Hair cells are the mechanosensory cells of the inner ear whose loss causes irreversible hearing loss in mammals. POU4F3, a POU-domain transcription factor, is only expressed in hair cells in the inner ear and is essential for hair cell terminal maturation in mice. In humans, it is essential for hair cell survival as all three identified families with a POU4F3 mutation show autosomal dominant adult-onset progressive hearing loss. The pathways by which POU4F3 has this effect are unknown as little is known about its target genes (only four such genes are known: Bdnf, Ntf3, Gfi1 and Lhx3). Therefore, identification of unknown POU4F3 target genes would aid the understanding of hair cell maturation and survival. Prior to starting my doctoral work, a subtractive hybridization was carried out in an inner ear cell line (UB/OC-2) which was manipulated to either over- or under-express POU4F3 to identify unknown POU4F3 target genes. The clones produced by this screen were first matched to their corresponding gene; the presence of these genes in POU4F3-expressing UB/OC-2 cells was then confirmed by western blot and/or reverse transcriptase PCR. Where possible, expression was also confirmed in cochlear hair cells by immunofluorescence microscopy. Subsequently, the interaction of POU4F3 with these genes was characterised. Bioinformatics software was used to identify putative POU4F3 recognition elements in target gene promoters, and these sequences were then synthesised and used in a DNA binding assay (EMSA) to assess POU4F3 binding. The ability of POU4F3 to regulate putative gene promoters was also tested using luciferase assays. By these methods, one gene (Nr2f2) was shown to be a direct target of POU4F3 in vitro and two other genes were shown to be likely POU4F3 targets. The identification of these genes improves our understanding of the mechanism of POU4F3 function and further investigation of these networks may yield therapies for deafness caused by hair cell loss.

Algorithms to predict speech intelligibility are useful tools for reliable, low-cost, and repeatable assessment of speech processing technologies, thus replacing costly and time-consuming subjective tests. Of particular practical interest is their application for hearing-impaired listeners, such as cochlear implant (CI) users, where speech enhancement is a possible way to reduce the gap to the higher scores achieved by normal hearing listeners. An additional feature that such measures may possess is the non-intrusiveness, which refers to the capacity of providing a speech intelligibility prediction without the need of a priori information, such as clean reference signals. This thesis includes behavioural and numerical experiments designed to investigate how speech intelligibility in unilateral and bilateral CI users can be predicted via objective measures. It is hypothesised that accurate predictions of speech intelligibility scores can be obtained by incorporation of CI-specific features into the proposed objective measures. In Chap. II, several intrusive measures originally developed for normal hearing listeners are employed to predict the speech intelligibility of unilateral CI users. A non-intrusive measure is also proposed and evaluated against other non-intrusive and intrusive measures, and shown to achieve good performance in noisy and reverberant environments. This measure relies on temporal envelopes and uses filterbanks that were derived for CI users of the Nucleus Freedom. In Chap. III, a non-intrusive measure of the bilateral advantage to speech intelligibility in CI is proposed and evaluated against simulated and publicly available data. As a first step, a non-intrusive model is developed that can predict the binaural advantage to speech intelligibility in normal hearing listeners. Behavioural data was collected in order to validate the accuracy of the model’s predictions. Both the normal hearing and the CI versions of the model measured the contribution of two psychoacoustic phenomena: the better ear effect, and the binaural unmasking. For the CI version of the model, the computation of these two contributions were modified to incorporate known limitations of the electric hearing, such as reduced time and frequency resolution, and the poor sensitivity to interaural time differences. With these modifications, also the CI version of the proposed model yielded accurate predictions of speech intelligibility scores obtained from simulated and subjective data. In Chap. IV, one of the unilateral intrusive measures described in Chap. II is used to optimise the parameters of a filterbank with respect to speech intelligibility in a simulated CI hearing environment. Across the unilateral and bilateral/binaural investigations, it can be concluded that temporal envelope modulations, amongst other factors, play a major role in predicting speech intelligibility for CI users. This is also the case for predictions for normal hearing subjects, and a few models already exist that use alterations in the temporal envelopes domain to predict the effect on speech intelligibility. However, this study shows that not all information employed by current models is necessary for such predictions. Non-intrusive models, such as those described in Chap. I and II, can be as accurate as models that require complete information of target and interferers signals. This means that that these measures could be applied to speech enhancement, providing a quick and inexpensive way to improve current settings in modern CI.

