Researchers Merge AI with Brain Stimulation to Dramatically Boost Mental Function

Novel Process May Lead to Treatments for Severe Mental Illness

Researchers from the University of Minnesota Medical School working with specialists from Massachusetts General Hospital have demonstrated that a combination of artificial intelligence and brain stimulation can dramatically improve a number of mental functions. This novel approach may result in breakthrough treatments for people suffering from a wide range of severe mental illnesses.


In a pilot study, researchers hoped to target a region of the human brain known as the internal capsule. This part of the brain oversees cognitive control, which is the ability to shift from one thought or pattern of behavior to another. A dysfunction in this brain region and related abilities is common across a range of mental illnesses.

“An example might include a person with depression who just can’t get out of a ‘stuck’ negative thought,” said Alik Widge, MD, PhD, an assistant professor of psychiatry at the U of M Medical School and the study’s senior author in a press release outlining the procedure “Because it is so central to mental illness, finding a way to improve it could be a powerful new way to treat those illnesses.”


With the help of Sydney Cash, MD, PhD, an expert in epilepsy research, and Darin Dougherty, MD, a specialist in clinical brain stimulation at Massachusetts General Hospital, Widge and his team observed the brains of 12 epilepsy patients. That’s because these patients were already having hundreds of tiny probes placed into their brains so doctors could help locate where seizures were forming, This unique brain map helped Widge and his team zero in on the fact that directing electrical stimulation to this one particular region of the brain helped those suffering from these types of symptoms to break the negative cycle and move on.

To accomplish this goal, the team had to write a specific algorithm that let them track the cognitive control of each of the 12 patients in the pilot study, looking at data from their actions in lab tests as well as their actual brain activity to see if they were falling into a negative cycle. Throughout the study, the AI software learned how to direct the boosts of electrical stimulation whenever a patient started to perform poorly on laboratory tests measuring cognitive control.

“This system can read brain activity, ‘decode’ from that when a patient is having difficulty, and apply a small burst of electrical stimulation to the brain to boost them past that difficulty,” said Widge. “The analogy I often use is an electric bike. When someone’s pedaling but having difficulty, the bike senses it and augments it. We’ve made the equivalent of that for human mental function.”


The next step for the research team is more clinical trials. According to Widge, this type of research can be done with existing tools, meaning that once a trial receives formal approval from the United States Food and Drug Administration, moving this to actual real world treatment may occur rather quickly.

“The wonderful thing about these findings is that we are now in a position to conduct clinical trials to further demonstrate effectiveness and then hopefully move to helping treatment-resistant patients who are in desperate need for additional interventions to treat their illnesses,” said Dougherty.

“This could be a totally new approach in treating mental illness,” added Widge. “Instead of trying to suppress symptoms, we could give patients a tool that lets them take control of their own minds. We could put them back in the driver’s seat and let them feel a new sense of agency.”

Published in Nature Biomedical Engineering, the research was supported by a grant from the Defense Advanced Research Projects Agency, among others.

Follow and connect with author Christopher Plain on Twitter: @plain_fiction