The neural representation of a second language in bilinguals has been shown to be modulated by factors such as proficiency, age of acquisition, and frequency of use. Today, papers investigating bilingual L1 and L2 language networks number in the hundreds. Yet conclusive findings from fMRI studies as to whether the neural resources recruited by L1 and L2 can be reliably differentiated is complicated by small sample-sizes and the wide array of experimental task designs implemented in the field. One method to increase the statistical power and generalizability of findings in individual fMRI studies is to conduct a meta-analysis that averages across multiple studies. There are currently no less than seven meta-analyses devoted to various aspects of bilingual language use (Cargnelutti, Tomasino, & Fabbro, 2019; Indefrey, 2006; Liu & Cao, 2016; Luk, Green, Abutalebi, & Grady, 2012; Sebastian, Laird & Kiran, 2011; Sulpizio, Del Maschio, Fedeli, & Abutalebi, 2019; Tagarelli, Shattuck, Turkeltaub, & Ullman, 2019). However, the specific nature of many research questions in bilingual neuroimaging research and a relatively early shift in the literature away from simple language tasks to other paradigms such as language switching have resulted in a multiplicity of research designs that defy easy categorization or comparison. Thus, despite the proliferation of bilingual neuroimaging studies, the number of these that utilize the same task paradigm remain quite small and insufficient to investigate the intersection of multiple factors like AoA, proficiency, degree of exposure, and task. This talk will critique previous meta-analyses devoted to L1 and L2 processing and present the results of the first meta-analysis to group studies strictly by task with the goal to illustrate how task differences may contribute to previous results.