Brn-3c (Brn3.1, POU4F3) encoding a POU domain transcription factor is a candidate gene for late onset sensorineural hearing loss, which is exhibited by a large proportion of the ageing population. To identify common sequence variants at the Brn-3c locus mutation scanning of the BrnSc cDNA, intron and 5'-flanking region was performed by PCR-SSCP analysis in 45 members of the general population. Seven polymorphic sites were identified of which five within the Bm-Sc 5'-flanking region appear common. A functional screening approach utilising in-vitro assays suggests that at least three common sequence variants in the Brn-Sc 5'-flanking region could have a functional affect: -566(GT)i7-23, -1391A>C and a complex multi-allelic poly-G polymorphism at - 3432 that exhibits multiple variations in length together with single base substitutions within the guanine repeat. The -3432poly-G polymorphism modifies the binding affinity of an OC-2 derived nuclear protein and there is convincing evidence that this is the transcription factor SP1. Use of purified human recombinant SP1 protein, in-vitro translated SP1 and in-vitro translated SP3 confirms that the -3432polyG polymorphism modulates a high affinity SP family binding site and evidence suggests that this alters the regulation of the BrnSc promoter when SP1 levels are limiting, p<0.05. Moreover, the data suggest a functional interaction between the -3432poly-G polymorphism and the -566(GT)i7.23 repeat which associate to determine the response of the Brn-3c gene to SP1. Similarly, evidence suggests that the variant allele, -1391C has a reduced affinity for an OC-2 derived nuclear protein and this is consistent with a significant decrease in basal activity of the Brn-Sc promoter, pC were genotyped for a pilot association study but allelic frequencies were not found to significantly differ between the patient and control populations examined (by %2 analysis). Further large-scale population studies are required to establish whether these common sequence variants are associated with late onset hearing loss.

Children with cochlear hearing loss are offered a range of intervention devices to manage their hearing impairment. The most common devices fitted are hearing aids, cochlear implants or a combination of both (bimodal stimulation with a cochlear implant on one ear and hearing aid on the other). The main goal of these devices is to improve listening and communication for speech and language development. However in more recent years additional focus has been given to non-speech sounds such as music. Pitch is an important aspect of music because it carries the melody; however it is represented differently by the different devices used. The impact this has on children’s musical ability is not fully understood. This thesis explores this area and aims to determine if groups of hearing impaired children who use different intervention devices have a differential impact on pitch perception, singing and general musical ability. The primary research question addressed within the thesis was, do differences exist between different groups of hearing-impaired children who use different amplification devices for general musical ability, pitch perception and singing ability?.Fifty seven children aged between 4 and 9 years old (15 Cochlear implantees, 21 hearing aid users, 8 children with bimodal stimulation and 13 normally hearing children) were assessed for pitch perception and singing while their parents completed a questionnaire on their general musical ability. Results indicated that children using purely electrical stimulation (bilateral cochlear implants) performed more poorly for pitch perception, than children using acoustic information either through bilateral hearing aids or bimodal stimulation. This result was not demonstrated for singing competency, however a reduced comfortable singing range and greater voice irregularity was observed for the cochlear implantees when singing. Normally hearing children performed better with respect to pitch perception and singing competency but did not show a significantly better score for musical enjoyment or involvement in comparison to all three hearing impaired groups. The results indicate that the bimodal configuration could provide some benefits for pitch perception for hearing-impaired children that have useable residual hearing. This doesn’t however extend to pitch production in terms of singing competency. The findings derived from this research study are important not only to build on current research literature but also to inform future clinical practice.

Adaptation of the spike rate of sensory neurones is associated with alteration in neuronal representation of a wide range of stimuli, including sound level, visual contrast, and whisker vibrissa motion. In the inferior colliculus (IC) of the auditory midbrain, adaptation may allow neurones to adjust their limited representational range to match the current range of sound levels in the environment. Two outstanding questions concern the rapidity of this adaptation in IC, and the mechanisms underlying it. I hypothesise that the descending, corticofugal system, a major pathway of the auditory system, plays a role in neuronal adaptation to sound level statistics in the IC. I recorded single unit responses in guinea pig IC to a broadband stimulus which switched every 5s between two distributions of sound level. I then deactivated auditory cortex bilaterally using cooling cryoloops. During cooling, adapted neuronal rate-vs.-sound level functions shifted rightwards, with thresholds tending to increase. This resulted in attenuation of the improvement in neuronal representation of mean sound levels associated with adaptation. In addition, the population of IC neurones adapted more slowly to the switching stimulus during cooling. The data suggest that cortex is not necessary for the generation of midbrain spike-rate adaptation, but that adaptation appears to be a property intrinsic to IC neurones and/or inherited from more peripheral sites. Nonetheless, corticofugal activity appears to be important in midbrain neural coding of sound level, altering adaptation and improving neuronal representation of sound levels around the most common levels in the stimulus.

