Professionals within the neurodiagnostic field are realizing that our brains’ electrical fields can help us understand various diseases in ways we never have — it has the  industry on the cusp of major transformation.

The brain changes before patients or doctors begin to see clinical symptoms, says Dr. William Bosl, professor of Health Informatics and Data Science at the University of San Francisco, affiliate faculty member at Boston Children’s Hospital and lecturer at Harvard Medical School.

“That opens up a window of opportunity for monitoring brain function over time to detect changes that might be pointing to emerging pathological conditions before we actually see the symptoms,” Bosl said. “In order for that to happen, we need to have some way of easily and relatively inexpensively measuring what the brain is doing. If it’s too expensive, we’re not going to do it routinely, especially if somebody has no symptoms. Why would we be spending a lot of money to measure something that may not arise?”

Bosl — who gave the Ellen Grass Lecture at the ASET 62nd Annual Conference in San Diego — is an advocate for research that would enable regular scans of brain activity at well-baby checkups. Research suggests early signs of emerging autism can be detected long before the infant displays symptoms.

Newly available EEG devices open new possibilities for quick and easy brain scans, he says, while the data they generate create new research opportunities, such as early detection of autism, monitoring epilepsy treatment, or early detection of other mental and neurological disorders.

Bosl has identified four emerging technologies which underscore a reimagining of the neurodiagnostic field:

  1. Electrodes and new EEG hardware

Low-cost, wireless electrodes are easier to use. A prevailing concern among technologists is that advancements in technology may lead to their displacement from specialized neurology and epilepsy settings, though Bosl does not see it that way.

“I really don’t think that’s going to happen,” he said. “These clinics require a skilled technologist to place them properly with the gels and so on, but you get high-quality signals. I don’t think that’s going to end quickly. The new devices might slowly diffuse into that setting and make the tech’s job a little easier, but they’re not going to be displaced.

“Rather, that technology opens up new opportunities in routine care settings where you don’t have time to do the full gel workup and exactly place the sensors — where time, ease of use, and quantity of measurement are paramount.”

  1. Ubiquitous electronic health records

As hardware progresses and its utilization increases, the information it generates must be gathered and stored. Recordings must be organized to be tracked over time. This requires electronic storage associated with each patient.

  1. Dynamical systems theory

The basic idea behind the theory, Bosl explains, is that systems — a collection of components that are connected in some way — change over time.

“A dynamical system, this network of components, changes according to some rule or function or law,” he said. “It turns out that the brain is, in fact, a dynamical system, a complex dynamical system, according to the fundamental definition of a dynamical system. Neurons are the fundamental components. The patterns of electrical fields of the brain generated by neurons change continuously in time.”

EEG recordings contain much more information than previously realized, Bosl said, and algorithms that come out of this theory tell us how to access it.

“There’s much more information about the brain than is currently being used,” he said. “Much of that, we’re starting to show, is relevant to all sorts of brain disorders, from psychiatric and mental disorders, neurodevelopmental disorders, to more traditional neurological disorders.”

  1. Artificial intelligence and machine learning

Advancements in AI allow us to find patterns within the complicated data EEG recordings generate.

“Dynamical systems theory, in principle, gives us all kinds of information about the brain, but it’s complicated. It’s a pile of numbers,” Bosl said. “It’s difficult for us to look at it and with our eyes make sense of it. But that’s what machine learning can help us to do. It can find patterns in that.”

Don’t worry about job displacement

Technological advancements in the neurodiagnostic field are not likely to create job displacement. If anything, Bosl says, it will create rapid expansion.

“But there is work to be done first,” he says.

In his keynote address, Bosl compared those working in the industry today to the “computers” from the early days of NASA. Unlike today’s machines, NASA’s computers in the 1950s were humans who would perform arithmetic on a card and pass it along to the next human computer.

When today’s computers displaced the professionals who performed computations by hand, those individuals recognized an opportunity: They possessed the skill set to program the new machines. They became the first generation of computer programmers — a profession far more numerous today than human computers were at the time.

“A new field developed called computer science that wasn’t just math. It wasn’t physics or engineering or any of those things, but it had elements of all of them,” Bosl said. “I see the analogy for neurodiagnostic technology to develop a new science of neurodiagnostics, which integrates all the technologies I talked about, but is more than any one of them.”

Adding ‘routine measurements into routine care’

Similar to how most doctor appointments include a blood pressure check, Bosl makes the case for incorporating regular brain scans into different areas of care. Doing so, however, is not currently easy and will require considerable research to develop and implement the algorithms and electronic health record infrastructure.

“There are bugs to be worked out,” he says. “Another big challenge is, how do we save that data with the patient medical record? New-generation EEG devices can wirelessly send data to a file at a nearby laptop over Bluetooth connections. Great. Well, a doctor doesn’t have time to fool around with this laptop now and copy the files over to an EHR. So, somebody’s going to have to figure out, implement software to make all this happen automatically, and save the data somewhere. Where does it get saved? How is it transferred? What are the data formats? There are a lot of nitty-gritty details.”

For this to become practical, Bosl adds, we must develop the aforementioned machine-learning algorithms to interpret the copious amounts of data routine scans would generate.

Then, the results have to be presented to clinicians and patients in a way that is appropriate for the audience. Neurodiagnostic technologists can contribute to all of this.

Listen to our interview with Dr. William Bosl below: