Dr. Frank Guenther Awarded NIDCD Grant to Study Speech Learning Through Deep Brain Stimulation

Frank Guenther, PhD, MS, professor of Speech, Language & Hearing Sciences and Biomedical Engineering at BU, has received a research grant funded by the National Institute on Deafness and Other Communication Disorders (NIDCD). The prestigious grant, offered by the National Institutes of Health (NIH), supports independent, health-related research and development for a five-year period. Dr. Guenther’s research, which began in September 2025, aims to develop an understanding of how the basal ganglia and cerebellum each contribute to speech sequence learning and production.

Deep brain stimulation (DBS) technology, serving as an essential component of the research, will help researchers examine how these brain circuits contribute to the learning and coordination of speech. The process of speech production requires the rapid coordination of many muscles and neural systems, yet an understanding of the neurological mechanisms underlying speech still has a long way to go.

“People are able to effortlessly move their tongues and lips and jaw in a very coordinated fashion to make sounds, and we don’t think twice about it.”

Using neural recordings, behavioral experiments, and computational modeling, Dr. Guenther and his team hope to advance the understanding of how regions of the brain interact to produce fluent speech. In examining the relationship between the basal ganglia, cerebellum, and speech learning, the research aims to clarify how these regions of the brain support the fluid, rapid production of speech.

Understanding the Neural Control of Speech

Speech, as Dr. Guenther explained, is one of the most complex motor behaviors performed by the body. Production of even a single word calls for rapid and precise coordination of the tongue, lips, jaw, respiratory muscles, and the sensory information communicated via the brain. The speech itself relies on the speakers awareness of how the speech sounds and feels, a process that is largely automatic for most. Nonetheless, underlying this deceptively simple process is a network of brain regions responsible for moving to produce speech, learning how to do so, as well as continuously adjusting the production as it occurs.

There are two main subcortical structures central to the research: the basal ganglia and the cerebellum. While they both are already understood as associated with the action and fine-tuning of movement and speech, precisely how they each serve speech learning and production remains an active area of scientific study and debate. Dr. Guenther differentiated between the two by crediting the basal ganglia with the rapid learning of new speech sequences, such as unfamiliar words, new combinations of speech sounds. On the other hand, the cerebellum, he credits with fine-tuning speech movements for smoothness, accuracy, and coordination.

Dr. Guenther’s interest in these questions developed from a long-standing and committed curiosity concerning how the brain controls movement. Trained as an engineer, before transitioning to speech and the vast neural processes that control it, his research included studying the neural process behind arm movement. His background in engineering led him to a highly specialized career in computational neuroscience and neural modeling. Dr. Guenther couldn’t credit his movement into the field to a particular event, describing the transition into a career devoted to speech neuroscience as gradual.

“It was more of a 10-year process, I guess, of moving solidly into speech research rather than a single event.”

That gradual curiosity has developed into a 30+ year career in researching brain mechanisms underlying speech, speech learning, and speech control. The current NIDCD-funded project builds upon that work, serving as an opportunity to address some of his longstanding questions. In executing new experimental methods with a greater access to direct measurements of brain activity, this grant is only the next stage in Guenther’s larger research trajectory.

Studying Speech Through Patients with DBS

A central feature of Dr. Geunther’s NIDCD-funded research are Deep Brain Stimulation (DBS) implants. These implants are primarily administered to treat movement disorders like Parkinson’s disease and Essential tremor, and the research examines two such groups of participants who have already received DBS implants.

The researchers operate out of, and in collaboration with, Massachusetts General Hospital. Working with neurosurgeon Mark Richardson, MD, PhD, as well as other clinicians based in the hospital, Dr. Guenther and his team are able to study how speech changes as stimulation to these circuits is manipulated. Patients who choose to opt in to the research generally complete approximately two days of testing, where researchers test their ability to learn and produce unfamiliar speech patterns with their DBS stimulators both on and off. Andrew Meier, PhD, a research scientist in the Department of Speech, Language, & Hearing Sciences and a member of the research team, noted that these sessions are longer than conventional speech experiments. The purpose, he says, is so patients are allotted with the time to learn the completely unfamiliar and invented syllable sequences before being tested.

“We’re looking at learning of speech, so we need basically the first half to teach people a novel speech sequence, and the second half to test them on various aspects of what they learned.”