The primary aim of this thesis is to answer the question why older adults often experience increased difficulties understanding speech in the presence of background noise, even in the absence of hearing impairment. The potential contribution of age-related declines in subcortical auditory processing, as measured by the auditory brainstem response (ABR) and the frequency following response (FFR), (spectro)temporal, and cognitive processing is explored. The data showed that older adults who fit strict criteria of normal hearing do not necessarily experience increased difficulties in the perception of speech in the presence of steady-state or amplitude-modulated noise. They only performed more poorly in the presence of two-talker babble. Older adults with slightly poorer hearing, however, did experience increased difficulties understanding speech in steady-state noise. The data furthermore showed age-related declines in subcortical auditory processing, working memory, processing speed, and sensitivity to spectro-temporal modulations. The normal-hearing older adults showed no age-related declines in behavioural measures of envelope or temporal-fine structure processing. These findings suggest that auditory neuropathy and cognitive declines associated with ageing do not necessarily lead to increased speech in noise difficulties. Normal-hearing older adults did not experience increased difficulties understanding speech in the presence of noise maskers, despite declines in subcortical auditory processing and cognition. Moreover, when the older adults did perform more poorly, this could not be explained by auditory neuropathy, cognitive declines, or reduced spectro-temporal modulation sensitivity. This thesis furthermore explores the effects of masker modulation type and sentence complexity on the fluctuating masker benefit. The results indicated that contextual information plays a greater role in dip listening than in the perception of speech in steady-state noise. Lastly, a new technique is proposed to more rapidly collect the FFR. Recording times can significantly be reduced by presenting stimuli continuously and averaging across a single cycle of the response.

The diagnosis of auditory processing disorder (APD) remains controversial. Quantifying symptoms in individuals with APD by using validated questionnaires may help better understand the disorder and inform appropriate diagnostic evaluation. Aims: This study was aimed at characterising the symptoms in APD and correlating them with the results of auditory processing (AP) tests. Methods: Phase 1: Normative data of a speech-in-babble test, to be used as part of the APD test battery, were collected for 69 normal volunteers aged 20–57 years. Phase 2: Sixty adult subjects with hearing difficulties and normal audiogram and 38 healthy age-matched controls completed three validated questionnaires (Amsterdam Inventory for Auditory Disability; Speech, Spatial and Qualities of Hearing Scale; hyperacusis questionnaire) and underwent AP tests, including dichotic digits, frequency and duration pattern, gaps-in-noise, speech-in-babble and suppression of otoacoustic emissions by contralateral noise. The subjects were categorised into the clinical APD group or clinical non- APD group depending on whether they met the criterion of two failed tests. The questionnaire scores in the three groups were compared. Phase 3: The questionnaire scores were correlated with the APD test results in 58/60 clinical subjects and 38 of the normal subjects. Results: Phase 1: Normative data for the speech-in-babble test afforded an upper cut-off mean value of 4.4 dB for both ears Phase 2: Adults with APD presented with hearing difficulties in quiet and noise; difficulties in localising, recognising and detecting sounds and hyperacusis with significantly poorer scores compared to clinical non- APD subjects and normal controls. Phase 3: Weak to moderate correlations were noted among the scores of the three questionnaires and the APD tests. Correlations were the strongest for the gaps-in-noise, speech-in-babble, dichotic digit tests with all three questionnaires. Conclusions: The three validated questionnaires may help identify adults with normal hearing who need referral for APD assessment.

The mammalian middle ear is composed of three bony ossicles, the malleus, incus, and stapes that function to conduct sound from the external to the inner ear. Normal development of middle ear structures is integral for transduction of sound, defects resulting in conductive deafness. Mice are reliant on both embryonic and postnatal developmental events to acquire hearing, and Eya1 mutant mice present with middle ear defects during both these developmental periods. Therefore, the aim of this project was to investigate the role of Eya1 in middle ear development and disease. Eya1 mutant mice on several backgrounds have previously been characterised with middle ear ossicle defects, however the role of Eya1 in regulating ossicle development has not been investigated. In this project, I characterise novel ossicle joint patterning defects of Eya1 mice of the C57BL/6 background, and show Eya1 to be expressed in the middle ear ossicles during embryonic development. I also investigate genetic regulators of joint development and show Gdf5 misexpression in the Eya1 +/- middle ear. I suggest Eya1 indirectly regulates middle ear joint patterning through a more general role in cartilage development. During postnatal development, Eya1 may further be required for the maintenance of joints. Transformation of Meckel’s cartilage results in separation of the jaw and middle ear, and is a characteristic feature of modern mammaliaforms. The Eya1 +/- mouse exhibits a delay in postnatal Meckel’s cartilage development, suggesting Eya1 as a regulator of this process. I investigate mechanisms associated with transformation of Meckel’s cartilage, and suggest Eya1 regulates this process indirectly through recruitment of TRAP positive cells. During adult stages of postnatal development, Eya1 +/- mice are predisposed to developing middle ear infections. With the use of microscopy and histological techniques, I characterise the Eya1 +/- mouse as a model of otitis media.

