A stratified, parallel-group, double-blind, randomized controlled trial of remotely delivered START treatment to individuals with severe-to-moderate stroke (with recruitment focused on individuals with low SES) will be conducted. Subjects and assessors will be blinded to the condition making the experiment double blind. Specifically, subjects will be told that we are exploring a new therapy that using different sounds to improve therapy. Parallel group design will ensure that subjects in the Control group are unaware that their "sounds" are softer than the START group. Trainers may become aware that a loud sound is present thus a unique Assessor will evaluate clinical performance before and after training making the study double-blind. Fifty-four subjects will undergo baseline testing in the laboratory to establish their capacity for functional and expressive speech as well as their self-reported health-related quality of life (power analysis below). Next, subjects will participate in a high-frequency, word-picture verification/ auditory-repetition treatment, 2 hr/day for 5 consecutive days focusing on expression of words of functional significance (e.g., water, fall). Subjects will either receive training with START or without (Control). Subjects will be re-tested immediately following training as well as one-month post to assess retention. Aim 1 will evaluate capacity of START to enhance SLT outcomes by assessing the % change in clinical assessment of functional and expressive speech. Our preliminary data points to a robust response [details]. Aim 2 will focus on the capacity of these changes to 1) be retained and 2) impact subject's reported quality of life. NOTE: While we are planning in-person baseline, end, and retention testing, in response to COVID, we have established remote clinical screening using peer-reviewed validated techniques for WAB and ABA-2 (see Alternative Solutions). All preliminary data collected for this proposal were collected remotely via no-contact protocols.

Inclusion Criteria:

• 18 years old
• Native English Speakers
• Capacity to provide informed consent
• Right-handed
• Corrected to normal vision
• Left hemisphere cerebral stroke at least 6 months prior to testing
• Presence of severe-to-moderate, non-fluent aphasia, i.e. Aphasia Quotients of 0-50 on the Western Aphasia Battery-Revised (PsychCorp, 2007)
• Corrected pure tone threshold (octave frequencies 250- 4000 Hz) norms for their age and gender41,42 NOTE: Audiometry data will be collected for all participants by lab personnel trained by an audiologist in a sound-attenuated booth. Based on our ongoing aphasia studies, we expect that ~30% of participants will use hearing aids; we will not exclude these individuals but rather include hearing aid use as a covariate in analyses.

Exclusion Criteria:

• Severe concurrent uncontrolled medical problems (e.g. cardiorespiratory impairment).