In the original Star Trek series that aired in the 1960s, creator Gene Roddenberry posited a handheld medical device that, when passed in close proximity to a patient, provided the doctor with an overview of the patient’s health. Scientists at GE’s Global Research Center are not yet working on such a device, but their research may be laying the groundwork for something similar.
The work entails detecting and decoding subtle electrical signals of the human body’s nervous system. Charles Scott Sherrington, who the Nobel Prize in Medicine in 1932, proposed that neurons, or nerve cells, carry signals between the brain and the limbs and organs in the form of electrochemical impulses. Many of these impulses run along the vagus nerve, which starts at the brainstem and runs through the neck and abdomen to most of the body’s organs. The vagus nerve is kind of like a central trunk cable in a telecommunications network: It is a bundle of hundreds of thousands of nerve fibers, each carrying on a “conversation” between the brain and one of the body’s organs. Eavesdropping on these conversations could yield valuable information about the state of the body’s health.
A team of scientists led by Dr. Kevin Tracey, president and CEO of the Feinstein Institute for Medical Research at Northwell Health, set out recently to do exactly that. Specifically, Tracey wanted to see if it was possible to infer something about the human body’s inflammation response — an aspect of the immune system that often indicates an infection or injury — by reading signals on the vagus nerves of lab mice. Tracey’s team implanted an electrode below the vagus nerve of each mouse and captured the electrical signals running through the nerve. Then they injected the mouse with cytokines — protein molecules that regulate inflammation — and monitored the signals for changes. “To have a baseline electrical signal that tells us what is healthy and what is unhealthy, that’s’s a completely new way of making a diagnosis,” says Sue Siegel, GE’s chief innovation officer who also runs GE Ventures and helped establish the partnership.
Siegel says the joint effort, which could lead to new insights into the triggers and treatments for inflammatory diseases such as rheumatoid arthritis, psoriasis, multiple sclerosis, lupus, celiac disease and other ailments, “shows a way to bring venture capital principles to the program management of an R&D project. It helps create value for everybody—corporations, government, private foundations, philanthropists, grant money,” she says.
Top image: By reading these signals, the team was able to “predict” the presence of cytokines — protein molecules that regulate inflammation — 83 percent of the time. “That’s a very high number in biology,” says Jeffrey Ashe (above) , a principal engineer at GRC. Image credit: GE Reports. Above: The team implanted an electrode below the vagus nerve of each mouse and captured the electrical signals running through the nerve. Image credit: Getty Images.
It fell to Peter Lorraine, a physicist and mathematician at GE, and Theo Zanos, head of the Feinstein Institute’s Neural Decoding and Data Analytics Lab, to analyze the signals.
The team started looking for “spikes” in the signal that generally indicate the presence of discrete impulses between the brain and an organ. The spikes are short pulses around a millisecond long but vary substantially depending on their function and location in the nerve bundle. The pulses also traveled at different speeds — from half a meter per second (mps) to 100 mps — and ran in both directions, to and from the brain. With the aid of machine-learning algorithms, which identified patterns, the team sorted the signals into families based on shape, amplitude (the height of the spike) and other characteristics. They found two to eight distinct families, depending on the individual mouse.
The most interesting signals to the researchers were those that correlated closely with the cytokine injections — likely to produce the same signals the body uses during an illness to trigger inflammation. By reading these signals, the team was able to “predict” the presence of cytokines 83 percent of the time. “That’s a very high number in biology,” says Jeffrey Ashe, a principal engineer at GRC. “With most things in physiology you’re looking at a predictive rate of between 70 and 80 percent.” The team published the results recently in the Proceedings of the National Academy of Sciences of the United States of America.
The hope is that one day, reading nerve signals could be a noninvasive way of making diagnoses in real time — without having to wait hours or days for test results. Lorraine and Ashe are also exploring ways of reversing the play: generating artificial nerve signals to alter the body’s physiology. The work suggests the possibility that therapeutic devices might one day generate signals that treat depression, arthritis or other ailments without drugs or surgical procedures. “Our ultimate goal,” says Lorraine, “is once we learn the language, we want to speak the language back to the nerves.”