However, EEG and MEG cannot provide a way to look beneath these patterns to see what neurons are actually doing to produce this electrical activity. Without this, according to April Levin, an autism specialist at Boston Children’s Hospital, researchers have been limited to studying patterns highlighted by previous research, with only a speculative understanding of how they are generated. 

“We look under the lamppost because that's where the light is, not because that's where we should be looking,” said Levin. 

After nearly two decades of methodical, mathematical work – weaving together her understanding of neural circuitry from animal studies with the physics of how MEG and EEG pick up and translate electrical and magnetic signals in humans – Jones created a free tool called the Human Neocortical Neurosolver, or HNN. The tool simulates the behavior of neurons inside a patch of human neocortical tissue and converts this activity into waveform EEG and MEG patterns. Researchers can adjust the parameters in HNN to generate a waveform that matches the one they’ve recorded in a human – and then “look under the hood” to see what neuronal activity would produce that particular pattern.

“She created a Rosetta stone for EEG and MEG,” said Christopher Moore, Adam and Margaret Korn Professor of Brain Science, who collaborated with Jones on the early stages of her research. “It’s all guesswork unless you have this tool. With it, EEG and MEG goes from being a largely observational modality, which can give patterns but not usually mechanistic insight, to allowing us to create a principled understanding of how specific cell types in specific brain areas contribute to a signal – and to the changes in that signal in disease. This kind of insight is a major advance, exactly the kind of detailed map that can allow us to target new drugs.”

Researchers across various fields are now using HNN to decode their data, test theories, and validate or generate hypotheses.

Fernando Maestú, a professor at Universidad Complutense in Madrid, studies Alzheimer’s disease. He has theorized that inhibitory neurons malfunction in the disease, causing excitatory neurons to become overactive – and drive the release of amyloid proteins that form the disease’s hallmark  “plaques.” According to Maestú, while much of Alzheimer’s research has focused on this toxic build-up of proteins, there has been less focus on how neurons behave in this toxic environment. 

When he saw Jones give a presentation about the HNN, he saw a perfect framework to test his hypothesis. The two labs have now completed initial studies on Maestú’s data – years’ worth of MEG recordings from patients experiencing early cognitive decline, some of whom later developed Alzheimer’s disease. The team has identified an activity pattern that predicts whether a patient with a diagnosis of mild cognitive impairment will go on to develop Alzheimer’s disease within two and a half years. With support from a Carney Innovation Award, they are now using HNN to model the neuronal behavior that underlies this pattern.

Meanwhile, for April Levin, a reliable biomarker for autism is the “holy grail.” Clinical trials often fail without an objective, measurable effect – even when patients report improvement, Levin says. When she first learned that the Jones lab had identified a specific pattern in an autism dataset where HNN potentially could provide a neural-level interpretation, she was wowed. “This was the first time that I was able to look at my data in a way that could really inform how we think about developing these biomarkers that are so very needed,” said Levin. 

Jones and Levin will be publishing together soon and are in conversations about further collaborations using HNN.

Researchers looking to test therapeutics in humans particularly see the HNN tool as a crucial step. Mohamed Sherif, an assistant professor of psychiatry and human behavior at Brown, studies treatment-resistant depression. Although medications can help patients, he says, it is often not clear how they work. “This tool allows us to get actual insights,” Sherif said.  Similarly, Dr. Ernie Pedapati, who treats children with Fragile X syndrome at Cincinnati Children’s Hospital Medical Center, knows firsthand the heartbreak of families when a new therapy ends up not working. 

“By using tools like the HNN, we are more confident that the treatments we are advancing into clinical trials are more than just a guess.”

THE HNN is user friendly. 

The tool provides an easy-to-understand interface that researchers or clinicians can operate without advanced training in mathematics and physics – a challenging but deliberate effort by Jones to boost the technology’s accessibility. To continue supporting the labor involved in providing an open-source tool for free, Jones and two of her postdocs are teaming up with pharmaceutical companies to create a commercial application that can be used to develop neurotherapeutics. 

Additionally, Maestú sees HNN as a transformative tool for low-income regions of the world, where hospitals have EEG machines but lack access to more expensive technologies, such as PET or MRI scans. “A platform where they can just upload their EEG to provide an answer, such as a certain percentage of risk for developing dementia, will be very impactful for them,” said Maestú. 

The growing impact of the Human Neocortical Neurosolver is something Jones says she never imagined back when she first wondered about the potential of a math-based approach. 

“When we make a prediction with HNN in our lab and then it's holding true [in the clinic], I get chills on my skin,” said Jones. “We might actually be able to help real people with real diseases.”

______

Jones’s work has received support from the Carney Institute’s Innovation Awards program. HNN was developed in part with support through the NIH’s Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative, a program aimed at accelerating the development of neurotechnologies.