Master's Thesis

Development of a voice user interface for Amazon Alexa in speech therapy for patients with aphasia and dysarthria


Rusbeh Nagafi

Processing Period:

01.04.2018 - 01.10.2018


In this thesis two therapeutic training programs for the Amazon Echo are to be developed. A repeating task is to be developed for patients with dysarthria. The patient is asked to repeat phrases back to Alexa. For aphasiacs, speech comprehension and word finding will be trained in the form of a quiz.

In the repeating task, the sentences to be repeated will be designed together with the speech therapist and tailored to the respective patient. This allows the patient to specifically train the sounds (e.g. fricatives or labials) that are necessary. Patients should be given a questionnaire on a regular basis in which they evaluate the progress and motivation of the training and the functionality of the system (speech recognition, VUI, functionality). For this project an open source template with the English Alexa voice will be translated into German and further developed.

For people with aphasia, a concept is to be created and implemented in the context of the work, which aims at the integration of playful elements to increase motivation in therapy. Special emphasis is placed on word finding and speech comprehension.  Here the system reads sentences to the patient, which the patient completes linguistically correctly. The system also sends image cards to the patient's smartphone or tablet (in the Alexa app) and prompts the patient to create the appropriate set for each image card. Here too, a questionnaire is to be handed out at regular intervals. Both programs are supervised/evaluated with the expertise of speech therapists.