In Canada, one individual suffers a stroke every seven minutes. Of those who survive, roughly 30% will have aphasia, a difficulty speaking with- or understanding others. Aphasia can have a significant negative impact on mood, quality of life and day to day functioning.One of the most frustrating symptoms of aphasia is a difficulty finding the right words for common objects, familiar people and places. For example, if an individual with aphasia wants to say the word "cup", they may make speech errors that are meaning-based (e.g., "glass") or sound-based (e.g., "pup"), or they may be unable to say anything at all. Fortunately, therapy for word-finding difficulty is helpful for many people, but some individuals don't improve - the reasons for this remain unclear. The goal of this project is to understand the critical factors that make therapy helpful, so that they can be harnessed to develop more successful therapy methods. The factor of interest in this project is cueing, which is commonly used in language therapy. For example, cues for the word "cup" can be: "you pour coffee in it", or "it rhymes with pup". Using computational modelling tools, we hope to identify the language processes being trained by therapy. The findings from this work can be implemented relatively quickly in clinics across Canada to help tailor therapy to the individual needs of stroke survivors, and to ultimately improve their communication and quality of life.