Background : Some individuals with severe non-fluent aphasia do not respond in a functional way to any form of communication therapy. Others show improved ability to communicate with treatment focused on alternative communication modalities such as drawing, gesturing, or using a computer. An important difference between these two patient groups may lie in their nonverbal executive function abilities. Executive functions refer to a range of cognitive abilities including goal formulation, planning, carrying out goal-directed plans, and monitoring the effects of actions. Aims : We aimed to determine whether individuals with severely restricted verbal output due to aphasia could significantly improve their functional communication skills by using an alternative communication computer program called C-Speak Aphasia. We examined several factors to determine how they related to individual patients' ability to communicate expressively using C-Speak Aphasia, including linguistic factors related to semantic processing and extent of executive system dysfunction. Methods and Procedures : Results from five patients who received at least 6 months of treatment to learn C-Speak Aphasia are presented. Communication skills on five untrained tasks were repeatedly probed throughout training to assess carryover of treatment effects. Response to treatment was then examined with respect to baseline measures of language and non-linguistic executive functioning. Outcomes and Results : Using C-Speak Aphasia three of the five participants communicated significantly more information on selected probe tasks than they did without the computer. Executive function skills were more relevant to treatment response than severity of aphasia or semantic knowledge. Subjects with more intact executive function skills responded better to treatment with this alternative communication method than subjects with relatively greater impairment in these skills. Conclusions : These results suggest that executive function impairments may underlie poor response to treatment of alternative modes of communication, and that non-linguistic measures of executive functioning should be part of every aphasia assessment when attempting to determine candidacy for certain types of treatment programmes.