The purpose of this study was to validate a functional magnetic resonance imaging (fMRI) paradigm to activate both anterior and posterior language areas while collecting accuracy and reaction time data on subjects' performance. The paradigm was based on alternating graphemic and lexical decision tasks. In line with the classical model of language organisation, based on lesion data, and with the results of previous neuroimaging studies, cortical activation associated with lexical decision-making was strongly lateralised to the left hemisphere and involved a network of regions in the frontal, temporal and parietal lobes. Single subject analysis demonstrated that the activation paradigm we propose is suitable for detecting language processing areas in humans for clinical studies.