Abstracts

RATIONALE:Several functional MRI (fMRI) language tasks are reported to successfully lateralize and localize language in children. This study compares the effectiveness of three of these paradigms. METHODS:12 patients under 18 years (3 left handed) were referred for fMRI evaluation usually for presurgical planning. All had fMRI language examination using three different language paradigms, namely verb generation(VG), word generation to letters(LF), and object naming (ON). fMRI studies were obtained using a GE 1.5T Signa clinical MRI system. A spiral T2*-weighted scanning protocol (TE=40ms,TR+640ms, FOV=30cm)was used.Paradigms were presented in block design. Data was analyzed using Stimulate software. Sites of activation for each patient and each paradigm were analyzed visually and independently by two reviewers. Sites of activation and degree of laterality were determined and compared. RESULTS:12 patients (6 females) were studied with all three paradigms. In one patient with a brain malformation activation sites and laterality of language could not be determined. Each task variably activated the inferior frontal gyrus, superior temporal gyrus, and the supplementary motor area. Object naming tasks were more often had bilateral activation. Hemispheric dominance for language could be determined in 10 patients using LF, 8 patients using VG, and 5 patients using ON. Combining two tasks facilitated language localization in one patient. CONCLUSIONS:VG and LF fMRI language paradigms tended to give more consistent language lateralization than ON in children.The use of ON in combination with either VG or LF facilitated determination of language lateralization.