The marvels of mathematics may open the door to a new, improved hearing test for newborns.
A mathematician and a recent electrical engineering doctoral graduate from Washington University in St. Louis have devised a hearing test that measures the auditory brainstem response (ABR) 20 times faster than current methodology. The technique allows for testing on small digital machines that takes just two minutes instead of the hour current methods take, and volunteers instead of medical personnel can administer it.
“This would be a hand-held device that gives a pass-fail for hearing — red light, green light,” explained Victor Wickerhauser, Ph.D., Washington University professor of mathematics in Arts & Sciences. “It can be used by very inexpensive employees instead of specialists. It could make universal screening for hearing defects in babies very practical.”
Wickerhauser said that the National Institutes of Health has mandated that all newborn babies be screened by ABR testing within the next three years. The test is critical, he said, because an estimated four out of every 1,000 newborns show some kind of hearing loss. While most problems can be amended with a hearing device, some are more severe. If problems are not addressed within the first two years of a child’s life, the child’s cognitive abilities are impaired, which has a big impact on society as well as the individual child.
The auditory brain stem response is a physiological response that can be evoked by a click in the ear. Any living person that can hear even slightly will develop an electrical response in the brain stem between 50 and 100 milliseconds after the click is administered. The problem is this response is very weak, especially compared with competing electrical activity, such as surface muscle activity. Such nonessential, nuisance activity scientists call “noise.”
Wickerhauser and Elvir Causevic, now a researcher at Everest Co. in St. Louis, used a mathematical tool called wavelet analysis; simply put, this is a signal processing technique that extracts signals from “noise,” or unwanted data.
In the 1950s and ’60s, diagnosticians performed tests with giant electroencephalogram (EEG) machines that worked in expensive acoustic isolation rooms; the operators had to repeat clicks, record and average lots of data to see if the response was there. These tests took an hour and were completely impractical for newborns.
Today, little digital EEG machines record signals and average 10,000 repetitions. It still takes up to an hour, but Wickerhauser and Causevic have a process that has sped up the averaging of the data. Their method gives a pass-fail grade based on just 500 repetitions. Instead of an hour, it takes two minutes.
Wickerhauser and Causevic used wavelet analysis, a sophisticated kind of harmonic analysis that is integral in analyzing and compressing data — video, sound or photographic, for instance — for a wide range of applications.
Wickerhauser is an internationally renowned specialist in wavelets. Causevic came to Washington University in part to tackle problems in hearing and sought out Wickerhauser for his expertise.
The two are co-inventors of the technique and have applied for a patent.