It might start when a person stumbles over a word. Within as few as four years, a patient with primary progressive aphasia, a neurological language disorder, could go completely mute, said Maya Henry, director of the Aphasia Research and Treatment Lab. 

According to an artificial intelligence study released by UT researchers on Feb. 6, a new AI tool can read brain activity and translate it into continuous strings of text. Alexander Huth, associate professor of neuroscience and computer science, said his team began by training AI models on brain scans of participants as they listened to podcasts. 

“Our normal computational models take the words in the podcast that you’re listening to and predict how your brain would respond,” Huth said. “What we were able to do once the models got good enough was actually reverse that process to take the brain responses and figure out what were the words that caused them.” 

The first iteration of the brain decoder tool appeared in a 2023 study from the same lab. However, Huth said this original decoder posed some practical limitations. First, for the model to be trained on the brain scans of a patient, the patient had to be able to actually process the language of the podcasts they were listening to — a capability some aphasia patients lack, Huth said. Second, even if the patient processed language, they would need to complete over a dozen brain scan sessions before the model was ready, Huth said.

Huth said the 2025 study was able to remove the need for long training sessions and any verbal processing from the patient by moving the models they had already built for one person onto another.

“We can take a little bit of data, like an hour of data from people watching short movies, and then use that to build a language decoder for that person,” Huth said.

Henry, an associate professor in the Department of Speech, Language and Hearing Sciences, said the next step for this research is to recruit patients with aphasia to participate in decoding experiments and to share their ideas for this technology. 

“Communication is how we define ourselves and how we connect with the world,” Henry said. “(This technology has) the potential to not just learn something about how the brain supports language and how we can decode, but to really impact people’s lives.”