HSL Seminar: Faisal Karmali, Ph.D. (9 a.m., 3/20, 33-206)
"Optimization of dynamics by the brain: Bayesian models of vestibular processing"
Faisal Karmali, Ph.D.
Harvard Medical School
Instructor in Otolaryngology
Massachusetts Eye and Ear
Co-Director, Jenks Vestibular Physiology Laboratory
Abstract: The vestibular system senses head motion using imperfect sensors with noisy outputs, and there is evidence that the brain is able to estimate head motion in a way that improves on the sensor characteristics. There is a rich history of applying mathematical models to understand this estimation process, including at the MVL/HSL. This includes Bayesian optimal models which propose that the brain adjusts the response dynamics according to the sensor dynamics and noise properties. In this talk I will discuss these Bayesian models and recent experimental findings that validate these models – including for known changes in dynamics with aging, stimulus amplitude, and peripheral damage.