Logo image
Mobile augmented reality game for auditory training: Feasibility and transfer to speech-in-noise performance
Abstract   Peer reviewed

Mobile augmented reality game for auditory training: Feasibility and transfer to speech-in-noise performance

Pooseung Koh, Jeongwoo Park, Sungyoung Kim, Inyong Choi and Hyojeong Lee
The Journal of the Acoustical Society of America, Vol.158(4_Supplement), pp.A73-A73
10/01/2025
DOI: 10.1121/10.0040174

View Online

Abstract

Traditional auditory training struggles with user engagement and accessibility, reducing long-term adherence and limiting clinical impact despite proven efficacy. We developed an augmented reality (AR) mobile application combining spatial audio rendering with adaptive auditory training targeting sound localization and speech-in-noise perception through selective auditory attention training. This feasibility study enrolled eleven adults (age: 53±3 years) who completed 4 weeks of at-home AR training using our mobile application. We measured training transfer using the Korean Matrix Sentence Test (KMST) at 0 dB, −6 dB, and −9 dB fixed signal-to-noise ratios (SNR) before and after training. Statistical analysis included paired t-tests with Bonferroni correction and reliable change index calculations. Significant improvements occurred across all KMST conditions: 0 dB (3.0 ± 2.9%, p = 0.006), −6 dB (7.3 ± 4.5%, p<0.001), and −9 dB (10.5 ± 8.0%, p = 0.001). Training showed dose−response relationships, with training duration correlating with −6 dB KMST gains (r = 0.547, p = 0.082) and in-game localization performance predicting challenging speech-in-noise conditions (p = 0.054 for −6 dB, p = 0.045 for −9 dB). This feasibility study demonstrates measurable AR training transfer to speech-in-noise assessments, with seven participants (64%) achieving reliable improvement. Results support ecological auditory rehabilitation approaches and warrant larger randomized controlled trials. [Work supported by the National Research Foundation of Korea (N01250813).]

Details

Metrics

1 Record Views
Logo image