SOUNDS IMPORTANT:
A blanket approach to helping babies. By Linda J. Ferrier and Harriet J. Fell
Compared to other animals, human babies look incredibly vulnerable, yet are born with so many amazing skills, not least of which is their ability to tune in the world of sound immediately after birth. Sixteen years ago we, Linda Ferrier and Harriet Fell, met as a result of a Children's Hospital research study on the language development of normal and premature infants. At the time, Linda was part of the research team and Harriet had just delivered her daughter, Tova, at the Boston Lying-in Hospital. Tova was enlisted as a subject. There began our joint interest in the development of speech and communication in children. Linda visited Tova and Harriet every month for three years and collected samples of Tova's prespeech babbles and other behaviors.
After three years, the study ended. Later, Harriet, a professor of computer science at Northeastern, happened to wander past an office door that showed a computer-generated graphic of a cup of tea and a familiar name. Linda, by then, had joined N.U.'s department of speech-language pathology and audiology. We renewed our friendship over lunch at the Faculty Center and soon began collaborating on interdisciplinary projects, developing software for children with severe motor problems. We did some exciting work devising educational software that could be used with different interfaces such as a joystick, a keyboard, or voice. While the design and development went on in computer science software design classes, speech-pathology students field-tested the software with children at the Kennedy and Massachusetts Hospital Schools. Both groups of N.U. students made initial visits to the classrooms to ensure that the systems would meet real needs.
We felt then that some of the children using our systems could have benefited from earlier intervention, but neither suitable input devices nor software for very young children existed at the time. We turned our attention to designing an interface and communication system for infants with severe motor disabilities.
Infants, virtually as soon as they are born, are more interested in speech than other sounds. It appears that they come "prewired" to be language users. They are also able to discriminate among the sounds of all the different languages of the world, an ability which disappears by the end of the first year as their brains become neurologically dedicated to distinguishing the sounds of their native language.
At birth, the infant vocal tract is similar to that of other primates. The tongue is farther forward in the mouth than in adults, and the neck is relatively short. (Look at a baby or two-they don't have much neck to start with.) At this point, there are structural constraints on the infant's ability to make speech sounds. As the neck grows in length, the larynx and the tongue base descend, and the infant is able to use the oral space as a resonator. In their first three months, babies make mostly short vowel sounds. By six months they are beginning to explore consonants, and by the end of the first year they are making sounds composed of consonant-vowel sequences such as "mama" and "dada" that sound like adult words. In doing this, infants are learning important discriminatory, linguistic, and motor skills. They can amuse themselves by exploring the different sounds the vocal tract is capable of producing. Parents eagerly join in these early vocal games by imitating their infants' sounds. By twelve months, most infants begin to use true words that mean something.
Unfortunately, some infants, such as those with severe cerebral palsy, are incapable of making these early sounds-even though they may have normal intelligence. These infants are deprived of what may be a critical stage in language development and perhaps important knowledge of the sound structure of the language. They are also left without an important tool for initiating social interaction. We sought to use technology to provide these infants a means of using whatever motor capabilities they have to produce and explore sound on their own. Infants incapable of vocalizing might activate recorded or electronic babbling sounds, as speaking infants do. They might also listen to music or environmental sounds such as a toilet flushing-a favorite sound of young children! We hoped they would eventually learn to control their body movements to activate switches, for communication, to affect their environment, to make choices, and, later, to produce words.
We needed sound and speech output and a soft, waterproof, noninvasive input device-some sort of pad or blanket that was sensitive to a child's movement. The system had to be simple to assemble and use if it was to have a place in the busy classroom or home. As scientists, we also wanted our system to collect data to document infants' learning. As we both like to sew, we considered constructing a pad ourselves, using our quilting skills to encase a collection of large pressure-sensitive switches. We learned about commercially available components from colleagues, however. We settled on using a Nintendo Power Pad (sold with a video game for teenagers) and an adapter (from Transfinite Systems) designed to transfer signals from the Nintendo device to a Macintosh computer. With the help of seed money from the Northeastern Research Scholarship Development Fund, we put together our prototype system and named it the "Baby Babble Blanket."
To prevent immobile children becoming bored with access to only one or two sounds, we included a total of twelve sounds, one of which played at random each time a switch was activated. We tested this product at the Boston College Campus School, which provides services to children with multiple handicaps. We began by selecting three developmentally delayed children who appeared to be constantly active: one rolled from side to side, one kicked constantly, and one flopped on his back and then sat up again. We entered sounds into the device that the teachers thought would be interesting to the children. At that time, our system was quite rudimentary and accepted only one sound per switch. We hypothesized that the children would increase their activity levels to hear the sounds. Interestingly, all three children reduced their levels of movement when placed on the blanket. We assumed that attention to sounds decreased their levels of stereotypic movements. Other children, whose normal activity levels were lower, showed an increase in switch activations when placed on the blanket.
With these encouraging results, we applied for funding from the federal Department of Education and were awarded two grants to test the prototype, and then to develop the system further and carry out more extensive field-testing. Again we used our working model of engaging computer science students to develop the code and speech-language pathology students to carry out field-testing in homes and classrooms. A great joy was the satisfaction our students got from working on real-life projects in a team format. We developed libraries of digitized sounds from which teachers and parents could select, as well as the capability of being able to record voices on the spot-everyday routines that parents and teachers say to children. We included early- and late-stage infant babbles and early words that most children use. The device was extended to activate battery-operated toys. We increased the sophistication of the data analysis to include easy-to-read graphs that teachers could interpret quickly.
We found that one five-month-old child, with hydrocephaly, club feet, and poor muscle tone, was able to switch from rolling his head in order to hear a recording of his mother's voice to raising his legs to do so-demonstrating an awareness of cause and effect. Another five-month-old, blind but with normal motor control, smiled, vocalized, and increased his level of activity when the blanket's sound was turned on, but cried when the sound was turned off-clear proof that he enjoyed it!
Our prototype had deficiencies, however. There were large spaces between switches where small babies might move to no effect. The switches in the Nintendo pad needed at least ten pounds of pressure to be activated; some of our infants were too light to consistently trigger them. Finally, teachers who wished to use the device in different locations needed a more portable system with fewer wires to come unplugged or for children to get tangled up in. We began to look for collaborators to help us with a new streamlined version. A serendipitous encounter with the president of Meeting the Challenge, an electronics manufacturer in Colorado Springs, Colorado (telephone 800-864-4264, e-mail <snd-explorer@mtc-inc.com>) moved us toward commercialization of the Baby Babble Blanket. Meeting the Challenge already had expertise in the manufacture of sound recording/producing technology in the form of "Pocket Coach," a device that prompts mentally retarded adults through a sequenced vocational task. Meeting the Challenge is now in the final stages of putting together a new, portable, more sensitive version of our product-renamed Sound Explorer. It became commercially available last month, at a cost of $350.
We are still interested in baby babbles, but are now taking another tack. Our new product is an Early Vocalization Analyzer, a computer system that automatically analyzes recordings of infants. We hope it will act as a screening test to warn health professionals when an infant may be at risk for later communication problems. The next time you walk past a baby in a supermarket, analyze those interesting babbling noises yourself. They represent a major leap forward in development and an interesting window into the young child's abilities.
Linda Ferrier is an associate professor of speech-language pathology in Bouvé College of Pharmacy and Health Sciences. Harriet Fell is a professor in the College of Computer Science.
