Woman and fNIRS machine

AI that can diagnose tinnitus through brain scans can improve treatment

By Donna Lu

Brain scans could help an AI diagnose tinnitus

Mehrnaz Shoushtarian

Artificial intelligence that can diagnose tinnitus based on brain imaging results, rather than subjective tests, can help improve treatment for the condition.

Mehrnaz Shoushtarian of the Bionics Institute in Melbourne, Australia, and her colleagues have developed an algorithm that can be used to determine whether and how severely a person has tinnitus.

The AI ​​can detect the presence of tinnitus with an accuracy of 78 percent and differentiate between mild and severe forms with an accuracy of 87 percent.


Chronic tinnitus affects around 15 percent of adults. The condition is usually diagnosed through a hearing test, self-reporting, or a subjective questionnaire.

First, the team used a non-invasive imaging technique known as functional near-infrared spectroscopy (fNIRS) on 25 people with chronic tinnitus and 21 people without the condition. It uses infrared light to measure blood flow and oxygen levels in specific brain regions that correspond to brain activity.

The team measured fNIRS signals while participants were presented with both visual and audible stimuli: a display of circular checkerboard patterns and 15-second noise segments.

According to Shoushtarian, many visual-auditory nerve pathways interact in both people with and without hearing impairments. Previous research has shown that people with tinnitus have decreased activity in the cuneus, a region of the brain that is involved in visual processing.

People with tinnitus were asked to rate how loud and annoying their condition was. These results were correlated with patterns of brain activity based on their fNIRS signals.

The researchers found that people with severe tinnitus had higher background connectivity between specific areas of the brain. In patients with louder tinnitus, the brain reactions to both visual and auditory stimuli were significantly reduced. The team believes this is because the increased background neural activity in people with tinnitus affects the brain’s responsiveness.

The researchers then trained an algorithm on the results of the fNIRS and tinnitus severity levels. The fact that the AI ​​can objectively differentiate between mild and severe tinnitus can help improve management of the condition, Shoushtarian says. It is currently difficult to gauge how successful treatments are because the results are based on a person’s subjective reports over time.

Journal reference: PLoS One, DOI: 10.1371 / journal.pone.0241695

Sign up for our free Health Check newsletter for a round-up of all the health and fitness news you need to know every Saturday

More on these topics: