We invite you to join us for our first colloquium of 2025 with Dr. Hamid Motallebzadeh.
TITLE:
Transforming Audiology with Machine Learning: Simulation-Based Inference for Personalized Diagnostics
ABSTRACT:
Machine learning is revolutionizing medicine, including audiology, by transforming diagnostic and therapeutic approaches. Effective machine learning depends on training algorithms with diverse datasets encompassing both normal and pathological conditions, as well as intersubject variabilities. However, in audiology, large datasets covering the full spectrum of middle-ear pathologies and their stages are often unavailable. Synthetic data provides a powerful solution by utilizing high-fidelity computational models of auditory systems to simulate a wide range of conditions and replicate clinical test outcomes.
Simulation-Based Inference (SBI) employs these computational models to establish associations between the parameters of auditory systems—such as material properties and anatomical geometries—and the clinical test outputs they generate, including wideband immittance measures. By training SBI with synthetic data that incorporates variability from both normal and anomalous physiological and anatomical conditions, it enables detailed, quantitative assessments of underlying abnormalities in clinical data. This capability facilitates objective, patient-specific diagnostics by uncovering subtle patterns in clinical data that conventional methods or human interpretation might overlook.
This presentation will explore the development of synthetic data generators, the training and optimization of SBI models, and the evaluation of their diagnostic performance. By addressing critical gaps in data availability and improving diagnostic precision, SBI exemplifies its transformative potential to advance personalized diagnosis and intervention planning in audiology.
BIO:
Dr. Motallebzadeh is an assistant professor in the Department of Communication Sciences and Disorders at California State University, Sacramento. He also holds an adjunct professor position in the Department of Biomedical Engineering at McGill University. Previously, he served as an instructor and investigator in Otolaryngology–Head and Neck Surgery at Harvard Medical School.
Dr. Motallebzadeh’s research focuses on hearing science, with expertise in biomechanics, diagnostics, and auditory health technologies. His work integrates computational modeling and machine learning to advance hearing assessment and intervention strategies. He is currently leading NIH-funded research to develop automated, objective diagnostic tools for middle-ear pathologies, leveraging machine learning for pattern recognition and classification. Additionally, Dr. Motallebzadeh holds a CIHR grant for his work on newborn hearing screening. He also directs a pre-clinical investigation of bone-anchored hearing implants, utilizing in-silico and in-simulacra modeling to enhance monitoring and improve device efficacy.
We look forward to seeing you on January 8th at 12:10pm! If you plan to join us virtually, please click here to RSVP for the Zoom session.