A Within-Subject Comparison of Hearing Aid Performance in Noise Based on Verbal Response Times

Abstract

Background: Verbal response times (VRTs) are among the suggested markers for cognitive load during word recognition tasks. Measurements of VRT during hearing aid fitting can be a useful tool to obtain information about listening effort with different amplification parameters. Material and methods:A software program was developed to easily measure VRTs in speech recognition tests. The system plays 50 randomly chosen recorded words out of a set of 700 disyllables. Speech material can be presented together with pre-selected noise samples at different speech-to-noise ratios and processed with low-pass filters with selectable cut-off frequencies. The test is carried out in free field. A voice activity detector measures the time between the offset of the presented word and the onset of the repetition by the subject, which allows VRT and speech recognition scores to be quickly assessed. Tests were carried out with a group of 8 normal-hearing subjects to evaluate the effect of different filter parameters and a second group of 8 normal-hearing people to evaluate the effect of different speech-to-noise ratios on VRT. Finally, a group of 15 adult hearing-impaired subjects who used hearing aids were fitted under different conditions and the VRTs were compared between fittings. Results:Reducing the low-pass filter cutoff frequency or adding noise to the speech signal increased VRTs in normal hearing people, suggesting an inverse relationship between VRT and ease of listening. In the hearing-impaired group, VRTs with different fittings of the hearing aid showed differences that can be used as an indicator of listening effort. Conclusions:Adding a measurement of VRT to a regular word recognition test during hearing aid fitting could be useful for adjusting parameters or deciding between models or processing strategies, especially if recognition scores are high.