Hardy, Alexander
(2019)
Towards an improved auditory brainstem response measure of hidden hearing loss in humans.
PhD thesis, University of Nottingham.
Abstract
Hearing loss is usually diagnosed through pure-tone threshold audiometry. However, some listeners who have normal hearing thresholds complain about difficulty communicating in environments where there are competing sounds and noise. Animal research has shown that this problem may arise from a loss of synapses between inner hair cells and auditory nerve fibres. This is known as synaptopathy and is thought to mainly affect medium to high-threshold auditory nerve fibres, which are responsible for encoding moderate to loud sounds, but do not respond to quiet sounds. If true, this would explain why hearing thresholds are not affected. This ‘hidden’ hearing loss may explain some of the difficulties that some listeners experience whilst having clinically normal audiograms.
Synaptopathy has been studied in humans using the auditory brainstem response (ABR), but so far, the results have been varied. One possible reason could be that auditory brainstem responses contain a high degree of inter-individual variability resulting from unrelated factors such as gender, head size, or cochlear length. These confounds may be alleviated by using relative rather than absolute ABR measures. This thesis investigated two ABR methodologies, based on relative measures, aimed at optimizing the response for detection of hidden hearing loss.
The first is the wave I to V ratio, which is well established in the literature. It compares the amplitude of early (wave I) and late (wave V) components of the ABR. If synaptopathy has occurred, wave I is expected to be decreased but wave V is expected to be unchanged, due to compensatory changes in neuronal gain (hyperexcitability). This should, leading to a decrease in the wave I/V ratio. However, wave I is reportedly mainly elicited from high frequency regions of the cochlea, while wave V is thought to originate from a broader range of cochlear frequencies. This means the wave I/V ratio may be frequency-dependent, and so, using it in order to measure synaptopathy requires careful matching of audiometric thresholds which would be impossible in the context of clinical assessment. This means it needs to be measured in a frequency-specific manner in order to be clinically useful.
In the first experiment, we aimed to optimize the methodology for recording the wave I to V ratio frequency-specifically. First, we compared stimuli with different temporal properties to improve neuronal synchronization across different frequency regions, which would lead to increased ABR waves. Next, we compared stimuli with different spectral compositions, with the aim of increasing contributions to the ABR arising from low-frequency cochlear regions. Finally, we compared mastoid electrodes to specialist tiptrode electrodes. Tiptrodes are designed to enhance the early ABR waves, particularly, wave I, by moving the electrode inside the ear canal and thus closer to the generator of this wave. The results showed that both waves I and V are enhanced by using stimuli that improve temporal synchronization across frequencies, but that this advantage is not maintained in the frequency-specific measurements. Both the spectral manipulations and the use of tiptrodes were found to improve wave I, but, if anything, have a detrimental effect on wave V. In a second experiment, we measured the frequency-specific wave I to V ratio, developed in the first experiment, in two groups of subjects, one with high levels of previous noise exposure and one with low levels of exposure. The hypothesis was that high levels of noise exposure may be associated with hidden hearing loss. This should therefore lead to a reduction in the ABR wave I. In contrast, the wave V amplitude is expected to be unchanged, as central compensatory mechanisms leading to neuronal hyperexcitability are thought to restore activity at higher processing levels. Therefore, if hidden hearing loss was present in our noise-exposed subjects, we would expect to see a reduced wave I to V ratio compared to the non-exposed subjects. Contrary to this expectation, we found no significant change in wave I, but a significant decrease in the amplitude of wave V, paired with an increase in the wave V latency.
The second approach used a novel method, based on the phenomenon of adaptation, which is well-studied in ABRs. Previous research suggests that auditory nerve fibres that encode louder sounds have different adaptation properties to fibres that encode fainter sounds, suggesting that synaptopathy should change the adaptational properties of ABRs.
We first performed an experiment in non-exposed subjects to validate a measure of adaptation obtained using maximum length sequences, which allow the recording of well-defined ABRs even at very short inter-stimulus intervals. Then, we compared this measure between noise-exposed and non-exposed subjects. Our results showed a difference between the two groups in the size of wave I, with the non-exposed subjects showing a larger wave I amplitude than the exposed subjects. There were no clear differences in adaptation properties that would have indicated a change in the distribution of auditory nerve fibres encoding low and high sound levels.
This thesis suggests that the most promising approaches for developing an ABR-based measure of hidden hearing loss in humans are (a) frequency-specific ABR recording and (b) use of different stimulus levels. Whilst our results, along with results from another recent study, conflict with the idea that early and late ABR waves are affected differently by hidden hearing loss, they offer promising new approaches for dealing with the high degree of unrelated inter-individual variability present in ABR recordings, and thus present an important first step towards developing an objective diagnostic hidden hearing loss test in humans.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
![[img]](/style/images/fileicons/application_pdf.png)