Optimistic Optimization

Neuromodulation is increasingly used to treat neurological diseases like depression, epilepsy, and Parkinson’s disease. “The most common treatments for these ailments are drugs, but many of them are very expensive, and they impact the rest of the body and other organ functions,” says Tay Netoff, professor in the Department of Biomedical Engineering (BME). “Neuromodulation has great potential to revolutionize neurologic therapies because it’s more precise in targeting specific parts of the brain. Electrodes are very localized and often patients don’t feel it.”

However, there is not enough research done on the optimal settings for both invasive and non-invasive neuromodulation therapies. And on the surface, the problem is impossible. With a deep brain stimulation device, for example, there are over 100 trillion possible settings. Plus, different patients can respond to the same settings in different ways. How to decide which one is best for a patient?

To address this issue, Netoff’s team – which is led by him, fellow BME professor Alex Opitz, and clinicians Gregory Worrell and Paul Croarkin from Mayo Clinic – is taking a patient-centered approach through Minnesota Personalized Neuromodulation Center, or MNPeNCe. They generally begin with 15-20 settings and ask the patient to rate each one. Sometimes, one of the settings is optimal, and sometimes, they try new settings based on ratings until they find one that works for the patient.

Image caption: Tay Netoff with patient (left) and implantable paddle electrode used to deliver electrical stimulation to the brain (right)  

The team is using the data from their patient interactions to develop software that guides clinicians on how to define optimal parameters for individual patients. “We are aiming to optimize neuromodulation therapies for a broad range of conditions, with each having its own desired outcome, such as improved motor function or reduced seizure frequency,” says Andrew Lamperski, an associate professor in the Department of Electrical and Computer Engineering. “While the details of the different treatments vary, the approach to optimization is similar: an algorithm takes information about which stimulation settings were used and how the patient responded, and predicts which settings will work best.”

This approach is no different than the algorithms used everyday on the internet, according to Lamperski. “They show up classically in product recommendations, personalized advertisements, and website optimization. In these cases, the algorithms collect your click data to help predict what they should display to keep you engaged.” Instead of genre preferences or popularity ratings, the MNPeNCe team uses factors like motor function, electrode stimulation amplitude, and subjective preference ratings, depending on the disease.

The team first came together through a Minnesota Partnership grant in 2023 to form the Minnesota Personalized Neuromodulation Center (MNPeNCe). Since then, MNPeNCe has secured other grants to conduct studies in partnership with industry, including a study on epilepsy with Medtronic’s Percept™ PC and spinal cord injury with Abbott’s Eterna™ SCS System. 

Now, MNPeNCe is working with IEM to secure a center-level grant. “IEM is really good at identifying and connecting research groups together. In the future, we want to continue partnering with IEM to grow our industry connections and recruit study participants,” says Tay. Alongside Micheal Lotti, program manager of IEM’s Center Accelerator Program, the team hopes to build a larger center by early 2025 to study neuromodulation optimization in epilepsy, depression, facial pain, Parkinson’s disease, and chronic pain. “We hope to maximize the impact on neuro devices for every patient,” says Netoff. “We want therapies to feel like a personalized selection for patients, what is best for their symptoms and how they feel.”