Investigating the reliability of the evoked pupil response in listeners with hearing impairment during a speech-in-noise task

Pupillometry is often used as a measure of listening effort associated with a task, where an enlarged pupil size is associated with increased mental effort. However, pupillometry has so far been mostly applied using group-level analysis with normal-hearing cohorts. This study investigated the individual reliability of the evoked pupil response among listeners with hearing impairment during a speech-in-noise task. The aim was to study which internal and external factors could impact the reliability of the pupil response in this group (i.e., hearing and cognitive abilities, normalization procedure, signal-to-noise ratio, SNR, and visit number, respectively). For this, two pupil features were measured (peak and mean pupil dilation) across three different visits. The results obtained in the current study showed that with increasing number of visits, the reliability of the pupil features increases, consistent with earlier studies with normal-hearing listeners. Moreover, the changes in the pupil response were reproducible after correcting for the individual's baseline pupil response and the individual dynamic range. There was evidence of a “give-up effect” among listeners with larger hearing loss in the more demanding tasks (lowest SNR). These findings make an important contribution to the use of pupillometry in assessment of listening effort on an individual level.