OROFACIAL MOVEMENT ANALYSIS IN INFANTS AND YOUNG CHILDREN
- NEW OPPORTUNITIES IN SPEECH STUDIES

by Jordan R Green Ph.D. ,
Assistant Professor, Department of Communicative Disorders, University of Wisconsin-Madison, Madison WI 53706, USA

The main research focus for Dr Jordan Green and his team at the University of Wisconsin is oral sensorimotor skill development for speech and feeding. Current research collaborations also involve investigations on the perception of visual speech, oromotor co-ordination in speech disorders, and acoustic-to-articulatory relations. Dr Green teaches graduate courses in dysphagia and the neural processes in speech, language and hearing.

The speech physiology lab is staffed by two doctoral students majoring in speech science (Erin Wilson and Rita Patel) and three masters' students majoring in speech pathology and audiology (Mellanie Belleau, Amanda Melby and Marie Birnbaum), one post-bachelorette in computer science (Allison Mccarthy), and a computer programmer (Dave Wilson). Erin is conducting a kinematic based study on the development of chewing and Rita is working on a kinematic and EMG study of Mobius syndrome, a rare syndrome that impairs facial mobility. Allison is working on driving computer generated facial models with the motion data obtained from a Vicon system.We are assisted in our work by a Vicon 250 system with five 60Hz IR cameras. Its flexibility and accuracy have made it an integral part of our research program, allowing us to overcome prior methodological barriers toward the direct study of naturally occurring oral movements in very young children including newborn infants.

The long-term objective of our research program is to provide a comprehensive account of oromotor development by combining descriptive behaviour analyses with fine-grained quantitative analyses of facial movement patterns. Speaking is among one of the most complex motor acts performed by humans. It requires the co-ordination of over 70 muscles serving the respiratory, laryngeal, and vocal tract systems. Moreover, speech is produced at a remarkably fast rate, approximately fifteen sounds per second. The development of speech production ensues over an extended period and appears to significantly lag the attainment of many associated cognitive/ perceptual capacities. Children typically do not master the sounds of their language until 8 years of age, and some features of speech do not exhibit adult-like consistency until adolescence. The vocal repertoire of young children is limited, in part, because they are not endowed with the sensorimotor control for producing the range of sounds in their language. In early speech, the probability that a sound will be correctly produced depends on the match between existing coordinative capabilities and those required by each sound. Young children challenged by their limited options for producing different sounds are obligated to adopt strategies for approximating adult-like speech. These early articulatory adaptations provide a window into the developmental status of the neuromotor system and cognitive/ perceptual processes. Careful study of these behaviours is providing new insights into the many processes involved in learning to speak.

Unlike other motor systems (e.g., reaching, locomotion), the developmental course of articulation is largely unknown, although it is a matter of fundamental importance for understanding the physiologic basis of both typical and disorder speech development. Research in speech motor development has been slowed by the absence of methods for obtaining physiologic measures of articulation in young children. Consequently, only 10-15 studies have been published in this area over the past decade and most of these studies have been based on older children, who have already obtained many fundamental skills for speech production.
In the absence of physiologic data, researchers have had to formulate their models of articulatory development primarily on acoustic and perceptual observations. Consequently, many of the most fundamental questions concerning the development of speech motor control have yet to be addressed. For example, what are the roles of reflexes and other extant neural circuits in the development of sensorimotor control for speech? What are the sensorimotor milestones of speech? How does the sequence of sensorimotor development influence the sequence of speech sound acquisition? Additionally, very little is known about disorders of speech motor development. In summary, although a great deal of information has been accumulated about the development of speech, much remains to be learned about the underlying physiologic and environmental factors that guide the course of speech development. Further knowledge of these factors is critical for generating biologically plausible models of speech motor development and advancing intervention and treatment programs for speech and feeding impairments.

Recent advances in computer-based motion capture technology, however, afford new research possibilities, offering a means to non-invasively record orofacial movements from very young children. We have recently used these techniques and developed custom analysis routines to study naturally occurring oral movements in young children (Green, Moore, Higashikawa, & Steeve, 2000; Green, Moore & Reilly, 2000) and newborn infants (Green & Wilson, 2002). In collaboration with Christopher Moore's research team at the University of Washington (http://faculty.washington.edu/spchphys/), we successfully recorded upper lip (UL), lower lip (LL), and jaw (J) movements from one-, two-, and six-year-old children and a group of adults. Figure 1 displays an upper lip, lower lip, and jaw trajectory produced by a subject in each age group. These examples illustrate the developmental changes in coordinative organisation that were observed across age groups. Adult subjects uniformly produced these movement sequences, with high levels of movement coupling among the different articulators. In contrast to the adult pattern, 1-year-old children tended to exhibit pronounced jaw displacements accompanied by excessive compression of lip tissues during oral closure. As displayed in Figure 1 (panel One) this compression was associated with oppositional movement (180 degrees out of phase) of the lips and jaw. Thus, closure of the mouth appeared to be primarily achieved by jaw movement at this age. In 2-year-old subjects (Figure 1, panel Two), the upper and lower lip displacements increased relative to those produced by the 1-year-olds, and jaw displacements appeared to decrease. For the 2-year-olds, the upper and lower lip displacement time-histories were often similar in form (e.g., "mirror movements") and frequently were characterized by a single rise-fall sequence extending across both syllables. The displacement patterns of 6-year-olds (Figure 1, panel Six) were similar to those of adults, but were more variable.

