App designed to predict aggressive behavior

By Catherine Robertson Souter
May 1st, 2017

In 2014, a group of parents in California sued a neighbor over the problem behaviors of their autistic child.

According to one report, the plaintiffs claimed that they were “not upset about him being autistic” but about his violence towards other children, claiming it made the neighborhood unsafe and even affected home sales.

A lawsuit may not be a common reaction to autism, but any parent with a child who exhibits violent or aggressive behaviors can understand what the parents of the autistic boy must have felt.

While experts say that aggressive behaviors are not a part of autism, but rather a by-product for a small minority, they change everything for those children and the adults who care for them.

“Besides the immediate danger,” said Matthew Siegel, M.D., director of the developmental disorders program at Spring Harbor Hospital in Westbrook, Maine, “the real problem is that aggression can change the trajectory of a child’s or adult’s life. If you are aggressive and unpredictable you get excluded from school, community and social settings, things that you need more than anyone else as an individual with autism.”

The main issue with these behaviors is that there are often no warning signs before an episode, which puts the person and caregivers, and anyone in the vicinity, in physical danger. The question then becomes, wouldn’t it be great to have a way to predict the behavior before it starts?

The answer first occurred to Matthew Goodwin, Ph.D., assistant professor of health informatics at Northeastern University, more than 20 years ago when heart rate monitors were all the rage with sports enthusiasts.

He began tracking cardiovascular arousal in non-verbal and intellectually impaired children and adolescents with autism hoping to link the physiological signals with behavior. Starting in 2008, he began working with Rosalind Picard, Sc.D., director of computing research at the MIT Media Lab to develop a device to track heart rate along with skin temperature and electrodermal activity to give a clearer picture of stress.

Goodwin then brought in Julie Lockner, an MIT graduate student and co-founder of 17 Minds, a startup software company, to help design an app that will allow caregivers to track behavior in real time.

Locker, who describes herself as an “inspired parent” of a son with non-verbal learning disorder enrolled in grad school at MIT’s Sloane School of Management to get involved in the project.

“We started looking at wearable technology,” she said. “And, as electrical engineers, my husband and I want to help solve these problems so I went back to school and 17 Minds is the corporation that I incubated while in the MBA program at Sloane.”

The overlay of the observable behaviors with the biological data should provide the ability to learn an individual’s particular physiological signals when in a neutral state and when in an aroused state and to ultimately predict the behavior.

The technology Goodwin developed, now available through a company called Empatica, and the app developed by 17 Minds are part of a research study with Spring Harbor Hospital and a team at University of Pittsburgh to gather data with adolescents.

“We are developing a different approach to aggression in autism,” said Siegel “We are working to shift the paradigm. All of the approaches we use at present are reactive. So, we are trying to manipulate this technology to where we can get a signal that there is a high risk of serious behavioral problems occurring.”

The team is working on both the research and a commercially available product at the same time. A useable product could be brought to market for early adapters by fall. Continued feedback and research will help to further advance the usefulness of the product and, with increased functionality will come an increase in marketability and an expected reduction in the cost of the $1,700 wearable. The timeline may seem rushed but, for Goodwin, it has been a painfully slow process.

“We’ve been at this a long time,” he said. “I’m frustrated because we could have had this 20 years ago but it has taken time to test, get the data, develop the algorithms, etc. And it is not that we didn’t have the technologies or the people. It’s all about resources. The project has been going in fits and starts because of funding.”

Posted in Articles, Leading Stories | Comments Off on App designed to predict aggressive behavior

Comments are closed.

Powered By MemberPress WooCommerce Plus Integration