Researchers combine robotics, machine learning and brain imaging to assist infants with CP

University of OklahomaNov 2 2012

Learning to crawl comes naturally for most infants, but those with cerebral palsy lack the muscle strength and coordination to perform the 25 individual movements required for crawling. With a $1.135 million, three-year grant from the National Science Foundation's National Robotics Initiative, University of Oklahoma researchers from the Norman and Health Sciences Center campuses are combining robotics, machine learning and brain imaging to assist infants with CP with the challenging, life-altering skill.

"Because infants with CP are unable to reliably perform the individual movements that make up crawling behavior, they learn to stop trying instead of continuing to practice these movements," said Project Leader Andrew Fagg, associate professor in the OU departments of Computer Science and Bioengineering and project leader. "This substantially delays their development of skilled crawling. In turn, cognitive development and other areas of development are delayed because they both rely on the infants being able to explore their surrounding world."

"In our previous study, we were able to capture many of the infant's actions and had a robot that could assist some of the infant's attempts at crawling. These assists serve as rewards that encourage continued practice of specific limb movements. This grant will allow us, among other things, to develop new robot platforms that can allow a greater range of infant mobility" said David Miller, professor in the OU departments of Aerospace and Mechanical Engineering and Bioengineering. "In the latter part of this grant, we will also start working with the transition from crawling to walking."

"This grant is also important because it builds on and expands our previous work that maximizes the interaction of robotics with what an infant can do," said Thubi Kolobe, professor of rehabilitative sciences at the OU Health Sciences Center College of Allied Health. "Infant learning is integral, and when infants stop trying, parts of the brain responsible for the skill are negatively affected. The next step of this research is to increase the level of help that infants with or at risk for CP are getting. We are looking for combinations of assists that result in the best incentives for these infants. We also want to see if there is a connection between what the infants are learning and what is happening in the brain."

Lei Ding, assistant professor in the OU departments of Electrical Engineering and Bioengineering, will then perform brain scans using electroencephalograph to determine how the infants' brains respond when they are assisted by the robotic device. The EEG technology will assess brain activity of infants during crawling and provide information about changes that occur because of robotics assists and infant efforts.

"Beginning in spring 2013, we will conduct clinical trials to test six infants without CP on the new crawling robot," says Kolobe. "Then, one year later, we will conduct clinical trials to test 24 CP infants on the crawling robot. Initial tests on standing and walking with infants without CP will be conducted by the end of the project. No CP infants will be tested on standing and walking in this grant, only healthy infants."

"This is groundbreaking research, and no one else in the world is doing it," says Kolobe. "We want to invite anyone with an infant who is at risk for CP or severe developmental delays, between four and eight months old, who is interested in participating in these clinical trials to contact Dr. Thubi H.A. Kolobe, at 405-271-2131 ext. 47121 or [email protected]."