Cochlear Implant Research Lab

Borjigin's

Cochlear Implant Research Lab (CIRL)

The mission of the Cochlear Implant Research Lab (CIRL) is to improve the quality of life for individuals with cochlear implants and more broadly, those with hearing loss.

About the Lab

Cochlear implants are among the most successful neural prosthetic devices, restoring access to sound for more than a million people with severe-to-profound deafness. However, despite substantial gains in speech intelligibility in quiet settings, cochlear implant users continue to face significant challenges in noisy environments—the conditions that define most real-world listening.

Our research therefore focuses on understanding how people navigate everyday listening situations, particularly in noise, and leveraging that knowledge to advance sound coding and signal processing strategies for cochlear implants.

Agudemu Borjigin working with a patient in the Cochlear Implant Research Lab

Our Research

Bringing Back the Missing Details of Sound

Many people with cochlear implants can understand speech in quiet, but still struggle in real-world situations—like finding where a sound is coming from, enjoying music, following conversations in echoey rooms, or keeping up without feeling exhausted. Our research focuses on a key piece of sound that today’s cochlear implants largely leave out: temporal fine structure (TFS).

What is TFS?

When sound reaches the ear, it has two parts: a slow “envelope” that carries intensity or loudness changes, and fast, rapid timing (or frequency) information called temporal fine structure. TFS helps the brain perceive pitch, locate sounds in space, make sense of speech in reverberant rooms, and manage listening effort. Most current cochlear implant strategies do not encode these fast timing cues well—meaning an important layer of sound is missing.

What We Are Testing

This project is to test whether adding TFS back into cochlear implant processing can improve four areas that are important for everyday hearing:

Spatial hearing – knowing where sounds are coming from
Pitch perception & music appreciation – hearing melodies and vocal emotion
Listening in reverberation – understanding speech in echoey spaces
Listening effort – how hard the brain has to work to follow sound

CI users will complete listening tasks with and without TFS-based processing, while we measure both performance and effort using behavioral and physiological tools.

Our Goal

Our goal is to build a clearer picture of how restoring these missing timing cues can make hearing feel more natural, less tiring, and more connected to the world.

Artificial Intelligence & Hearing

Everyday listening is full of challenges—background noise, overlapping voices, and fast-changing sound environments. For people with hearing loss and cochlear implants, these challenges can make communication exhausting and frustrating.

Our lab uses artificial intelligence (AI) to teach computers how to “listen” the way humans do, so they can improve sound before it reaches the ear or a cochlear implant. We design deep-learning systems that automatically separate speech from noise, enhance important sound features, and adapt to different listening environments in real time. By combining AI with hearing science, our goal is to make speech clearer, reduce listening effort, and improve communication in the real world—whether in a busy restaurant, a classroom, or at home.

Our Goal

Our goal is to ultimately make everyday communication easier and less tiring for CI users, and make cochlear implant care more precise, personalized, and transparent.

Our Projects

Deep Neural Network (DNN)–Based Noise Reduction for Cochlear Implants

Background noise is one of the biggest barriers to understanding speech with a cochlear implant. Traditional noise-reduction methods rely on simple rules and often fail in complex, real-world environments. Our lab develops deep neural networks (DNNs) trained on large collections of real-world sounds to learn how to separate speech from noise. These AI models can recognize patterns that traditional algorithms miss and can clean up the signal before it enters the cochlear implant. This work aims to:

Improve speech clarity in noisy environments
Preserve important sound cues for natural listening
Reduce mental effort during conversation

Predicting Cochlear Implant Outcomes with Machine Learning & Big Data

Cochlear implant outcomes vary widely—some people hear very well after implantation, while others continue to struggle. Currently, clinicians have limited tools to predict who will benefit most from a cochlear implant before surgery. We are developing machine-learning models that analyze large clinical datasets to uncover patterns linking pre-operative factors (such as hearing history and test results) to post-implant outcomes. Our goal is to create a clinical decision-support tool that can: