Wireless Smart System of Vocalization Sensors for Convergent Evidence of Post-Discharge Communicative Participation in Everyday Life Contexts

"1- Overview
• Research team
o Transdisciplinary research-and-development team; based at the University of North Texas; comprised of collaborators representing expertise in:
 Engineering
 Electrical
 Materials
 Computational
 Acoustic processing and speaker diarization
 Communication sciences and disorders
 Acquired neurogenic language disorders (aphasia)
 Acquired neurogenic speech-motor disorders (dysarthria)
• Goal
o To develop engineered rehabilitation solutions for patients with acquired neurogenic communication disorders
o To maximize communicative functionality and rehabilitation outcomes of discharged patients in naturally situated, communicative contexts of everyday life

2- Problem Statement
• The impairments of people with communicative disorders caused by stroke or other brain injury often persist even after inpatient discharge
• Associated communicative participation restrictions in everyday life may lead to:
o negative socio-emotive outcomes for the client
o substantial financial costs to the healthcare system and to society
• Participation restrictions may be mitigated by post-discharge assessment of the client’s everyday communicative participation that is both feasible and ecologically valid
o Current post-discharge assessments: limited primarily to subjective, face-to-face, rating scales and interviews:
 costly to administer
 provide only subjective, point-in-time feedback and input to the clinician
o The proposed engineered system would improve the feasibility and validity of assessment of everyday communicative participation-in-context
 would be automated, cost-effective, and user-friendly in its implementation
 would provide objective, continuous, internally cross-validated, ecological, and accessible feedback and input to both clinician and client
 would generate converging evidence of communicative participation, when combined with traditional assessment methods
• Target populations:
o Patients with acquired neurogenic disorders of communication, following inpatient discharge, including patients with:
 aphasia—estimated U.S. prevalence: 2.46 million adults
 dysarthria—affects an estimated 26% of adults following stroke, 90% of people with Parkinson’s disease, and 46% of adults with other neuromuscular diseases (including muscular dystrophy and ALS)

3- Solution/Technology
• Rehab Tech product
o Engineered system of two, cross-validated, wireless smart vocalization sensors, time-synchronized in their output; worn by the client for extended periods in their everyday contexts of communication
 Dermally adhered, external, para-laryngeal vocalization sensor: capacitive vocalization sensor with improved environment-independent sensitivity along with on-chip read-out circuitry with low-power radio for wireless data telemetry
 A microphone-based application which detects and differentiates voicing of the client from voicing of their conversational partners (speaker diarization), which records only extracted acoustic signal features to maintain full confidentiality of conversational content
o Machine-learning-based estimation of tailored and user-accessible vocalization-quantity-and-concentration index based on a multi-factor analysis of acquired time-series metrics from the acquired signal
• Stage of development: Early
o Para-laryngeal sensor: electrical and materials engineering design and testing in process
o Acoustic data simulating naturally occurring conversation among adults provides the sampling context for assessment of reliability and validity of speaker diarization

4- Competition/Competitive Advantage
• Individual, proof-of-concept components (para-laryngeal vocalization sensors; pediatric-adult speaker-diarization) exist in some clinical-research contexts
• Competitive advantages of the targeted system include:
o cross-validated synchronization of the dual-system output signal
 indexes both amount and speaker-diarized proportion of client vocalization within natural, adult conversations
o user-accessible output indices"