Thesis (M.S.) -- Boston University. / BACKGROUND: The continual increase in the number of bi/multilingual aphasic patients has given rise to the question of efficacy of treatment across languages. One question at the forefront of current research is the extent to which language control interacts with cross-language facilitation treatment in these patients. Theories of bilingual language processing suggest that there exists bidirectional and asymmetrical relationships between the two lexicons (e.g., Revised Hierarchical Model, Kroll et al., 2010). Such a model allows for the prediction of cross-language generalization resulting in improved facilitation of translations and semantically related translations, a finding observed in treatment studies of rehabilitation of bilingual aphasia (Edmonds & Kiran, 2006; Kiran & Roberts, 2010). Recent studies examining the nature of bilingual language processing and lexical access, however, have hypothesized that cognitive-linguistic control plays a central role in selecting language representations. The neurocognitive model of language control, proposed by Green and Abutalebi (2007), states that the interplay between cortical (e.g., pre-frontal cortex, anterior cingulate cortex) and sub cortical regions (e.g., basal ganglia) sustains the intensive cognitive demand of managing two languages. This neural circuit regulates such tasks as appropriate language selection and language switching that allows for inhibition of potential cross-linguistic competitors during language production (Luk et al., 2011). Recent studies of bilingual aphasia have begun to provide evidence for impaired language/cognitive control and interference (e.g., Goral et al; 2006; Green et al., 2010). AIMS: In order to better understand the potential for cross-linguistic generalization and interference in multilingual aphasic patients, this current case study follows a trilingual woman with aphasia through two periods of rehabilitation. Several research questions are posed: (1) Does training in the weaker language (French) result in generalization to semantically related items in the target language as well as trained and untrained items in the stronger language (English); (2) Does a second period of treatment in the stronger language (English) reveal differences in treatment gains and/or in crosslinguistic generalization patterns; and (3) What is the effect of cognitive control on cross-linguistic generalization during rehabilitation of lexical access? METHODS AND PROCEDURES: The participant was a 59-year-old trilingual woman (Amharic, L1-English, L2-French, L3) who presented with a fluent aphasia secondary to left frontal tumor resection in 2008. Post-surgery CT and MRI revealed a left frontal infarct over the pre-central gyrus with extension into the basal ganglia. A detailed language use questionnaire was used to obtain information regarding the patient’s language use for each of the three languages. A single subject case study design was implemented following procedures previously developed in Edmonds & Kiran (2006) and Kiran & Roberts (2009). Following an assessment of the patient’s current language and cognitive abilities, the patient completed a 10-week treatment period in French, followed by a 10- week treatment period in English. RESULTS: Results demonstrated overall improvement on trained items in the target language across treatment periods in both languages. Within-language generalization to semantically related items and cross-linguistic generalization to translations of trained and semantically related items were not observed. In addition, error patterns revealed a considerable increase of interference of the treatment language into the non-treatment language on trained items relative to the respective treatment phase. Although the patient did show learning of new items (as evidenced by an increase in conceptual scores), as treatment progressed in one language, the patient’s ability to inhibit this language during non-treatment language probes decreased substantially. In addition, a non-linguistic flanker task targeting interference suppression demonstrated impaired non-linguistic cognitive control. Evidence from this case study suggests that facilitation may sometimes be overridden by language interference and provides support for the model of neurocognitive language control.