral and mesial, using the more recently developed Free- the particularly influential findings were reported several Surfer Magnetic Resonance Imaging (MRI) processing meth- decades ago using what methodologies were available then odology (Fischl and Dale, 2000; Fischl et al., 2004). The (Geschwind and Levitsky, 1968; Galaburda et al., 1978; LeMay particular focus of this paper is to ascertain any systematic and Culebras, 1972). Subsequent research confirmed these differences in cortical space allocation to heteromodal versus structural asymmetries (Foundas et al., 1994, 1995; Anderson modality-specific cortices in the two hemispheres. et al., 1999; Watkins et al., 2001) but demonstrated that the relationship between structural asymmetries in the planum temporale and language lateralization is not nearly as strong or 2. Methods as direct as asserted earlier, and the very existence of such a relationship has been scrutinized (Beaton, 1997). Other 2.1. Participants structural asymmetries have also been described and subse- quently confirmed, notably “Yakovlevian torque” (Yakovlev, Structural MRI data from adults (N ¼ 39) aged 19e40 ¼ e ¼ 1972; Yakovlev and Rakic, 1966; Watkins et al., 2001; Narr (Mage 27.75, standard deviation SDage 6.12; 19 females et al., 2007) characterized by the right frontal and left occip- and 20 males) were analyzed. Participants were all right- ital protrusions, whose possible relationship to any functional handed as determined by the Edinburgh Handedness Inven- asymmetries remains unclear. Regional hemispheric asym- tory (Oldfield, 1971) with scores ranging from 40 to 100. They metries both in cortical thickness (Luders et al., 2006) and were all free of neurological, psychiatric, or neuro- volume (Good et al., 2001), both in gray and white matter developmental disorders based on screening interviews. They (Penhune et al., 1996; Takao et al., 2011) have been reported. were recruited as part of a community-based normative Any morphometric comparison of the two hemispheres reference sample at NYU Comprehensive Epilepsy Center. may be complicated by individual variability, which is particularly pronounced in certain structures, e.g., anterior 2.2. Imaging data acquisition cingulate and paracingulate cortex (Fornito et al., 2004; Huster et al., 2007). Furthermore, there is a growing appreciation of Two T1-weighted volumes (TE ¼ 3.25 msec, TR ¼ 2530 msec, sex-linked differences in regional brain morphology TI ¼ 1.100 msec, flip angle ¼ 7, field of view (FOV) ¼ 256 mm, (Witelson, 1989; Habib et al., 1991; Crespo-Facorro et al., 2001), voxel size ¼ 1 Â 1 Â 1.33 mm) were obtained for each partici- including hemispheric asymmetries (Luders et al., 2009; Raz pant on a 3T Siemens Allegra scanner, acquisition parameters et al., 2004), as well as age-related hemispheric differences optimized for increased gray/white matter contrast, rigid body (Raz et al., 2004; Shaw et al., 2009). co-registered, and common space-reoriented. Images were Our understanding of the functional differences between automatically corrected for spatial distortion, registered, the two hemispheres has also been refined beyond the classic averaged to improve signal-to-noise ratio, and processed with distinction between verbal and visuo-spatial asymmetries. the FreeSurfer (4.0.2) software (http://surfer.nmr.mgh. Additional functional differences have been described, notably harvard.edu). Each T1-weighted image took 8:07 min. those linking the right hemisphere to cognitive novelty and exploratory behavior and the left hemisphere to cognitive 2.3. Imaging data processing familiarity and routinization. Since this functional asymmetry was first proposed (Goldberg and Costa, 1981; Goldberg et al., Averaged volumetric MRI images were used to model each 1994a), it has found support with various neuroimaging tech- subject’s cortical surface with an automated procedure niques, including PET (Gold et al., 1996; Shadmehr and involving white-matter segmentation, gray/white matter Holcomb, 1997), fMRI (Henson et al., 2000), and high- boundary tessellation, inflation of folded surface tessellation, frequency EEG (Kamiya et al., 2002). It has been argued that and automatic topological defect correction (Dale et al., 1999; the “novelty-routinization” functional hemispheric asymme- Fischl et al., 2001). try is fundamental and irreducible to the more commonly Automated analysis was performed on a 156 node invoked language-visuospatial asymmetry, since it is present computing cluster and took approximately 32 h per scan. Each in a wide range of mammalian species (Vallortigara, 2000; analysis was then manually inspected which took, depending Vallortigara and Rogers, 2005; Vallortigara et al., 1999). on segmentation