Nonlinear frequency compression is a signal processing technique used to increase the audibility of high frequency speech sounds for hearing aid users with sloping, high frequency hearing loss. However, excessive compression ratios may reduce spectral contrast between sounds and negatively impact speech perception. This is of particular concern in infants and young children, who may not be able to provide feedback about frequency compression settings. This study explores use of an objective cortical auditory evoked potential that is sensitive to changes in spectral contrast, the auditory change complex (ACC), in the verification of frequency compression parameters. We recorded ACC responses in adult listeners to a spectral ripple contrast stimulus processed with a range of frequency compression ratios (1:1 to 4:1). Vowel identification, consonant identification, speech recognition in noise (QuickSIN), and behavioral ripple discrimination thresholds were also measured under identical frequency compression conditions. In Experiment 1, these measures were completed in ten adult normal hearing individuals to determine the effects of this type of signal processing in individuals with optimal hearing. In Experiment 2, these same measures were repeated in ten adults with sloping, high frequency hearing loss, which is the clinical population for whom this signal processing technique was intended. No significant relationship of compression ratio and vowel identification was found in the normal hearing listeners, though a significant negative effect of increasing frequency compression ratio was observed in the hearing impaired group. Significant effects of compression ratio on ACC amplitude, consonant identification, ripple discrimination threshold, and speech perception in noise were found for both the normal hearing and hearing impaired groups. These results indicate that the ACC response, like speech and non-speech perceptual measures, is sensitive to frequency compression ratio. Further investigation of this technique with hearing impaired individuals is needed to determine optimal stimulus and recording parameters for the ACC in both adults and children.

Les enfants sourds congénitaux ne peuvent pas acquérir le langage oral. L'acquisition du langage après réhabilitation montre des différences individuelles qui peuvent être liées à la réorganisation corticale consécutive à la surdité. Le fonctionnement auditif cortical peut être exploré par les méthodes électrophysiologiques. Notre première étude a permis de montrer une relation entre les réponses auditives temporales (onde N1c) et les capacités langagières chez des enfants avec une réhabilitation audioprothétique. Dans une seconde étude, nous avons mis en évidence une réponse spécifique à la voix humaine chez les enfants implantés cochléaires ayant un bon niveau de langage. La distribution de cette réponse est cependant différente de celle obtenue chez les enfants normo-entendants. La réduction de l'artefact dû à l'implant cochléaire a été une étape préliminaire indispensable à l'analyse des potentiels évoqués auditifs corticaux. L’identification de tels marqueurs permettra de définir des facteurs pronostics et de mieux repérer les enfants nécessitant une adaptation précoce de la rééducation. / Congenitally deaf children cannot acquire spoken language. Language development following auditory rehabilitation varies widely from child to child; this variability could be explained by cortical reorganisation following congenital deafness. Electrophysiological techniques make it possible to investigate cortical auditory processing. Our first study identified a relationship between auditory temporal responses (N1c wave) and language performance in children fitted with hearing aids for moderate congenital deafness. Our second study investigated cortical voice processing in children with a cochlear implant (CI) and with good language performance. Minimisation of the CI artifact was a preliminary step to analyse cortical auditory evoked potentials. A specific response to voice was found in children with CI, broader and more frontally distributed than this recorded in normal hearing children. The identification of such markers shoult it make possible to identify prognostic factors and to better indentify children in need of early care management.

Behavioral perceptions and cortical auditory evoked potentials (CAEPs) from native English (N=10) and Spanish speakers (N=10) were recorded for speech stimuli that constitute phonetically relevant categories of voicing. The stimuli were synthesized bilabial stop consonant-vowel syllables in a continuum ranging from/ba/to/pa/that varied in voice onset time (VOT) from 0 to 60ms. Different behavioral perceptions were evidenced by significantly different categorical phonetic boundaries between the two groups (p.05). Peak P1, N1, and P2 response latencies and P1–N1 and N1–P2 amplitudes increased significantly with increasing VOTs (p

