A feasibility study on the use of audio-based ecological momentary assessment with persons with aphasia
Hester, Yolanda, Billah, Syed Masum, Shinohara, Kristen · 2023 · Proceedings of the 25th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) · doi:10.1145/3597638.3608419
Summary
This paper presents a feasibility study investigating whether audio-based ecological momentary assessment (EMA) can be used to evaluate word-finding difficulties (anomia) in people with aphasia outside clinical settings. Aphasia, a language disorder typically resulting from stroke, affects approximately one million Americans and significantly impacts communication abilities. Traditional assessment of anomia relies on in-clinic picture-naming tasks that capture only a snapshot of a person's abilities, missing the variability that occurs across different times, contexts, and emotional states throughout daily life. The researchers developed a smartphone and smartwatch system that delivers picture-naming prompts to participants throughout their day. The system used two protocols: a single-image microinteraction EMA (μEMA) delivered 36 times daily via a Fossil Gen 6 smartwatch, and a nine-image standard EMA delivered four times daily via a Motorola G Power smartphone. Six participants with mild to moderate aphasia completed the three-week study, each responding to 108 unique target images drawn from the Bank of Standardized Stimuli (BOSS). The study examined compliance rates, audio data quality, and participant experience to determine whether this approach could realistically extend anomia assessment into everyday environments.
Key findings
Participants demonstrated strong overall compliance, with approximately 80% of prompts receiving responses across both protocols. The smartphone-based nine-image EMA achieved higher compliance (84.6%) than the smartwatch-based single-image μEMA (76.8%), likely due to the smartphone's familiarity and ease of use. However, audio data quality presented significant challenges: only 68.4% of recorded responses were intelligible, while 17.5% were missing and 11.6% were unintelligible due to background noise or unclear speech. The smartwatch protocol was particularly affected, with lower intelligibility rates attributed to the device's small microphone and participants' difficulty positioning it correctly. Technical issues were substantial — Bluetooth connectivity between the watch and phone was unreliable, and smartwatch battery life required daily or twice-daily charging. Despite these challenges, participants generally found the study manageable and were willing to continue. Importantly, the research confirmed that anomia assessment data can be collected in naturalistic settings, though the technology needs refinement. Participants with hemiparesis (one-sided weakness, common after stroke) found the smartwatch especially difficult to operate.
Relevance
This study breaks important ground in extending clinical speech-language assessments into real-world contexts for people with aphasia — a population often excluded from technology research due to communication barriers. For accessibility practitioners, it highlights critical design considerations when building technology for people with acquired cognitive-linguistic disabilities: interfaces must minimize text, use aphasia-friendly language, and account for co-occurring motor impairments like hemiparesis. The finding that smartphones outperformed smartwatches in both compliance and data quality challenges assumptions about wearable convenience. The work also demonstrates the importance of involving the target population in feasibility testing before scaling up, as issues like Bluetooth reliability and battery management created barriers that would not affect neurotypical users as severely. Future iterations should explore on-device processing to eliminate Bluetooth dependencies and investigate automated speech recognition tuned for atypical speech patterns.
Tags: aphasia · ecological momentary assessment · smartwatch · audio data collection · anomia · assistive technology · feasibility study · speech-language pathology