Grant: Project Grant

Hearing problems are among the most common birth defects. If hearing problems are not detected, it seriously delays language development and causes many other problems. Therefore, early screening, diagnosis and treatment are very important. However, because of temporary conditions in their ears immediately after birth, babies often fail their screening test even when they do not actually have a permanent hearing problem that requires treatment. This leads to a lot of unnecessary follow-up testing and parental anxiety. 'Tympanometry' is a clinical test that involves applying pressures to the ear at the same time as sounds. It can help with screening because it tells us about conditions in the middle ear, but its results are not well understood, especially in babies. Our Aim 1 is to study tympanometry by making special pressurized vibration measurements at multiple locations on the eardrum and in the middle ear in experimental animals. Aim 2 is to develop good mathematical models that will help us understand pressurized middle-ear function. The models will be checked against the measurements from Aim 1. Aim 3 is to study how to use artificial intelligence to interpret the results of hearing tests. The mathematical models of Aim 2 will be used to produce artificial data that can be used to 'train' a 'neural network'. This training takes a lot of time, but the neural network could then be used to quickly analyze measured data for any ear. This project will give us a better understanding of how the ear works, and of how artificial intelligence could be used for a new test that would tell us whether a baby has a hearing problem, and what kind. This could ultimately lead to a fast screening device that would be much more accurate than the devices used now.