San Antonio researchers will use artificial intelligence to test whether subtle language patterns can help uncover mild traumatic brain injuries in veterans.
Researchers with UT San Antonio, the South Texas Veterans Health Care System and UT Health San Antonio are collaborating on the yearlong study, which is supported by $200,000 in grant funding from the San Antonio Medical Foundation.
If the researchers find language patterns that can reliably be linked to cognitive functioning, the findings could lead to new clinical assessments for concussions, according to Vivian Ta-Johnson, UTSA psychologist who specializes in natural language processing and lead researcher on the study.
“Language is a really rich source of behavioral information,” she said. “And with the recent developments in AI, we’re finally able to fully take advantage of that rich data and information that it provides us.”
According to Dr. Monica Verduzco-Gutierrez, chair of the Department of Physical Medicine and Rehabilitation at UT Health San Antonio, there are shortcomings to current tools for diagnosing concussions, which can have subtle symptoms and often go untreated.
Standard assessments entail understanding how the injury occurred, whether the patient lost consciousness and whether they had immediate cognitive or memory issues. But these assessments don’t tell doctors how patients are functioning outside of the clinic, Verduzco-Gutierrez said. And CT and MRI scans for concussions can come back normal even when treatment is needed.
Additionally, she said, concussions have historically not been treated as seriously as they deserve, though the tide is changing.
“A lot of people might hear of the word concussion and don’t realize that that’s actually a mild traumatic brain injury,” she said. Verduzco-Gutierrez, who is not involved in the new study, helped craft new concussion care standards that emphasize the seriousness of even mild TBIs.
Concussions can have long-lasting consequences, including persistent issues with cognition, depression, anxiety and headaches, among others.
“Traumatic brain injury has been described as one of the signature wounds of war for the post-9/11 veteran cohort,” said Alicia Swan, director of rehabilitation research at the South Texas Veterans Health Care System. “We’re now over half a million of that service cohort that have any type of traumatic brain injury, and approximately 85% of all traumatic brain injury is mild in severity.”
The presence of post-traumatic stress disorder and mental health challenges can make diagnosing mild TBI in the veteran population even more complicated, Swan said.
Over the next year, Ta-Johnson and her team will use AI to analyze the data already being collected at the South Texas VA through a separate Pentagon-funded study. That study has a language component that involves recording stories from veterans who have been diagnosed with a mild TBI.
The local researchers will use the recordings to quantify subtle language patterns like topic drift, word and sentence fragmentation, repetition and coherence, among other metrics.
“Using language and just speaking to tell a story requires actually quite a bit of cognitive work all at once,” Ta-Johnson said. “And those are many of the same cognitive processes that are affected or can be affected by (mild) TBI.”
UT Health’s Rocio Norman, who researches mild TBIs, will help determine if the patterns they identify are clinically meaningful, potentially paving the way for a new clinical tool.
Researchers who study early signs of neurodegenerative diseases like Alzheimer’s have also looked at language as a potential tool to catch subtle changes in cognition.
Swan said that the new research could also have different applications beyond mild TBI alone.
“This could serve a lot of other clinical complaints that are related to the many that you might experience as a byproduct of exposure to traumatic brain injury,” she said. “It can have a lot of ripple effect beyond just this specific population.”
