Cleft palate, which is caused due to a split in the roof of the mouth, can cause problems in speaking and eating, in addition to development, learning, and social challenges in children. This condition can be corrected with surgery, but may require a second surgery if the child affected has continued speech problems.
“A lot of these surgeries are based on anatomy textbooks. However, there are potential variations in children based on the individual’s anatomy, gender, and race. These [factors] are not usually considered,” said Brad Sutton, professor of bioengineering and technical director of the Beckman Institute’s Biomedical Imaging Center. “We are hoping to provide an MRI technique that will be able to visualize the muscle movements associated with a particular patient.”
The HHS-funded project “Using atlas-driven imaging for determining variations in velopharyngeal function among children with cleft palate and hypernasal speech” will receive $2 million over four years to investigate imaging techniques for cleft palate patients.
One of the problems with current MRI techniques is that the imaging process is slow. The patients need to lie still for three to four minutes just to get a single image, which can be difficult for children. In the last few years, Sutton along with Zhi-Pei Liang, professor of electrical and computer engineering, have been trying to speed up the imaging process.
“In 2010, we could capture 20 frames per second by just looking at a single part of the speech process,” Sutton said. “Now we can take multiple 3D images, with high resolution, at 166 frames per second.”
The increased resolution allows the researchers to visualize how the velum (which is a soft tissue at the back of the roof of the mouth), lips, and tongue move during particular phrases of speech. However, this high-resolution technique still requires 10 minutes to complete.
“For this grant, we are collaborating with Massachusetts General Hospital that takes atlases into account,” Sutton said. An atlas averages the 4D movies (3D plus time) of how a group of people say a particular phrase. It represents how the muscles move, on average, during a certain speech sample.
The researchers will look at how a group of people, who do not have a history of cleft palate, move their speech muscles during certain phrases. Researchers will use that information to see what parts differ in the patients.
“If our imaging algorithm already knows how the normal subject moves, we can take that into account, and then we only need to sample enough information to see how the patient is different from those motions,” Sutton said. “We don’t need to sample as much information and that’s how we are planning to speed up the process.”
The atlases will be used in cleft palate patients in clinics in collaboration with Jamie Perry from East Carolina University and former Beckman graduate student. “Dr. Perry looks at what types of surgical interventions are helpful for cleft palate patients and how information from imaging can inform surgical planning,” Sutton said.
The grant will also help the researchers develop atlases that take gender and race into consideration. The effort will be aided by Ryan Shosted, associate professor of linguistics, and David Kuehn, professor emeritus of speech and hearing.
“We are excited because a lot of people want to use the imaging technology we have already developed,” Sutton said. “We are now trying to create new tools that should run on any MRI scanner.”