In 2009, Frank Guenther collaborated with a team of researchers on a game-changing device that measures signals from the area of the brain controlling speech and translates those signals to a computer in real time.
They implanted electrodes in the brain of a patient with locked-in syndrome—the patient was paralyzed but could move his eyes. Then, they used the device, called a speech brain-computer interface (BCI), to translate signals from the electrodes into acoustic signals, allowing the patient to create vowel sounds on the computer. The more the patient used the BCI, the more adept he became; eventually, he learned to control the computer synthesizer just by thinking about producing a certain vowel sound.
“As far as I know, to this day, this patient was the first and only who was implanted [with electrodes] specifically in the area of the motor cortex that controls speech output,” says Guenther, a professor of speech, language, and hearing science and biomedical engineering. “The reason we only could do vowels at that time was we had a very low capacity electrode system compared to what’s available now. There were only three wires that were implanted in the cortex that we recorded from, so we could only pull out so much information.”
These days, there are electrode systems with more than 100 wires. Still, Guenther says, the project highlighted the potential for restoring communication in people with locked-in syndrome. He is eager to see what could be accomplished with FDA approval of new research on such a device given the advancements in electrode systems. “Even with just those three wires, the patient was able to learn to produce vowels with this synthesizer in real time so you could hear it coming out of the speaker as he was producing it.”
Today, Guenther, a professor of speech, language, and hearing sciences and of biomedical engineering, is looking into the potential to use electrode system implants and BCIs to address speech disorders such as stuttering, which his lab has studied in recent years.
“Stuttering is very poorly understood,” says Guenther. Over time, he says, researchers have settled on a set of structures in the brain called the basal ganglia as the site of impairment for both Parkinson’s disease and stuttering. “We know for sure it’s [where the impairment is] for Parkinson’s because dopamine going to the basal ganglia is depleted, and that’s what causes the motor problems. With stuttering, there’s something else going on with the basal ganglia, but something not nearly so obvious.”
Using the Technology to Treat Stuttering
According to the National Institutes of Health, around 3 million Americans stutter. Guenther and his lab have created a neurocomputational model, which they are using to get to the bottom of what is happening in the brain when it comes to stuttering. The model is designed as a neural network, with processing nodes that are assigned specific locations in the brain. A set of equations govern how these nodes interact, imitating how the brain works. In researching stuttering, Guenther and his team have sometimes purposely damaged parts of the model to try to mimic what is happening in the brain with such a disorder. “We can not only see what happens to the speech—because the model actually does the speech output—but we can also see how the brain activity changes as a result of what’s happening,” he says.
They have used the model to make predictions about different types of stuttering, and they are comparing those predictions to tests and brain imaging on people who stutter versus those who don’t. “We also try to look for correlations between certain stuttering behaviors and certain brain characteristics,” Guenther says. For example, some people repeat the same part of a word, while others have blocks, meaning they get caught up on saying a word. “There is one set of people who have mostly blocks we are looking at, and we think there might be a particular subsystem within the basal ganglia that is impaired.”
Guenther points to a procedure for Parkinson’s called deep brain stimulation—similar to a BCI—that gives him hope for using electrode system implants to treat stuttering. “The basal ganglia have two pathways,” he says. “One is excitatory for movement and the other inhibits movements, and they have to be in the right balance. Parkinson’s disease is when the excitatory system is too weak and the inhibitory system is too strong.”
With deep brain stimulation, doctors place an electrode in a particular part of the basal ganglia of a patient with Parkinson’s to impair the inhibitory path and balance it with the excitatory path. “That can restore function for quite a long period of time, depending on how fast one’s Parkinson’s is evolving,” Guenther says. He and his lab have been working with doctors at Massachusetts General Hospital who perform deep brain stimulation surgeries to learn more about the procedure’s impact on the basal ganglia and to do tests on a neurocomputational model. In July 2023, Caroline Brinkert joined the speech, language, and hearing sciences faculty as a lecturer, and will be assisting Guenther in recruiting subjects for stuttering research.
“It’s not out of the question that in a decade or two, there may be some option for people who have a severe stutter, such as a deep brain stimulator or a drug that affects the excitatory pathways, that takes care of the stutter,” Guenther says. “I think our modeling may help figure out how to design something along those lines. The possibilities are exciting.”