Coupled middle- and inner-ear model for model-based hearing diagnostics based on objective audiometric measurands

• Applicants: Dr.-Ing. Ernst Dalhoff; Dr. Lars Kirsten; Professor Dr.-Ing. Michael Lauxmann
• Subject Area: Biomedical Systems Technology ; Acoustics ; Medical Physics, Biomedical Technology
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 567964272

Worldwide, 400 million people suffer from hearing loss requiring intervention, on trend rising. A specific and effective development of new drugs for protection and rehabilitation of the inner ear, as well as early therapy and planning of middle-ear surgeries, requires hearing diagnostics more precise and effective than today. The overarching aim of this project is to obtain a highly precise, frequency- and place- specific objective diagnosis of the hearing signal path - from the ear canal pressure up to the excitation of the inner ear hair cells. To this end, the innovative clinically applicable objective audiometric measurement methods wideband-impedance tympanometry (WBT) and pulsed distortion-product otoacoustic emissions (DPOAE) as well as optical coherence tomography (OCT) of the tympanic membrane serve as the basis for measurement data characterizing an individual subject. For the analysis of this data, both working groups have developed highly sophisticated physics- and biology-based models of middle- and inner-ear, which will be combined numerically. The acousto-mechanical finite-element model of the working group Lauxmann comprises the external ear canal, eardrum, and ossicles, with all joints and ligaments. The hydromechanical inner-ear model of the working group Dalhoff models the cochlear amplifier, comprising the outer hair cells and their energy supply. The coupled model thus contains all diagnostically measurable input and output quantities before conversion into neural signals. Using the reflex arc of the stapedius muscle, properties of the neural pathway up to the brainstem can be assessed as well. Methods of parameter identification such as, for example, machine-learning-based Bayes inference will be used to identify clinically indicative model parameters based on individual comparisons of measurements in patients and model predictions. The methodological approach based on WBT and OCT will be validated in a first step in post-mortem preparations of the human temporal bone using invasive laser-Doppler-vibrometry measurements and μCT imaging. Subsequently, the model-based analysis of measurement data complemented with DPOAE data and using the coupled middle- and inner-ear model will be transferred to measurements in human subjects. In these subjects, the model-based prediction of middle-ear transfer functions can be validated using DPOAE level maps based on relative changes due to static pressure variations. It is expected that the ability to predict middle ear transmission with high precision in a model-based manner will have positive implications beyond the project in many use cases of hearing diagnostics, especially for newborn hearing screening, hearing diagnostics of young children, but also the detection of rare hearing diseases, such as semicircular dehiscence or even non-invasive measurement of intracranial pressure.

• DFG Programme: Research Grants