Modulating Intensity and Dosage of Aphasia Scripts (MIDAS)

National Institute on Deafness and Other Communication Disorders (NIDCD), Northwestern University, University of Colorado, Boulder

The purpose of this study is to evaluate how changing conditions of speech-language treatment (namely, amount of repetition and distribution of practice schedule) affects the language outcome of participants with aphasia following a stroke. Using a computer based speech and language therapy program, participants will practice conversational scripts that are either short or long. Participants will practice for either 2 weeks (5 days a week) or for 5 weeks (2 days a week).

Determining the optimal intensity of treatment is essential to the design and implementation of any treatment program for aphasia. Yet, treatment intensity is a complex construct and information on the variables modulating it remain ambiguous and limited. Studies reported in the neuroscience and clinical literature support the need for intensive treatment to induce long-term neuroplastic changes while the cognitive psychology literature suggests that learning is best maintained with distributed schedules. A few studies have looked at dose parameters for single word naming tasks, but there is limited evidence regarding dose parameters for treatments that focus on training the production of larger units, such as sentences or even connected discourse. One approach that is frequently used clinically and has evidence for its efficacy is script training. Little is currently known regarding the optimum dose of script training (i.e., the number of repetitions over time of each sentence within the script) that is required to promote the best outcomes. This study investigates the effects of modulating stimulus variables, specifically stimulus practice distribution and stimulus repetition. We use a baseline script treatment that has experimental support regarding its efficacy, and that allows the manipulation of these variables. To ensure independence and fidelity, treatment is provided in a controlled computer environment (desktop and tablet). To avoid clinician-related variables such as expertise and personality factors that may influence treatment, sentences are modeled during treatment by an anthropomorphic agent with high visual speech intelligibility and affective expressions. With regard to "best outcomes", generalization is the ultimate goal of any treatment approach. Therefore, the primary outcome is a generalization measure of conversation. Secondary measures address short-term acquisition, longer-term maintenance, and stimulus and response generalization for assessing gain over baseline, differential effects, and interactions. A mobile-connected wireless wearable laryngeal sensor allows tracking of talk time at home and in the community as a measure of treatment effectiveness and transfer. For privacy, it does not record audio. Results and computational models of learning (generalization, short-term acquisition, and longer-term maintenance) will contribute new evidence to fill critical gaps in current scientific understanding regarding the effectiveness and clinical application of aphasia treatment approaches. More generally, findings will help to inform clinical practice and treatment of neurologic communication disorders; the virtual clinician guided intervention that the proposal develops has the potential to reduce costly clinician-client time otherwise required for long-term rehabilitation.

Stroke, Speech and Language Therapy, Speech and Language Treatment, Communication Disorders, Language Disorders

Number of correct information units produced during conversation with communication partner (speech-language pathologist) using the lines of the trained script.

Number of correct information units produced during conversation with communication partner (speech-language pathologist) about the broad topic of the trained script.

Percent accurate words produced on trained script during oral reading probes provided via the computer program.

Rate of production (words per minute) of trained script during oral reading probes provided via the computer program.

Number of correct information units produced during conversation with communication partner (speech-language pathologist) using the lines of the trained script.

Number of correct information units produced during conversation with communication partner (speech-language pathologist) about the broad topic of the trained script.

Percent accurate words produced on trained script during oral reading probes provided via the computer program.

Rate of production (words per minute) of trained script during oral reading probes provided via the computer program.

Inclusion Criteria: Men or women with diagnosis of aphasia Left-hemisphere stroke Aphasia Quotient score between 40-80 on the Western Aphasia Battery-Revised At least 6 months post stroke Completed at least eighth grade education Premorbidly literate in English Visual acuity no worse than 20/100 corrected in the better eye Auditory acuity no worse than 30 dB HL on pure tone testing, aided in better ear Not receiving speech-language therapy at the time of study Exclusion Criteria: Any other neurological condition (other than cerebral vascular disease) that could potentially affect cognition or speech, such as Parkinson's Disease, Alzheimer's Dementia, traumatic brain injury Any significant psychiatric history prior to the stroke, such as severe major depression or psychotic disorder requiring hospitalization; subjects with mood disorders who are currently stable on treatment will be considered Active substance abuse