Hemispheric dominance for language is an important issue in functional neuroimaging, particularly driven by efforts to overcome the need for the invasive Wada test, which is all the more pressing in children. Here, we aimed at developing new paradigms for functional magnetic resonance imaging (fMRI) for assessment of language dominance that can be used in younger children and allow for performance monitoring. Two new tasks (letter and animal task) were developed and compared to two reference tasks (synonyms and verb generation task) from the literature. Overall, 23 healthy children participated (13 boys, 10 girls, 10.2 +/- 2.5 years, range 6.1-15.3 years). Analysis was done using statistical nonparametrical mapping (SnPM2) on SPM2. Both reference tasks show activation in a number of left-frontal brain regions. The letter task induced a very localized activation in the left hemisphere's Broca's region, while not activating other frontal brain regions. Lateralization (as assessed in different anatomically and functionally defined regions) was consistent and strong. The animal task failed to activate frontal brain regions and was not suitable for assessing language dominance in children in this form. We conclude that while both reference tasks are useful for determining language dominance, they coactivate a number of task-related frontal areas not directly involved in language processing. Additionally, one task is not applicable in young children while the other does not allow performance monitoring. The letter task allows to selectively activate language areas in the dominant hemisphere and is applicable even in the very lowest age group amenable to fMRI investigations while still allowing performance monitoring. It may thus be a useful tool in assessing normal and pathological language organization.