A physics approach to stroke rehabilitation

CONTACT: Molly Goaley

Sarah Hulbert George is a biophysicist whose research focuses on stroke rehabilitation
Physics alumna Sarah Hulbert George focuses her research on stroke rehabilitation and recovery.

We often hear stories about what happens to the body during a stroke. The sudden loss of oxygen to the brain, the onset of paralysis in the face or limbs, blurred vision and slurred speech, the alarming disorientation and need to seek help. With one out of every six people experiencing a stroke at some point in their lives, a survivor’s personal account can be not only gripping and terrifying, but all too often relatable.

What we hear less of are the stories about what happens after a stroke. For survivors who retain long-term or permanent damage, recovery and rehabilitation can be a grueling process. Imagine the most automatic tasks becoming impossible to perform, with words and thoughts swimming around somewhere in the mind but never quite making their escape to the surface.

Consider for example the use of a fork: your brain must send a signal for your arm to reach the handle, your hand to close around it, your arm to raise it, your mouth to open, and your arm to move the fork into your mouth. Simple. Except, your arm and hand won’t comply. And when you try to explain this to someone else, you can’t readily recall the name for what you’re eating.

Regaining speech, vision, memory and muscle movement takes a significant toll on the body and mind. While recovering from a stroke comes with small rewards in each daily success, it also comes with frustration, exhaustion and a host of other emotions and physical states when rehabilitative therapies do not go as planned.

To improve treatment, it’s imperative that scientists understand how the brain and body connect during every step of recovery. This is the type of neurophysiological activity WMU physics alumna Sarah Hulbert George B.S. ’13 examines. Specifically, she studies brain activity during movement to identify the unique characteristics that indicate whether a person will execute that movement, such as lifting an arm, well or poorly.

Taking a physics approach to understanding the mysteries of stroke has unique advantages. Because physics extracts the fundamental, underlying information at the root of complex problems, it can offer a new perspective into how the brain functions during and after a stroke.

Hulbert George uses her undergraduate training in physics from WMU to examine the neurological phenomena that control movement, approaching the process like a complicated physics problem. “I did my undergraduate work in physics because I wanted to understand the way things work at the base level,” she says. “In my opinion, physics addresses the questions of how and why better than any other field.”

By the time she was a senior at Western, Hulbert George also realized she wanted her research to impact people first and foremost. She chose to pursue biophysics because it takes the quantitative, rigorous approach of physics and applies it to biological problems. Now a Ph.D. candidate in biophysics at The Ohio State University, Hulbert George employs physics-based principles on a daily basis in her current project, titled “Predicting Your Next Move: Real-Time EEG and Kinematic Analysis Motor Neurorehabilitation and Neurofeedback.”

“I use biophysics to approach my neurophysiological and neurorehabilitative research from a quantitative, objective standpoint,” she says. “The physics of electricity and magnetism are the foundation for electroencephalographic (EEG) and single-unit recording, which are key components of my work.”

Because a big part of Hulbert George’s research involves examining brain activity during movement, she spends a substantial amount of time investigating kinematics – such as position and velocity – which are the basis of mechanics, the branch of physics dealing with the study of motion.

“You can imagine that for a person who has had a stroke and now suffers from poor motor control in one of their arms, as often happens after stroke, they might try some rehabilitation to gain back the lost function,” she says. “This is where my research comes in – if we can identify the features that indicate a person’s quality of movement just before they make the movement, then we can provide them with real-time neurofeedback during their rehabilitation practice.”

Hulbert George collects kinematic and brain signal data by having participants play Recovery Rapids, a video game designed to deliver motor practice for the weaker hand and arm following a stroke. Participants’ body movements are captured by the Microsoft Kinect sensor to power a digital kayak downstream. Meanwhile, Hulbert George and her colleagues use EEG to simultaneously record their brain activity.

“After the session, we process both sets of data and extract the features from EEG signals using machine learning,” Hulbert George says. “Once we extract the features, we can do a correlation analysis to answer questions like, ‘does this feature always, or almost always, show up right before a good movement?’”

If during rehabilitation a person is about to perform a “good quality” movement, Hulbert George can detect the characteristics using EEG and provide visual feedback that sends the message, “great job, keep up the good work.” This type of positive reinforcement, she says, increases the efficiency and quality of therapy for stroke survivors.

By approaching the research from a physics perspective, Hulbert George is uniquely positioned to weed through complex information and find its most revealing details – the neural signatures that indicate the quality of a person’s movement.

“We’re not really sure what we’re looking for in terms of the features themselves,” she says. “The beauty of unsupervised machine learning for feature extraction is that we do not need to know 'a priori' what features will come out. Using these methods, we can let the signals themselves tell us what features uniquely characterize them.”

Employing the quantitative rigor of biophysics to reveal how the brain functions after a stroke – and using that knowledge to improve rehabilitative therapy – models what Hulbert George hopes to continue throughout her career. Her ultimate goal, she says, is to help people achieve a better quality of life.

“The most exciting thing about my work goes back to why I turned to biophysics in the first place,” she says. “Because it has the potential to impact people, and not just tangentially, but in a very direct and immediate manner.”

View this story and more in the 2018 issue of WMU's Arts and Sciences Magazine.