Authors: Kavya Sreekumar¹, Prashanth Prabhu¹
¹All India Institute of Speech and Hearing (AIISH), Mysuru
Background: The Audible Contrast Threshold (ACT) is a language-independent clinical tool used to predict aided speech-in-noise performance by measuring spectro-temporal modulation sensitivity. While it serves as a valuable suprathreshold index of auditory processing, its test–retest reliability—crucial for longitudinal monitoring and clinical research—remains under-investigated.
Method: Thirty normal-hearing adults (18–25 years) underwent five ACT measurements across two sessions, spaced one week apart. The protocol included three intra-session and two inter-session recordings. Reliability was assessed using a two-way mixed-effects Intraclass Correlation Coefficient (ICC), Cronbach’s alpha for internal consistency, and Bland-Altman analysis to identify systematic bias.
Results: Analysis revealed moderate single-measure reliability (ICC = 0.612) and good-to-excellent average-measure reliability (ICC = 0.887). Cronbach’s alpha indicated excellent internal consistency (α = 0.903). Bland-Altman analysis showed a negligible inter-session bias (−0.20 dB), confirming clinical stability over time, though intra-session results suggested a mild practice effect with slightly wider limits of agreement.
Conclusion: ACT demonstrates robust test–retest reliability and excellent internal consistency, particularly when thresholds are averaged across multiple recordings. The near-zero inter-session bias validates ACT as a stable, reliable measure for clinical use. To ensure optimal stability in practice, clinicians are encouraged to average repeated recordings. Further research is warranted to confirm these findings in older populations and those with sensorineural hearing loss.