Les populations neuronales ont la capacité de s’organiser en réseaux qui produisent spontanément différents patterns d’activité électrique coordonnée. Au travers de leur activité synchrone, les réseaux immatures combinent les informations génétiques et les influences environnementales et contrôlent plusieurs processus neuro-développementaux dont la plasticité synaptique. Pendant le développement prénatal et post-natal précoce, ces activités électriques synchrones peuvent être générées au sein des cortex sensoriels eux-mêmes, des structures sous-corticales ou être évoquées par l’activité des organes sensoriels, elle-même spontanée ou provoquée par les stimuli sensoriels. Dans des travaux antérieurs, nous avons montré qu’un pattern EEG typique du prématuré, le « Delta-brush » (DB), qui associe une onde lente et des oscillations rapides, pouvait être évoqué dans les cortex primaires sensori-moteur, visuel et auditif par les mouvements spontanés et par les stimuli sensoriels correspondants. L’objectif de la présente étude, dédiée aux réponses corticales aux stimuli auditifs (click), était de préciser les caractéristiques spatio-temporelles des DBs évoqués ainsi que leur rapport avec des potentiels évoqués auditifs corticaux. Pour cela, les enregistrements EEG ont été réalisés en haute résolution (32 électrodes) chez 30 nouveau-nés prématurés de 30 à 38 semaines d’âge gestationnel sans risque neurologique et la position des électrodes a été recalée sur des IRM 3D acquises chez d’autres prématurés représentatifs des âges étudiés. L’analyse de population a montré une augmentation significative de la puissance spectrale après le stimulus dans toutes les bandes de fréquence allant du delta à gamma et située au niveau de la partie moyenne et postérieure du lobe temporal. Ces réponses du cortex temporal avaient une prédominance droite, étaient plus amples dans le sommeil calme et diminuaient en puissance avec l’âge. Le moyennage des réponses EEG a révélé que la composante lente du DB était une onde lente négative de grande amplitude qui culminait dans les régions temporales moyenne et postérieure à environ 550 et 700 ms respectivement. L’analyse temps-fréquence a confirmé la présence d’oscillations rapides dont les oscillations gamma, à partir du pic de l’onde lente et cohabitant avec cette dernière durant environ 700 ms. Ces résultats suggèrent que le DB évoqué par les stimuli auditifs correspond en fait à la composante lente tardive du potentiel évoqué auditif cortical du prématuré et qu’il regroupe des oscillations dans toutes les bandes de fréquence, gamma y compris, fréquence identifiée ici pour la première fois en réponse à un stimulus sensoriel chez le prématuré humain. Nous avons obtenu des résultats préliminaires similaires aussi dans une autre modalité sensorielle ; l’activité oscillatoire du DB évoqué contribue donc probablement à la maturation des cartes corticales et représente un biomarqueur potentiel du fonctionnement normal des cortex sensoriels chez le prématuré. / At the early developmental stages, during the third trimester of gestation in humans and the first post-natal weeks in rodents, sensory neocortical areas reveal similar patterns of spontaneous correlated neuronal activity. In vitro and in vivo experiments indicate that these spontaneous activities are generated from neuronal networks in the cerebral cortex, in subcortical structures or in the sensory periphery (retina, limb jerks, whiskers). Spontaneous, periphery-driven and also sensory evoked activity is relayed to the developing cerebral cortex via the thalamus and the neocortical subplate, which amplifies the afferent sensory input. The patterns of sensory evoked activity were extensively studied in rodents, but in humans their spatiotemporal dynamics still remain elusive. In humans this developmental process happens during the second half of gestation: the major growing afferents from the thalamus spread within the transient subplate zone, relocate in the cortical plate, and form functional synapses with both transient and permanent neuronal populations. Characteristic immature activity patterns of “delta-brushes” (DBs) have been reported in the preterm temporal cortex following auditory stimuli. However, the spatiotemporal dynamics of these auditory-evoked DBs remain elusive. Here, we explored the electrophysiological responses evoked by click stimuli using 32-electrode EEG recordings in thirty premature infants from 30 to 38 postmenstrual weeks (PMW) of age. Electrodes position was digitalized and registered to 3D reconstructions of preterm heads and brains computed from MRI images of other age-matched groups. Population power spectrum analysis within the 2 seconds after stimulation revealed significant increase in all frequency bands from delta to gamma, located on the middle and posterior temporal regions with a right predominance and higher power increase in the quiet sleep. Time-frequency wavelet analysis also showed fast oscillations including gamma that begin at the peak of the delta waves and co-occur with it during a period of around 700 ms. Power of auditory evoked responses significantly decreased from 30 to 38 WPM in delta to alpha bands. These are the first report of gamma oscillations in preterm sensory evoked responses. Furthermore, average cortical auditory evoked potentials (CAEP) (processed with a mean reference and a 0.16 Hz high-pass filter) revealed high amplitude delta negative waves peaking successively from the middle to posterior temporal regions at around 550 and 700 ms. Altogether these results suggest that the auditory-evoked DBs in premature infants are a slow late component of the CAEP covering temporal regions and grouping fast oscillations notably gamma oscillations.

