Falling at the intersection of hearing sciences, cognitive sciences, and neurosciences along with data sciences, machine learning and engineering, CogHear seeks to bring researchers from across these disciplines together. Activities include online discussions and in-person workshops aimed to advance scientific inquiry and computational tools for decoding brain activity and assessing cognitive functions in order to control assistive listening devices used in everyday environments.
Over the past two decades, communication aids (e.g. hearing aids) have seen tremendous leaps due to adoption of digital technologies and advances of powerful computations that leverage basic findings from hearing sciences and digital communication (e.g. directionality, adaptation to environments). Still, these technologies remain focused on the signal that is delivered to the ear, with a sensory-centric, bottom-up view of communication aids. A great deal of auditory perception, however is top-down, heavily engaging the cognitive system in shaping what and how we perceive sounds. This feedback encompasses understanding how complex processes such as attention, memory, language, listening effort, auditory scene analysis affects and interact with hearing function. One of the challenges of moving research in these new directions is that communities exploring hearing and cognitive function are disjoint with minimal opportunities to interact. CogHear activities address this gap.
With both communities reaching a maturity level that can translate into meaningful technologies, we believe the time is ripe for a scientific meeting that pushes forth new trends in these unusual communication technologies. We designed the scope of this meeting as a hands-on workshop to jump start this interaction and pushes for more concrete interdisciplinary projects that we envision could translate into effective technologies over the next 5 to 10 years.
CogHear situates itself at the intersection of hearing sciences, cognitive sciences, and neurosciences along with data sciences, machine learning and engineering. It aims to bring together researchers across these disciplines that would normally not have a natural venue to interact. We aim to foster a spirit of collaboration that allows researchers to interact closely. With recent advances in engineering capabilities, both in terms of computing power, miniaturized deployable systems as well as the ability to tackle previously intractable problems using deep learning, the timing is ripe to consider translational impact of basic science research on communications aids. Conversely, a number of advances in basic science are now becoming possible with more sophisticated computing capability (e.g. decoding of multivoxel brain activity).
The organizers,
Mounya Elhilali (Johns Hopkins University)
Shihab Shamma (University of Maryland)
Malcolm Slaney (Stanford)