Hair cells are the sensory receptors of the inner ear, converting sound and accelerations into neuronal signals. Unlike mammals, birds are able to regenerate their sensory function after auditory or vestibular trauma. Sensory regeneration requires a composite programme of epithelial repair, hair cell production and functional reinnervation. The purpose of this thesis was to explore the mechanisms of epithelial repair in the avian inner ear, and examine their contribution within the broader process of sensory regeneration. An epithelial cell culture model of the chick utricle (Gallus gallus) was developed to facilitate in vitro experimentation. Initial investigations utilised laser ablation to create wounds in the sensory epithelium. Epithelial wounds were found to heal within 4 hours using a contractile acto-myosin cable in combination with lamellipodia driven cell crawling. Nuclear incorporation of BrdU was also used to assess support cell proliferation at the sites of epithelial wounds. Proliferation was significantly increased within 48 hours of laser ablation, and was spatially restricted to the areas of epithelial trauma. The increase was strongly reduced by pharmacological blockade of PI3K, but only partially so using inhibitors of the ERK and JNK MAP kinases. The response of the avian sensory epithelium to aminoglycoside ototoxicity has also been investigated. Streptomycin sulphate was used to induce extensive hair cell death in explant cultures of the chick utricle. During this process, support cells were found to remodel and eliminate the hair cell from the epithelial surface. Time-lapse microscopy of p-actin-EGFP revealed that support cells formed a cable of f-actin around the hair cell neck. The cable subsequently constricted to repair the epithelial defect, and in doing so severed and ejected the hair bundle. Support cells adjacent to a dying hair cell also extended pseudopodia basolaterally to form a distinctive calyx of f- actin around the soma. Simultaneous time-lapse recordings of p-actin-EGFP with TOTO-3 revealed that the support cell calyx was a phagocytic structure, which culminated in the engulfment of the hair cell. These results demonstrated that support cells have a proactive role in both maintenance of epithelial integrity and the removal of corpses during hair cell death.

Connexin 26 (Cx26) and Cx30 are the two predominant gap junction constituents expressed in the mammalian cochlea. Mutations in either gene cause hereditary deafness, indicating an essential role for connexins in auditory function. Gap junctions consisting of Cx26 and Cx30 have been implicated in several cochlear processes; however, the precise functions and life-cycle of connexins in the cochlea are poorly defined. Three aspects of inner ear connexin biology were investigated. Most connexins traffic to the plasma membrane (PM) via the conventional secretory pathway. Conflicting data exists for Cx26 trafficking, whereas that of Cx30 had not been previously studied. Trafficking of Cx26 and Cx30 were investigated using stably-transfected HeLa cell lines. Treatment with brefeldin-A (a Golgi-disrupting drug) did not prevent targeting of Cx30 to the PM, whereas Cx26 was strongly inhibited. These data suggest that Cx30 may traffic to the PM via a Golgi-independent pathway, which is in contrast to a Golgi-dependent pathway for Cx26. Gap junctional intercellular communication (GJIC) pathways are hypothesised to support K+ recycling in the cochlea. This study investigated the development of GJIC in the lateral wall (LW) of live rat cochlear slices. Cx26 and Cx30 immunofluorescence revealed a progressive increase of gap junction expression from postnatal day 0 (P0) to P7-P8. Dye-coupling was compartmentalised between P2-P5, but was extensive by P7. These data suggest that GJIC matures several days in advance of hearing onset and provides anatomical evidence of a putative K+-recycling pathway. Finally, Cx30–/– mice are deaf and fail to develop an endocochlear potential (EP). This study investigated the expression of proteins involved in EP generation and found that the potassium channel Kir4.1 was noticeably reduced in the stria vascularis (SV) of Cx30–/– mice. In contrast to a separate study, the SV endothelial barrier was intact. In addition, anatomical analysis was consistent with loss of Cx30 retarding maturation of SV.