UTSA Researchers Explore AI Speech Analysis for Mild Brain Injuries in Veterans

SAN ANTONIO — At UT San Antonio, researchers are investigating whether artificial intelligence (AI) can help detect communication problems in veterans with mild traumatic brain injuries (mTBI), with the potential to enhance diagnosis and improve access to care.

Backed by a $200,000 grant from the San Antonio Medical Foundation, the one-year collaborative study will apply natural language processing, a form of AI, to examine and quantify speech and language patterns tied to mTBI, building a tool designed to bring language analysis into everyday clinic appointments.

The project, funded by the San Antonio Medical Foundation, uses natural language processing to analyze recordings of veterans describing their injuries and life events — hunting for patterns standard assessments miss, UT Health San Antonio reported. It brings together speech-language pathology, psychology and veteran rehabilitation expertise across UT San Antonio, UT Health San Antonio and the South Texas Veterans Health Care System.

The effort is headed by three researchers: Rocío Norman, PhD, CCC-SLP, associate professor in the Department of Communication Sciences and Disorders at UT San Antonio’s School of Health Professions; Vivian Ta-Johnson, PhD, assistant professor of psychology in the College for Health, Community and Policy; and Alicia Swan, PhD, an mTBI expert and director of rehabilitation research at the South Texas Veterans Health Care System within the Department of Veterans Affairs.

What makes the study matter now is the diagnostic gap it targets.

CT and MRI scans for concussions can come back normal even when treatment is needed, and standard assessments don’t show how patients function outside the clinic, according to Dr. Monica Verduzco-Gutierrez, chair of the Department of Physical Medicine and Rehabilitation at UT Health San Antonio, who is not involved in the study and spoke to the San Antonio Report.

The signature wound standard tests miss

“Traumatic brain injury has been described as one of the signature wounds of war for the post-9/11 veteran cohort,” Swan told the San Antonio Report. “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.”

Norman, a speech-language pathologist who has worked with veterans for 18 years, told UT Health San Antonio that mild TBI differs sharply from severe injury. “Many of the problems associated with mild TBI involve higher-level tasks,” she said.

“That differs from severe TBI, in which someone may need to relearn how to walk or live independently. With mild TBI, the difficulties may emerge when someone goes to college, returns to paid employment or tries to communicate meaningfully with loved ones.”

She has seen the gap since treating Iraq and Afghanistan war veterans: many struggled to find words, organize thoughts or keep up in noisy, fast-paced conversations, even when standard assessments showed no impairment. Existing adult language assessments, built largely for stroke-related aphasia or severe brain injury, don’t capture that, she told NR Times.

“The work is rooted in clinical experience with patients whose concerns could not be validated objectively,” Norman told NR Times. “Objective measurement is fundamental to health care. We use the term ‘biomarker’ because we hope speech could become a biomarker for mild TBI.”

Turning stories into data

The team will analyze recordings already collected under a separate Pentagon-funded study Norman leads at the San Antonio VA, in which veterans diagnosed with mild TBI describe how they were injured and recount an important life event. Natural language processing converts those transcripts into measurable data: coherence, emotional tone, organization, detail, topic drift, fragmentation and repetition, according to the San Antonio Report.

“Natural language processing gives us a way to turn patterns in language into measurable data,” Ta-Johnson told UT Health San Antonio. “Recent technological advances have made some NLP tools powerful and accurate enough to capture and measure these facets of language in ways we could not before.”

Telling a story out loud, she told the San Antonio Report, demands attention, memory, planning and executive function — the same cognitive processes mild TBI can affect. Researchers studying early signs of Alzheimer’s disease have pursued speech as a similar early marker, a precedent the San Antonio Report notes.

The outlets frame the leadership differently: the San Antonio Report calls Ta-Johnson the lead researcher, while UT Health San Antonio names Norman as the project’s lead. Both describe the same three-person team.

Complicating diagnosis further, Swan told the San Antonio Report that post-traumatic stress disorder and other mental health challenges make mild TBI harder to pin down in veterans — and concussions have historically not been taken as seriously as they deserve, Verduzco-Gutierrez said, though she added 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 told the San Antonio Report.

Support for clinicians, not a replacement

The researchers stress the technology would support clinicians rather than replace them, UT Health San Antonio reported. Detailed language analysis today requires time-consuming transcription and coding by multiple reviewers; an automated tool could flag patterns during a routine office visit, NR Times reported.

The immediate goal is whether the method can identify communication differences and how they relate to symptoms such as anxiety, post-traumatic stress, depression and sleep difficulties, UT Health San Antonio reported. If the pilot produces meaningful findings, the team plans to pursue larger federal grants, expand to broader, potentially multi-site populations and study more specific cognitive deficits reflected in language.

Swan told the San Antonio Report the work could reach beyond mild TBI, serving other clinical complaints tied to traumatic brain injury exposure. “It can have a lot of ripple e”

Photo of author

Newsroom

Newsroom at Archyde.

U of T Rugby Player Helps Lebanon Win Asia Rugby Sevens Trophy