Dissertations, Theses, and Capstone Projects

Date of Degree


Document Type


Degree Name



Speech-Language-Hearing Sciences


Mira Goral

Committee Members

Loraine K. Obler

Ofer Tchernichovski

Subject Categories

Cognitive Science | Speech and Hearing Science


Aphasia, Aphasia treatment, Nonfluent aphasia, Recursive functional learning, Recursive self-feedback, Telehealth


People with nonfluent aphasia (PWNA) require feedback from an external agent as well self-feedback to facilitate performance of language production tasks. The extent to which PWNA can use self-feedback alone to improve their task performance is unknown. In addition, it is argued that self-feedback may reinforce or minimize performance errors over time. To test whether either is the case, we introduce an original learning mechanism, recursive functional learning, which uses self-feedback loops to optimize recursively multiple subsystems (cognitive and linguistic subsystems) that PWNA engage during task performance. We used this mechanism to underpin a novel automated procedure we developed, recursive self-feedback, that exposes PWNA to audio playbacks of their own performance recursively when producing personalized scripts.

Through two experiments, we examined and demonstrated the effects of recursive self-feedback on four PWNA’s production of trained scripts, untrained scripts, as well as on connected speech elicited through untrained narrative tasks. Furthermore, we explored the role of their cognitive abilities in their script performance. Recursive self-feedback improved speaking rate and speech initiation latency on trained and untrained scripts; this generalized particularly to improvement in proportion of grammatical sentences produced during untrained narrative tasks. In addition, all participants demonstrated relatively preserved general cognitive ability, executive function and attention which may have facilitated their script production. Importantly, the beneficial effects of this recursive self-feedback suggest that some active ingredients of script-based interventions e.g., speech unison and repeated exposures to written scripts, may be optional for facilitating script production in PWNA. In conclusion, the evidence suggests that PWNA can improve their spoken language performance through recursive functional learning.