Clinical auditory profiling and profile-based hearing-aid fitting

Abstract

In this thesis, the basis for "precision audiology" was explored. The prerequisites for implementing precision treatments are 1) that the diseases must be heterogenous, 2) that there exist multiple options for treatment and 3) that there are "markers" that associate certain characteristics of the patient to specific treatments. The sources and consequences of a sensorineural hearing loss (SNHL) are diverse and the hearing devices, especially hearing aids, have multiple configurations that can be adjusted for specific needs. The present thesis focused on the investigation of auditory biomarkers that allow the link between auditory perceptual deficits and hearing-aid settings. Data-driven auditory profiling has the potential to identify hidden patterns in the data and subpopulations of hearing-impaired people with distinct differences in terms of their perceptual hearing deficits. The first study of this thesis showed that a method for auditory profiling can provide meaningful results. The results supported the hypothesis of having four profiles along two independent dimensions of perceptual deficits resulting from different auditory distortions. However, the main limitation for drawing strong conclusions was the selection of the analyzed datasets. In the second study, a new test battery for characterizing auditory deficits was proposed and tested on 75 listeners with various degrees of hearing loss and speech discrimination. The test battery assessed the auditory processing abilities of the listeners covering aspects such as audibility, loudness perception, binaural processing abilities, speech perception, spectro-temporal modulation sensitivity and spectro-temporal resolution. The dataset was then analyzed with an iterative data-driven profiling method based on the aforementioned profiling method. This robust auditory profiling yielded four clinically relevant subgroups of listeners. Importantly, the results were consistent with previous approaches of hearing loss characterization leading to the following conclusions: 1) The listeners’ hearing deficits were characterized by two independent auditory distortions, a “speech intelligibility related distortion”, that affected listeners with audiometric thresholds above 50 dB hearing level (HL) at high-frequencies, and a “loudness perception related distortion”, exhibited by listeners with audiometric thresholds above 30 dB HL at low frequencies; and 2) The four profiles (A-B-C-D), showed similarities to the audiometric phenotypes provided by Dubno et al. (2013) suggesting that Profile B might be considered a sensory loss and Profile D a metabolic loss. Finally, a proof-of-concept study was performed, where listeners evaluated different compensation strategies using a realistic hearing-aid simulator. The results suggested that listeners belonging to different profiles might prefer different compensation strategies. Listeners with high degree of loudness-related deficits might benefit from different forms of gain prescription, whereas listeners with speech intelligibility-related deficits might benefit fromsignal-to-noise ratio improvement. Different approaches for precision audiology may be implemented in the near future, which should have implications for hearing-aid development, hearing loss characterization and the quality of service in the hearing-care clinic towards an evidence-based audiological practice.