Clinical Investigative Study Determination of Language Dominance: Wada Test and fMRI Compared Using a Novel Sentence Task Klemens Gutbrod, PhD, Dominik Spring, MD, Nadia Degonda, PhD, D¨ orthe Heinemann, PhD, Arto Nirkko, MD, Martinus Hauf, MD, Christoph Ozdoba, MD, Armin Schnider, MD, Gerhard Schroth, MD, Roland Wiest, MD From the Department of Neurology, Inselspital, Bern University Hospital, and University of Bern, Switzerland (KG, DH, AN); Institute of Diagnostic and Interventional Neuroradiology, Inselspital, Bern University Hospital, and University of Bern, Switzerland (DS, MH, CO, GS, RW); Department of Neuropsychology, Swiss Center for Epilepsy, Z¨ urich, Switzerland (ND); Division of Neurorehabilitation, University Hospital, Geneva, Switzerland (AS). Keywords: Language dominance, Wada test, functional magnetic res- onance imaging, review, sentence task. Acceptance: Received November 15, 2010, and in revised form April 29, 2011. Accepted for publication June 14, 2011. Correspondence: Address correspon- dence to Roland Wiest, MD, Insti- tute of Diagnostic and Interventional Neuroradiology, Bern University Hospi- tal, and University of Bern. Inselspital, CH-3010 Bern, Switzerland. E-mail: roland.wiest@insel.ch. J Neuroimaging 2012;22:266-274. DOI: 10.1111/j.1552-6569.2011.00646.x ABSTRACT BACKGROUND AND PURPOSE This study aimed to develop a new linguistic based functional magnetic resonance imaging (fMRI)-sentence decision task that reliably detects hemispheric language dominance. METHODS FMRI was performed in 13 healthy right-handed controls and 20 patients at 1.5 T prior to neurosurgery. The main components of language were assessed with different paradigms (rhyme, synonym, and sentence). In controls, activations were quantified by a volume of interest analysis. Four neuroimagers tested a visual rating score in the patients group. Interrater agreement and concordance between fMRI and Wada test were calculated. RESULTS In healthy controls, the frontal language area was activated by the sentence and synonym task in 100% and in 73% by the rhyme task. The temporal language area was activated in 100% by the sentence-, in 64% by the synonym, and in 55% by the rhyme task. In the patients group, interrater agreement was .90 for activations in the inferior frontal and .97 in the superior temporal gyrus. Correlation between the WADA test and fMRI was .86 for the sentence, and .89 for the synonym task. CONCLUSIONS The sentence task provides robust activations in putative essential language areas and can be used for visual analysis of predefined areas to facilitate interpretation of clinical fMRI. Introduction Determination of language dominance is crucial prior to epilepsy surgery and any other neurosurgery close to language cortex in order to avoid postoperative deficits. Historically, the Wada test has been the standard for determination of language dominance. 1 Although still considered as a “gold standard” the Wada test has several important limitations because of its in- vasiveness, 2,3 difficulties related to its application and interpre- tation (eg, in the case of arterial cross flow, excessive or to low sedation, emotional reactions of the patients) 4-7 or the absence of spatial resolution. Methodological drawbacks of the Wada test include the limited time window to explore language func- tions during the procedure, the lack of normal control data and rather weak test-retest reliability. 4,8,9 Compared to alternative noninvasive approaches such as functional magnetic resonance imaging (fMRI) the Wada test is time consuming, requires more staff and is less cost-effective (up to three times). 10,11 Considering all these drawbacks, the role of the Wada test as the clinical standard for determination of language dominance seems to be rather weak and has been increasingly challenged in recent years. 12-14 Therefore, the need to replace this proce- dure with less invasive and more reliable techniques has long been recognized. Due to its high availability, one of the most promising methods is fMRI. The major difference between the Wada-Test and fMRI lies in the fact that the Wada-Test is an inactivation method blocking the function of one hemisphere, thus allowing to test the func- tion of the nonanesthetized hemisphere. It therefore designates the hemisphere that is essential or nonessential for language. By contrast, fMRI is an activation method providing informa- tion about which hemisphere shows areas of increased BOLD (blood oxygenation-level dependent) responses during the lan- guage task relative to a control task or rest. It may be that some of these hemispheric responses are superfluous, that is, not es- sential for language processing. 15 Therefore, the optimal fMRI design should not only be highly sensitive but also specific in activating the language dominant hemisphere, that is, should reliably activate all essential language areas. 266 Copyright ◦ C 2011 by the American Society of Neuroimaging