During the experiment, patients wear headphones and speak into a microphone and repeat unfamiliar speech patterns displayed before them on a computer. The first day of testing places an emphasis on learning sequences of syllables. Rather than using words familiar to the participants, the team generates multiple unfamiliar syllables organized into longer sequences, with a focus on repetition until each sequence becomes learned. The second day focuses on learning at a smaller speech level: individual unfamiliar syllables and difficult consonant combinations, some absent from ordinary English. This, Meier says, is to simulate a challenge created by sounds participants did not learn in early English-language development.

Dr. Guenther noted a key aspect of the implant is that it can be turned on and off, even via the patient’s phone. While this may not be utilized regularly by the patients themselves, the device can serve a fascinating purpose in an experimental setting. As the degree of stimulation is adjusted by the researchers, participants will adjust the speech in real time. Regulating the stimulation these devices provide throughout active testing allows the researchers to measure the resulting changes in learning, accuracy, and speech production. Changes in participant performance under different stimulation changes gives the researchers the opportunity to distinguish functions of the basal ganglia from those of the cerebellar circuit. Beyond observing these behavioral changes, the DBS implants also allow the team to record neural activity directly from the basal ganglia or thalamus while participants complete the activity.

Building on the DIVA Model

The current research project builds on a computational model developed by Dr. Guenther more than 30 years ago. Known as the DIVA (Directions Into Velocities of Articulators) Model, was developed to to explain the coordination among different brain regions during the processes of speech learning and production. Represented as a neural network, the model simulates connections between neural populations to describe how the brain translates intended speech sounds into coordinated speech movements.

“It’s meant to provide a relatively simple description of how the brain regions work together to control speech.”

When Dr. Guenther first developed the DIVA Model in the 90s, there was limited technology available to directly test many neurological predictions. Functional magnetic resonance imaging (fMRI), which became a tool in his later research, only first became available around the completion of his doctoral work. Subsequent advances in neuroimaging, neural recording, and measuring human brain activity allowed Dr. Guenther and his collaborators to test the model, and it’s been evolving ever since. Alongside technological advances, the model has undergone continuous growth, like matching parts of the model with specific brain structures.

The NIDCD-funded R01 research grant also serves as an opportunity to further refine the model, with the addition of behavioral measurements and greater information about the subcortical structures. Dr. Meier noted that the current research project “will likely inspire a new version of the DIVA model, which includes a simulation of the basal ganglia and the thalamus.” The potential modeling could also account for changes to such brain structures relevant to Parkinson’s disease and essential tremor.

Advancing Speech Research and Patient Care

While Dr. Guenther’s research primarily focuses on advancing an understanding of the neural mechanisms underlying speech, it also carries important implications for recipients of deep brain stimulators. DBS is primarily used to target tremor, rigidity, and other movement symptoms, but Dr. Guenther highlighted a potential unintended consequence: worsening speech following the implantation. That said, he stated an additional goal for his research is a greater clinical sensitivity to the preservation of speech while continuing to address other motor symptoms for patients with Parkinson’s and essential tremor.

“The goal is that when you implant somebody you’re not just improving their tremor. But you’re also improving their speech. You’re not degrading their speech with the surgery.”

In further identifying the distinct roles of the basal ganglia, cerebellum, and individual neural circuits, the research could inform electrode placement, reduce the negative effects on speech, and preserve the benefits of other motor symptoms. Dr. Meier also pointed to how the findings could impact speech therapy and rehabilitation, something he notes most patients with Parkinson’s seek out and/or require. In understanding the neural activity that supports speech learning, the research could inform speech therapy on how to better assist patients in learning or relearning speech motor patterns.

As the project progresses over five years, students at BU are also contributing to Dr. Guenther’s research. The Speech Neuroscience Lab, which Dr. Guenther directs, brings together students across disciplines of biomedical engineering, speech, language, & hearing sciences, neuroscience, and computer science. The labs research through computational modeling, behavioral experiments, neuroimaging, and neural recordings supports their broader mission of understanding how brain controls speech. For students on Dr. Guenther’s team, this means an opportunity to work alongside researchers and clinicians at Massachusetts General Hospital.

For Dr. Guenther, the current research project represents the continuing questions that have shaped over three decades of research. His present research gives him the opportunity to answer decades-old questions as technology advances, and he can acquire increasingly direct measurements of human brain activity.

“I’ve been modeling how the brain controls speech for thirty years. And we finally have technologies that allow me to answer some of the most detailed questions I have about that.”

Throughout the five-year project, Dr. Guenther and his colleagues will develop a greater understanding of how speech works in the brain while, for patients with speech motor impairments, prioritizing the protection of their ability to communicate.

Story by Dylan Nuñez