What can speech tell us about brain health?

Spanning topics from singing to AI, UBC’s Integrated Speech Research Laboratory (ISRL) has a diverse set of interests. Most recently, they have been looking into what speech can tell us about our brains — potentially contributing to a better understanding of the detection and treatment of neurodegenerative diseases.

“When people have done previous research on speech, oftentimes it has … taken the approach that speech is just sounds,” said Dr. Bryan Gick, founder and current director of the ISRL.

“In our lab, we take the approach that speech is movement.”

The work of the Speech and Neurological Disease research group is particularly significant in learning more about the embodiment of speech. Founded by PhD candidate Arian Shamei, the group focuses on the effects of neurological diseases on fine motor skills and their manifestations in speech.

“When something goes wrong with your body, that affects how that communication system works, and that’s the conceptual underpinning behind all of the work that we do in the lab,” said Gick.

Gick noted that speaking involves the same kind of mechanisms that we use for any other movement, but the signals our bodies are generating are much more complex. Spatial functions like walking are easier to visualize, so we don’t have that same understanding of speech and its processes yet.

Shamei began his research journey with a focus on neuroscience, but realized he was especially drawn towards linguistics and the effect of brain injuries on speech. Growing up, he witnessed the impact of aphasia — trouble with speaking resulting from a stroke — in his grandfather’s life, which inspired his interest in the field.

As he studied linguistics throughout his undergraduate years, Shamei was exposed to various theories of speech and motor control, but found them to be outdated. The up-and-coming field of state detection — identifying potential links between changes in speech and the mental state of the speaker — was what truly intrigued him.

On the brink of cannabis legalization, Shamei grew fascinated with the way that intoxication can affect the sound of one’s voice. Determined to explore this further, Shamei and his supervisor at the University of Victoria, Dr. Sonya Bird, found that cannabis reduces muscle tension and impairs one’s ability to coordinate movement in the mouth.

Around the time he started his PhD at UBC, Shamei noticed how machine learning was taking off — he optimized these advancements, finding ways to merge them with his interest in state detection. He started with intoxication detection machinery, but his interests started to gravitate towards disease states, like Alzheimer’s and Parkinson’s disease.

Because people are constantly talking, it’s an effective medium for continuous monitoring.

“The thing that they’re the most practiced at doing happens to be one of the most demanding things you can do,” said Shamei.

Shamei, Gick and PhD candidate Yadong Liu published an article earlier this year on the reduction of vowel space in individuals with Alzheimer’s disease.

“People with Alzheimer’s disease have really big impairments with movement trajectories throughout the body. But no one had really looked very hard at fine motor skills, like vowels-based articulation overall,” said Shamei.

The researchers accurately predicted that a general loss in bodily function caused by Alzheimer’s would be reflected by a decrease in the ability for tongue movement.

These findings are a step towards improving the way we diagnose and treat neurodegenerative diseases, possibly working towards modes of prevention and delay.

“As we learn more about the disease itself … we can detect those motor changes earlier and earlier, which enables early disease detection systems,” Shamei said.

—With additional reporting from Elena Massing.

This article is a part of The Ubyssey's 2023 language supplement, In Other Words.