ASU researcher is your daily steps accountability partner

Daily exercise is as essential for a healthy body as an oil change is to a properly operating car. It improves performance, increases clarity of thinking and can enable you to live longer. Yet, in the United States alone, people on average are physically inactive for almost eight hours each day, according to the National Library of Medicine. While arguably more comfortable, sedentary behaviors such as spending long hours seated in an office, playing video games or binge-watching television shows are a leading cause of 5.3 million deaths worldwide each year.

Mohamed El Mistiri, a postdoctoral research scholar in the School for Engineering of Matter, Transport and Energy, part of the Ira A. Fulton Schools of Engineering at Arizona State University, has received the Dean’s Dissertation Award from Dean Kyle Squires for applying his control systems engineering skills to develop personalized, just-in-time behavioral interventions to increase physical activity of sedentary adults.

From basic electronic devices like microwaves and cameras to advanced systems likes electric vehicles and spaceships, control systems engineering, also known as control engineering, is used to regulate the behavior of these devices and ensure that they deliver the desired output.

“Control systems engineering applications in behavioral science are an emerging research direction that my supervisor, Daniel E. Rivera, and the Control Systems Engineering Laboratory, or CSEL, at ASU, has been part of pioneering,” El Mistiri says. “In recent years, I have had an opportunity to contribute to various CSEL interventions that are now being evaluated in clinical trials.”

In the process of writing his doctoral dissertation in chemical engineering, he collaborated with Rivera and Eric Hekler, a professor at the University of California San Diego, to conduct multiple National Institutes of Health-funded clinical human trials aimed at behavior change via control engineering approaches. One of their main research projects is called YourMove, a 12-month-long intervention to help inactive individuals 25 years and older walk more to help reduce the risk of various types of cancer and other chronic illnesses.

The team developed a mobile and smartwatch app through which they would track participants’ behaviors and intervene to guide them toward increasing how much they walk.

Mobile health, or mHealth, interventions aren’t a novel approach to behavior change. What makes El Mistiri and his collaborators’ approach distinctly different is that it is rooted in the understanding that people are unique and that context matters.

"Traditionally, in behavioral medicine, people use static models,” El Mistiri says. “Researchers run an experiment for a period of time and just look at the status of participants at specific milestones.”

The biggest issue with this approach is that it doesn’t account for the fact that people’s lives are dynamic and daily situations, or life changes, can significantly affect how they behave. Therefore, most mHealth interventions are not useful for people trying to make physical activity a habit or change another negative behavior, such as smoking.

Using principles of control engineering, El Mistiri and the team developed a framework to automate personalized day-to-day behavioral interventions for everyone involved in the study.

“The Control Optimization Trial, or COT, which is the conceptual basis for YourMove, is the first study of its kind, as far as we know, that has ever been clinically implemented,” El Mistiri says.

The COT framework has three vital stages. In the baseline stage, El Mistiri and the team tracked and observed participants’ physical activity behaviors without intervening in any way. After collecting some preliminary data, they then started deploying behavioral goals, using principles of system identification that are uniquely designed to fit the individual’s situation.

An example of a goal might be walking 7,000 steps today. Additionally, in this stage of the framework, the team designed positive reinforcement mechanisms to encourage the participants to keep making progress. If a person received a goal to walk 7,000 steps and achieved it, they would earn reward points at the end of the day, which could be accumulated and redeemed for health-related products like water bottles.

Most importantly, they utilized control systems principles to automate this process. Using the collected data, they created a model predictive control algorithm, which through the app on participants’ smartwatches, acts as a health coach autonomously guiding them to achieve their physical activity goals throughout the day.

“The distinctiveness of the COT framework is that we do not need to look at the population level. We focus on modeling behaviors for each individual because people are different, context matters and things change,” says El Mistiri. “Our algorithm is run on a cloud server and adjusts goals and rewards based on how each participant responds.”

Developing the COT framework was an enormous task that required the complementary expertise of control systems engineers, data scientists, behavioral scientists and computer scientists from multiple schools, including the University of California San Diego and the University of Michigan.

“I’m really proud of it because the results exceeded my expectations,” says El Mistiri, emphasizing the effectiveness of the framework. “In a related work funded by the National Library of Medicine, our team was able to predict the most suitable time to send intervention notifications for 91% of the participants, which is impressive considering that other studies could not exceed 20%.”

For these projects, El Mistiri had to transcend disciplinary boundaries.

“As an engineer, of course, I didn’t have the behavior science domain knowledge. I worked with the behavioral scientists on the research team to ensure that we were designing effective interventions,” El Mistiri says.

Developing an algorithm that provides participants with goals that fit their physical activity level is not a simple task.

“We couldn’t just give a person a goal to walk 15,000 steps per day when they walk 2,000,” he says. “So we designed the algorithm to deliver ‘ambitious but doable’ goals in a way that allows people to sustain that healthy behavior.”

Reflecting on his collaboration with El Mistiri, Hekler says, “It has been an honor and privilege to work with Mohamed and to both see his development as a top-notch engineer and have the opportunity to learn from and with him now and into the future.”

Hekler adds, “There is no greater joy than seeing a former student mature into a true thought partner and colleague who can bring his own knowledge, skills and perspective into creating new ideas and possibilities. I look forward to continued work and collaborations as he continues to grow into a leader in this emerging field of applying control systems engineering to support human behavior change.”

El Mistiri says his contribution to the YourMove project would not have been as effective without the support of his doctoral advisor, both before and throughout the process. Rivera says he is proud of El Mistiri’s research and contributions to society.

“In 34 years at ASU, I have had the privilege of working with many excellent students. Mohamed El Mistiri stands out among this group for not only his strong engineering and computational abilities but his perseverance, communication skills and demonstrated leadership,” says Rivera. “He has been critical to the success of YourMove and other interventions in the lab and is eminently deserving of the Dean’s Dissertation Award.”

From smoking and drug abuse to phone addiction and other negative habits, El Mistiri’s work has applications beyond getting people to be more physically active. Recently, he took a postdoctoral position at CSEL, where he plans to continue leveraging the power of control engineering, artificial intelligence and machine learning to develop models that discern patterns in a person’s behavior and intervene at the most opportune times.

He says being a part of this project was, above all, a great learning experience and he can’t wait to see it come to fruition.

“I learned a lot of things about clinical trials and scientific collaborations,” he says. “I want to be there on the last day to celebrate, analyze the data and use the findings to guide my next projects.”

El Mistiri also received an inaugural Outstanding Graduate Accomplishment Award from the School for Engineering of Matter, Transport and Energy’s director, Anthony Waas.

El Mistiri says is excited to have received the two awards and for the impact of his research.

“I am honored and very humbled by the recognition,” he says. “Some of the findings in my dissertation have helped in forming the basis for other interventions that our lab is currently running, among them developing weight loss maintenance strategies with the Miriam Hospital. These efforts will be helpful for my research career in the long run.”

Waas emphasizes that supporting students in their research endeavors is critical to their success and discovering innovative solutions to high-priority problems.

“Seeing Mohamed El Mistiri apply his engineering skills to solve societal problems brings me great joy,” Waas says. “I encourage all students to pursue their dreams of impacting society in profound ways, and the school will support them in this quest.”