How AI Could Upgrade Brain Stimulation Therapies

How AI Could Upgrade Brain Stimulation Therapies
Stereotactic Neurosurgery operation, Pasteur 2 Hospital, Nice, France. A patient with Parkinsons disease is being treated with deep brain stimulation by implanting electrodes in brain and modulating cerebral electrical activity. (Photo: BSIP/Universal Images Group, Getty Images)
To sign up for our daily newsletter covering the latest news, features and reviews, head HERE. For a running feed of all our stories, follow us on Twitter HERE. Or you can bookmark the Gizmodo Australia homepage to visit whenever you need a news fix.

The human brain, just like whatever you’re reading this on, uses electricity to function. Neurons are constantly sending and receiving electrical signals. Everyone’s brain works a bit differently, and scientists are now getting closer to establishing how electrical activity is functioning in individual patients’ brains and how to stimulate it to treat neurological and psychiatric disorders. Some scientists are even using advanced AI predictive technology to enhance their brain stimulation therapy methods.

The idea of brain stimulation is not new and carries with it the cultural infamy of One Flew Over the Cuckoo’s Nest, a particularly brutal depiction of electroconvulsive therapy (ECT). Though ECT is still in use today to treat depression and other disorders, it’s come a long way since its invention in the 1930s.

Beyond ECT, there are therapies like vagus nerve stimulation (VNS), transcranial magnetic stimulation (TMS), magnetic seizure therapy (MST), transcranial current stimulation (tCS), and deep brain stimulation (DBS). Unlike noninvasive brain stimulation, which is done from the outside without any surgery involved, deep brain stimulation involves a surgery to actually implant electrodes in the brain. It’s typically used for treating serious neurological disorders like Parkinson’s disease. TMS, which is noninvasive, is currently used to treat depression.

Flavio Frohlich, director of the Carolina Centre for Neurostimulation at the UNC School of Medicine, told Gizmodo that other brain stimulation techniques we use today typically use much less electricity than ECT.

“The cultural connotation comes from electroconvulsive therapy, or ECT, which has been shown to be one of the most efficacious treatments for really very severe depression. But it also has a not-to-be-disregarded side effect profile,” Frohlich said. “The bottom line is, these noninvasive brain stimulation approaches that we’re studying today use about a thousandth of the energy of ECT.”

Though there are many exciting new advancements in brain stimulation techniques, the experts we spoke with for this article noted that brain stimulation typically shouldn’t be the first therapy someone attempts to treat a disorder. Deep brain stimulation involves surgery, which comes with its own risks, and the stimulation itself can potentially cause side effects ranging from speech problems to mood changes. Noninvasive brain stimulation can also cause side effects, which are usually relatively mild, such as the person experiencing headaches or pain at the treatment site.

But to millions of Americans who suffer from neurological and psychological disorders that don’t respond well to other therapies or medication, brain stimulation could become a very important alternative therapy. As things stand, many are living with these disorders without any effective way to relieve their symptoms.

“To come up with good ways to dynamically respond to the brain, we need to have a good model or a good understanding of what that response will be. We need to be able to predict it. The brain is very complex,” Frohlich said. “There is no mathematical model where you can say, if you put this in, this will happen. That’s really what we need to understand.”

What’s new and exciting about brain stimulation is that researchers are getting closer to being able to personalise these therapies to make them work better for patients. Frohlich said the idea is to monitor the brain while it’s being stimulated and observe how that particular person’s brain is responding to stimulation. Based on that information, you can adjust where you’re stimulating, the intensity of the stimulation, and more. This is called “closed-loop stimulation.” If a small amount of stimulation is producing the ideal response in a certain region of the brain, then you don’t need to increase the stimulation.

Essentially, we need to be able to predict how the brain will respond to stimulation, and we haven’t been able to do that. That’s what scientists are working on, and research published in the journal Nature Biomedical Engineering earlier this year showed how machine learning can be used to accomplish this goal.

Maryam Shanechi, an assistant professor of electrical and computer engineering at the University of Southern California, led the team that conducted this research. The team used a novel electrical stimulation wave to monitor brain activity and a machine learning model to predict how the brain would respond to deep brain stimulation. In this case, they were doing deep brain stimulation, and their system worked quite well. They found they were able to accurately predict how brain activity would change based on how they were stimulating the brain.

Let’s say a patient came in who was suffering from depression. The person conducting the therapy would measure their brain activity, and the machine learning algorithm would decode what the current symptom levels are based on activity levels and then determine how much stimulation would be required to reduce the symptoms. Shanechi said the process “repeats in real time until the symptom level is fully alleviated.”

“As far as the machine learning goes, really nothing we do requires invasiveness,” Shanechi told Gizmodo. “So the same kind of techniques we developed for deep brain stimulation… you can change the dataset to do noninvasive brain recording.”

Shanechi said figuring out how to effectively do personalised brain stimulation like this is important, because right now we’re applying the same kind of stimulation to everyone regardless of how their brain works, and disorders manifest differently in different people’s brains.

“When you don’t have personalisation, you just stimulate all the time. You don’t track the symptoms,” Shanechi said. “One thing that we want to do is develop machine learning algorithms and decoders that can, based on brain activity, decode in real time how somebody’s symptoms are changing, such that then you can stimulate as needed and provide the minimal amount of stimulation.”

In order to make personalised brain stimulation available widel, scientists will have to conduct clinical trials and get these types of therapies approved by the FDA. We don’t yet know if these therapies will simply be another way to treat neurological and mental disorders or if they could actually cure some disorders. Frohlich and Shanechi both said they’re hopeful that the brain’s capacity to rewire itself — an ability known as neuroplasticity — could mean that brain stimulation could cause a permanent change in how someone’s brain is wired and cure their ailment.

“Alleviating symptoms is quite valuable, and most groups aim at alleviating symptoms,” Shanechi said. “They hope that by alleviating symptoms in the long run, if you stimulate, you might also reduce the reliance of the patient on the therapy and potentially cure them by engaging these brain circuits and maybe engaging some adaptive processes.”

The various brains stimulation methods discussed above could provide crucial alternative neurological therapies. We still have quite a bit to learn about how these therapies really work and how effective they will be for the people who receive them. The real-life data on this medical technology is still in flux, but it is evolving and, hopefully, offering hope to millions of patients with treatment-resistant conditions.

Thor Benson is an independent journalist who has contributed to Gizmodo, The Daily Beast, The Atlantic, Rolling Stone, Wired and many other publications. Find him on Twitter at @thor_benson.