Spectro-temporal modulation (STM) detection performance is strongly correlated with speech-in-noise abilities in listeners with hearing loss. However, the underlying mechanisms for STM detection are yet to be revealed. We predicted STM thresholds from two recent studies (Zaar et al., Hearing Research, 2023, 427:108650; 2024, 443:108949) using a physiological model for the subcortical auditory system that analyzes dynamic pitch cues associated with STM stimuli. The sensitivity of the peripheral model was adjusted to match the audiograms of the listeners, covering a range of sensorineural hearing losses. Model predictions were close to the behavioral data on the group level. At the individual level, accurate predictions were obtained for a subgroup of listeners, whereas model thresholds were substantially lower than individual listener thresholds for another subgroup. Further analyses were conducted to shed light on the reasons for the observed mismatches between measured and predicted STM detection performance. The long-term goal of this work is to better understand the factors underlying performance on diagnostic STM tests and their relationship to speech intelligibility such that they can inform our physiological-modeling strategy for predicting speech intelligibility in individual listeners. Work supported by NIH-DC010813, William Demant Foundation (23-0720), and Innovation Fund Denmark (3130-00010B).
Carney et al. (2025) studied this question.