Jamie Perry perryja@ecu.edu

East Carolina University

College of Allied Health Sciences
Dept. of Communication Sciences and Disorders
East Carolina University
Greenville, NC 27834
(252) 744-6144

Presented Monday morning, June 26th, 2017
As part of a speaker panel session, “New trends in visualizing speech production”
173rd ASA Meeting, Boston

Cleft lip and palate is the most prevalent birth defect in the United States. Despite advances in surgery, 25-37% of children with a repaired cleft palate continue to have nasal sounding speech and require multiple surgeries (Bicknell et al., 2002; Lithovius et al., 2013). This relatively high failure rate has remained unchanged over the past 15 years.

A critical barrier to understanding surgical outcomes and decreasing failure rates is the lack of imaging studies that can be used on young children to understand the underlying anatomy. Current imaging techniques used to study cleft palate speech use either radiation (e.g., x-ray or computed tomography), or are considered invasive (e.g., nasopharyngoscopy). None of these traditional imaging methods provide a view of the primary muscles needed to have normal sounding resonance.

Our research laboratory from East Carolina University (Greenville, NC) has been working with a team, including Bradley Sutton and David Kuehn at the University of Illinois at Urbana-Champaign, to establish an imaging tool that can be used to examine the underlying anatomy in a child with cleft palate.

With the support of a team of experts in cleft palate and bioimaging, we described a method for obtaining dynamic magnetic resonance images (MRI) of children during speech. Using dynamic MRI, we are now able to view the muscles inside the speech mechanism. Figure 1 shows images along the sequence of the dynamic images. Images are obtained at 120 frames per second and allow investigators to study a three-dimensional dataset while simultaneously capturing speech recordings (Fu et al., 2015, Fu et al., 2017). With a leading expert in computational modeling from the University of Virginia, Silvia Blemker, we have been able to build a model that can simulate the anatomy in cleft palate. We are then able to study how surgical techniques impact speech.


Fig. 1

Specifically, we used computational modeling (Inouye et al., 2015) to simulate function of the mechanism for producing normal resonance, called the velopharyngeal mechanism. In 2015, Inouye and colleagues used this computational model to predict how much levator veli palatini muscle overlap is needed to produce normal function.

Using these and other types of computational models, we can predict outcomes based on surgery techniques. Through these series of investigations, we are able to advance our understanding of speech in children with cleft palate and to find ways to improve surgical outcomes.



Bicknell S, McFadden LR, Curran JB. Frequency of pharyngoplasty after primary repair of cleft palate. J Can Dent Assoc. 2002;68(11):688-692.

Fu M, Barlaz MS, Holtrop JL, Perry JL, Kuehn DP, Shosted RK, Liang Z, Sutton BP. High-resolution full-vocal-tract 3D dynamic speech imaging. Magn Reson Med. 2017;77:1619-1629. Doi: 10.1002/mrm.26248. PMID: 27099178.

Fu M, Bo Z, Shosted RK, Perry JL, Kuehn DP, Liang Z, Sutton BP. High-resolution dynamic speech imaging with joint low-rank and sparsity constraints. Magn Reson Med. 2015;73:1820-1832.

Inouye JM, Perry JL, Pelland CM, Lin KY, Borowitz KC, Blemker SS (2015). A computational model quantifies the effect of anatomical parameters on velopharyngeal function. J Speech Lang Hear Res. 58;1119: doi: 10.1044

Lithovius RH, Ylikontiola LP, Sandor GK. Frequency of pharyngoplasty after primary repair of cleft palate in northern finland. Oral Surg Oral Med Oral Pathol Oral Radiol. 2014;117(4):430-434. doi: 10.1016/j.oooo.2013.12.409.

Share This