A ce jour, six millions de français sont atteints de troubles de l'audition. Face à ce problème de santé publique, des outils performants d'exploration de la fonction auditive sont indispensables. La Speech ABR (Speech Auditory Brainstem Response ou Réponse du tronc cérébral auditif à la parole) est un outil prometteur, comme marqueur électrophysiologique fin de l'encodage neuronal de la parole. Cependant, sa méthodologie reste peu développée, son origine neuronale incertaine et elle n'a jamais été enregistrée chez le malentendant porteur d'aides auditives. Le premier axe de cette thèse porte sur les générateurs neuronaux de la Speech ABR. Le développement d'une méthodologie de recueil topographique de cette réponse jusqu'alors décrite comme strictement sous-corticale, a d'abord suggéré la possibilité d'un générateur cortical. Une étude en stéréo-électroencéphalographie a ensuite confirmé l'existence d'une activité Speech ABR dans les cortex auditifs primaires bilatéraux. Ce résultat apporte un éclairage nouveau sur la représentation des sons de parole par système nerveux auditif. Le second axe concerne l'étude de la Speech ABR chez le malentendant appareillé. Après avoir développé une méthodologie de stimulation acoustique directement au travers des aides auditives, nous avons étudié la plasticité neuronale induite par le port d'aides auditives. Les résultats montrent une amélioration de l'identification des phonèmes amplifiés, liée à une représentation corticale modifiée et à un encodage fréquentiel rééquilibré. Ces toutes premières preuves de plasticité neuronales dès les 4 premiers mois d'utilisation des aides auditives ouvrent de nouveaux espoirs thérapeutiques / To date, six million French are hearing impaired. To address this public health issue, efficient tools for exploration of the hearing function are essentials. Speech ABR (Speech Auditory Brainstem Response) is a promising tool, being a fine electrophysiological marker of the neuronal encoding of speech. Though, its methodology remains underdeveloped, its neural origin is still uncertain, and it has never been recorded in hearing aid users. The first axis of this thesis focuses on the neural generators of Speech ABR. The development of a methodology for recording topographies of this response, up to now described as strictly subcortical, first suggested the possibility of a cortical generator. A stereo-electroencephalography study then confirmed the existence of Speech ABR activity in bilateral primary auditory cortices. This result sheds a new light on the representation of speech sounds within the auditory nervous system. The second axis concerns the study of Speech ABR in hearing aid users. After having developed a methodology of acoustic stimulation directly through hearing aids, we investigated neural plasticity induced by hearing aid use. Results show an improvement in the identification of amplified phonemes, linked to an altered cortical representation and a rebalanced frequency encoding. This very first evidence of neural plasticity as soon as the first four months of hearing aid use opens up new therapeutic hopes

In individuals claiming compensation for occupational noise induced hearing loss, a population with a high incidence of nonorganic hearing loss, a reliable and valid behavioural pure tone (PT) threshold is not always achievable. Recent studies have compared the accuracy of behavioural PT threshold estimation using the slow cortical auditory evoked potentials (SCAEP) and auditory steady-state responses (ASSR) but there is no consensus regarding recommended technique. A review of the literature indicated that no comparison has been completed on the use of SCAEP and a single frequency ASSR technique. A research project was therefore initiated with the aim of comparing the clinical effectiveness (accuracy) and clinical efficiency (time required) of SCAEP and ASSR for behavioural PT threshold estimation in adults exposed to occupational noise. Adult participants were divided into a group with normal hearing (behavioural PT thresholds < 20 dBHL; n = 15) and a group of participants with hearing loss (n = 16 adults), the latter of which were recruited from individuals referred for audiometric screening, as part of hearing conservation programs, and who were, therefore, exposed to occupational noise. The GSI Audera electrophysiological system was used for both SCAEP and ASSR threshold measurement at 0.5, 1, 2 and 4 kHz. Use was made of tone burst stimuli for the SCAEP (rise and fall of 10 ms with 80 ms plateau), while amplitude and frequency modulated (AM/FM) stimuli was used during ASSR testing. The system’s 40 Hz protocol was chosen for use during ASSR recording while participants slept because this led to lparticipants. ASSR thresholds could not be measured in two of the three sleeping participants in the preliminary study using an 80 Hz modulation rate due to excessive noise. The mean SCAEP difference scores (SCAEP threshold minus behavioural PT threshold) for both participant groups were -0.2+10.2, 2.8+10.1,5.8+9.7, 0.5+10.4 at 0.5, 1, 2, and 4 kHz respectively, while ASSR difference scores were 25.3+12.8, 21.7+11.3,32.3+12.2, 27.1+13.8. The SCAEP correlations with behavioural PT thresholds across frequencies (r = 0.85) were also stronger than ASSR correlations (r = 0.75). Therefore, with regard to proximity of auditory evoked potentials (AEP) to behavioural PT thresholds and consistency of this relationship, the SCAEP, rather than ASSR, is the AEP of choice. However, the SCAEP took on average 10.1 minutes longer to complete than the ASSR. Clinical effectiveness was given comparably more weight than the clinical efficiency of the AEP technique to estimate behavioural PT thresholds due to the impact on overcompensation for occupational noise induced hearing loss. As such, the study acknowledged the SCAEP as the AEP of choice for the purpose of behavioural PT thresholds in adults exposed to occupational noise. It is important to note that the conclusion reached in the current study arose from the comparison of the SCAEP with a specific ASSR technique. Accuracy of ASSR estimation of behavioural PT thresholds is strongly influenced by stimulus and recording parameters of the system used, and by the participant variables. Copyright / Dissertation (MCommunication Pathology)--University of Pretoria, 2010. / Speech-Language Pathology and Audiology / Unrestricted