Comparison of app-based and clinically administered matrix sentence tests in cochlear implant users

This is the authors' abstract. We don't add key points for this paper.

Objective To evaluate the feasibility and reliability of a smartphone-based, self-administered matrix sentence test for assessing speech perception in cochlear implant (CI) users, and to compare its outcomes with those of the standard clinical Oldenburg Sentence Test (OLSA) in free-field conditions. Methods Nineteen adult CI users (12 experienced, 7 newly implanted) completed a standard open-set OLSA testing and a closed-set version using a research smartphone app with direct Bluetooth streaming to their hearing devices in a sound-treated room. The app presented five response alternatives per sentence and estimated speech reception thresholds (SRTs) at 60% word recognition to account for increased chance level. Test-retest differences and between-group effects were analyzed using mixed-model ANOVA and post hoc t-tests. Results All participants were able to complete the smartphone-based test independently. SRTs obtained via the app showed strong agreement with clinical measures, particularly among experienced users. Larger deviations were observed in three newly implanted participants. The ANOVA revealed a significant effect of experience level ( p = 0.02), but no effect of test method and no interaction. Usability was rated high, and the simplified five-option interface was well tolerated across age groups. Conclusion The results demonstrate that smartphone-based matrix sentence testing with direct audio streaming is a feasible and reliable method for assessing speech-in-noise perception in CI users. This approach offers potential for remote monitoring and self-assessment beyond the clinical setting.