Robust localization and lateralization of human language function: an optimized clinical functional magnetic resonance imaging protocol.

An optimized clinical functional magnetic resonance imaging (fMRI) protocol with a total scanning time of 8 min is presented that localizes Broca's and Wernicke's areas robustly and determines hemispheric dominance. Language function was visualized using two different sentence generation (SG) and word generation (WG) tasks. Block designed blood oxygenation level dependent (BOLD) fMRI was applied in 14 right-handed volunteers at 1.5 T during visual stimulation. BOLD-clusters were assessed individually for anatomical localization. Reference data are provided for the maximum correlation of the measured BOLD-signal time course to the applied reference function (r(max)), for the maximum relative signal change (dS%), cluster size and Euklidian coordinates of Broca and Wernicke activation and of the anatomical homologues in the right hemispheres. Statistical means and a lateralization index (LI) were calculated. Broca activation focussed on the inferior frontal gyrus, and Wernicke activation on the superior temporal, supramarginal or middle temporal gyri. Mean BOLD-signals for Broca ranged from 1.53% (SG) to 2.56% (WG), and for Wernicke from 1.47% (SG) to 1.80% (WG). LI indicated left language dominance. The data provided further evidence for the high anatomical variability of language areas, which underlined the relevance of an individual language localization and lateralization prior to brain surgery.

Effects of covert and overt paradigms in clinical language fMRI.

RATIONALE AND OBJECTIVES: The aim of this study was to assess the intrasubject and intersubject reproducibility of functional magnetic resonance imaging (fMRI) language paradigms on language localization and lateralization. MATERIALS AND METHODS: Fourteen healthy volunteers were enrolled prospectively and underwent language fMRI using visually triggered covert and overt sentence generation (SG) and word generation (WG) paradigms. Semiautomated analysis of all functional data was performed using Brain Voyager on an individual basis. Regions of interest for Broca's area, Wernicke's area, and their contralateral homologues were drawn. The Euclidean coordinates of the center of gravidity (x, y, and z) of the respective blood oxygenation level-dependent (BOLD) activation cluster, and the correlation of the measured hemodynamic response to the applied reference function (r), relative BOLD signal change as BOLD signal characteristics were measured in each region of interest. Regional lateralization indexes were calculated for Broca's area, Wernicke's area, and their contralateral homologues separately. Wilcoxon's signed-rank test was applied for statistical comparisons (P values < .05 were considered significant). Ten of the 14 volunteers had three repeated measurements to test intrasession reproducibility and intersession reproducibility. RESULTS: Overall activation rates for the four paradigms were 89% for covert SG, 82% for overt SG, 89% for covert WG, and 100% for overt WG. When comparing covert and overt paradigms, language localization was significantly different in 17% (Euclidean coordinates) and 19% (BOLD signal characteristics), respectively. Language lateralization was significantly different in 75%. Intrasubject and intersubject reproducibility was excellent, with 3.3% significant differences among all five parameters for language localization and 0% significant differences for language lateralization using covert paradigms. CONCLUSIONS: Covert language paradigms (SG and WG) provided highly robust and reproducible localization and lateralization of essential language centers for scans performed on the same and different days. Their overt counterparts achieved confirmatory localization but lower lateralization capabilities. Reference data for presurgical application are provided.