HealthMental Health


AI helps researchers develop personalized depression treatment

Researchers are studying how to use AI to personalize depression treatment and potentially predict a depressive episode before it happens.
Posted at 8:56 AM, Feb 09, 2023
and last updated 2023-02-09 08:56:11-05

Diane Davis of La Jolla, California, has lived with depression much of her life.  

"I think that people don't realize, for someone who struggles with depression, just how hard things can be. And it's simple things," she said. 

Her depression hit hard as she watched the news about the Uvalde, Texas, school shooting in May 2022. Nineteen elementary school students and two teachers were killed. 

" I just went into this deep dive. I was so depressed. Trying to get out of bed was difficult," she said.

Davis signed up to participate in a study at Neural Engineering and Translation Labs, or NEATLabs, at nearby University of San Diego. The study would be life-changing.  

"It really helped a lot. It made a big difference," she said. 

Scripps News visited NEATLabs and met co-director Jyoti Mishra.   

The team is studying how to use AI to personalize depression treatment and potentially predict a depressive episode before it happens.  

They put Scripps News' Lindsey Theis through the process to show how it works. 

The study participants undergo an EEG. Researchers painstakingly soak dozens of white leads in salt water and place them against the participant's scalp through the holes in a cap.  

The cap is hooked up to computers that measure electrical activity in the brain. That activity shows up as wavy lines. 

During the EEG, participants complete a series of "brain games" and different exercises. The two together are like a checkup for the brain.  

Dealing with the mental health impact of the Tyre Nichols video

Dealing with the mental health impact of the Tyre Nichols video

Expert says there has not been a lot of time for people to process the video because people might be experiencing various forms of trauma.


Based on those brain waves, Mishra and her team can tell how the regions of the brain in charge of emotion are doing.  

"We can record cognition in a very scalable manner. That is, we can take the tool to a clinic or a community setting and really gather data on many different aspects of cognitive functioning within a 30-minute appointment," Mishra said. 

Participants then use a smartphone app for a month. They check in three to four times a day, and self-report how they're feeling, what they ate, their social interaction, exercise and how well they thought they slept. 

Researchers also used smartwatches to track objective data on sleep, activity levels, diet and stress markers like heart rate and breathing.  

"All of that together forms this suite of assessments and gives us a lot of multi-dimensional data," Mishra said. 

Then, AI in the NEATLabs app analyzes the collected data and suggests lifestyle changes to lessen the impact of a depressive episode.   

For Diane Davis, it meant more exercise.  

"Using the tools they gave me and then having a doctor to talk to about what was going well, what wasn't — the tool itself is really good because it gives you feedback right away," Davis said.

Back at NEATLabs, we got a look at the next thing they’re testing.  

The researchers are using the same AI to predict when a depressive episode may come, and how long it might last. 

"What it's plotting is in blue is how their actual mood state is, how they feel in terms of how happy they are or how sad they feel in any given moment. And in red is how our models would predict their data," Mishra said. 

The initial signs are promising, Mishra says. But it's early in the research phase.  

The NEATLabs app is not meant to replace a doctor or mental health provider or their care, but to assist in treatment.