Figure 2 presents the results of an analysis that was developed to quantify developmental changes in the proportion that each articulator contributed to closing the mouth for the production of /b/ in "baba." In comparison to adults and older children, 12 month-old subjects relied more on the jaw to produce speech than on the lips. This developmental sequence has been recently corroborated in a follow-up experiment that showed jaw movement patterns to achieve adult-like quality sooner than upper and lower lip movement patterns (Green et al., 2002).
In summary, one result that has emerged from these experiments on early speech development is that co-ordination of the mandible precedes that of the lips. The findings also suggest that the young child has a limited ability to independently control the upper and lower lips. These constraints in early articulatory co-ordination and control have predictable consequences on infants' sound producing capabilities and may explain why children with differing experiences acquire early sounds in a similar order. These findings probably reflect extensive changes in the biomechanical composition of vocal tract structures and their neuromotor pathways. More general principles of skilled movement acquisition may also account for some of the changes observed. Additional studies are needed to reveal the relative importance of each of these factors.
In a current NIH-NIDCD funded project, we are studying the developmental course of orofacial coordination and control from birth to 12 months of age using the Vicon 250. For capturing facial movement, each subject is fitted with a 15 skin-mounted marker array. During data collection, infants are secured in a parent's lap or in an infant seat, while a wide array of facial movements are recorded during spontaneous and purposeful behaviors (i.e., chewing, sucking, speech, vocalizations, facial gestures). Figure 3 shows a four-month-old subject during a facial capture session and the computer generated facial model. EMG is recorded from five targeted facial muscles: orbicularis oris superior, orbicularis oris inferior, anterior digastric, masseter, and temporalis. EMG signals are transduced using miniature fine wire electrodes that have been designed in our laboratory specifically for recording from small anatomic regions. EMG signals are amplified (Grass Instruments, model P511), antialias filtered (2kHz), and digitized using Vicon acquisition hardware (fs = 4000Hz). A full-face video recording of each subject is used to transcribe vocalizations and classify different oromotor behaviours. Analyses are directed toward quantifying the diversity of orofacial movements as well as developmental changes in co-ordinative organisation. The combined information from movement and electromyographic data will provide a means to (a) describe the developmental course of orofacial co-ordination, (b) assess the plausibility of shared co-ordinative infrastructure among existing and emerging orofacial behaviours and (c) describe the rigidity of lip and jaw movement patterns.

Preliminary analyses of these data are beginning to reveal some interesting features of immature orofacial control. At all ages of study, infants exhibited a remarkable quantity and variety of spontaneous jaw movements. Although there appear to be large individual differences in movement characteristics, several task and age effects have been detected. For vocal-related behaviours, movement distance and speed appear to increase with age. At 7 months of age, we have observed a sharp increase in range of motion for spontaneous and vocal-related orofacial movements. It is interesting to note that these changes are occurring during the canonical babbling stage of early vocal development (Oller, 1980; Stark, 1980) and parallel previous reports of the emergence of stereotypes in the hands and arms at this age (Ejiri, 1998; Thelen, 1979).
Once completed, this description will provide much-needed information about the "initial" conditions from which speech motor control emerges and the features of immature orofacial control that shape early speech. The long-term objective of this work is to provide empirically based guidelines for evaluating the developmental status of the orofacial system in children with suspected speech or feeding delays, and for advancing underlying movement competencies in children with speech motor delays.

References
Green, J.R., Moore, C.A., Ruark, J.L., Rodda, P.R., Morvee, W. & VanWitzenburg, M. (1997). Development of Chewing in Children from 12 to 48 Months: a longitudinal study of EMG patterns. Journal of Neurophysiology, 77, 2704-16.
Green, J.R., Moore, C.A., Higashikawa, M., & Steeve, R.W. (2000). The physiologic development of speech motor control: lip and jaw coordination. Journal of Speech, Language, and Hearing Research, 43, 239-56.
Green, J.R., Moore, C.A., & Reilly, K.J. (2002). The sequential development of jaw and lip control for speech. Journal of Speech, Language, and Hearing Research, 45, 66-79.
Green, J.R. & Wilson, E. A kinematic description of oromotor behaviours during the first year of life. Presentation given at Conference on Motor Speech, Williamsburg, VA, 2002.