Modeling Electrical Conduction in the Cochlea for Implant Engineering

点击翻译按钮获取中文摘要

CONTEXT: A cochlear implant (CI) is a complex electronic device used to provide the sensation of sound to deaf individuals. Hearing outcomes are highly variable and may arise in part due to anatomical differences amongst implant recipients. There are no methods to measure detailed voltage patterns produced by CIs; however, this information would help in understanding how individual anatomy affects performance. Here, we will use computer simulation to compute voltages with the aim of understanding the mechanisms underlying patient-specific CI performance. OBJECTIVES: My long-term objective is to use engineering methods such as imaging and simulation to improve the performance of CIs. The specific aims of this proposal are to: (1) develop anatomically accurate computer models from images to calculate electrical voltages generated by a CI and quantify how they vary based on individual anatomy, (2) develop mathematical methods to estimate tissue electrical properties, (3) develop software to automatically generate high-resolution models from low-resolution diagnostic images to facilitate modeling of patients' ears, and (4) develop and apply mathematical methods to optimize CI design. SCIENTIFIC APPROACH: My team and I have collected unique images of cadaveric ears and heads using a synchrotron, a type of linear accelerator, and X-rays to depict all structures of interest in modeling. We will define the detailed geometry of computer models using these images and simulate voltage patterns using the powerful finite-element method (Aim 1). These models will be used to investigate how using simplified geometry as in past modeling studies affects computed voltages. For Aim 2, a mathematical optimization method will be developed to find the "best" electrical property values for each individual ear. Ultimately, we would like to create an easy to use, patient-specific modeling technique to optimize CI settings on a per-patient basis. Synchrotrons cannot be used on patients, so we will develop computer vision software as part of Aim 3 to generate high-fidelity models from lower resolution diagnostic scanners. Our synchrotron images will be used to train the proposed algorithms. Lastly, in Aim 4, we will develop mathematical methods and software to optimize CI design. IMPACT: Novel technical contributions of this research include (1) methods to generate high-fidelity computer models along with error estimates, (2) algorithms for electrical property estimation, (3) software to move our modeling methodologies into routine use in hospitals, and (4) methods to achieve optimal CI designs and settings. The tools developed in this work will improve hearing outcomes and support CI manufacturers, many of which have operations in Canada. Trainees working on this project will learn advanced imaging and computer simulation tools that will make them highly employable in several industries, not just the hearing implant industry.