Unless the test subject's shoulder injury "clunks," bio-engineers Dr. Patrick Atkinson and Robert Kargus don't hear much when they "listen" to the body in motion as part of their research using wavelet analysis, a form of broad signal analysis.

Atkinson, associate professor of Mechanical Engineering, and Kargus, a graduate student in Mechanical Engineering from St. Louis, Mo., are using wavelet analysis to "listen" to sounds made by the human shoulder in test subjects who have volunteered to the let the pair eavesdrop using digital technology.

Their goal is to establish a data base of digital signals that identify wear and tear on cartilage in the shoulder for use in medical diagnoses. Atkinson and Kargus are using very sensitive accelerometers to analyze the sound coming out of the shoulder joint.

"Wavelet analysis involves analyzing short wavelengths of electrical signals coming from the accelerometers. In simple terms what we come up with is a 'squiggle' on the screen or on a print out," said Atkinson, "and we have to determine what that squiggle means."

Image removed. Image removed.

The idea is that the squiggle for one injury will look different from another type of injury or from a non-injured shoulder, according to Atkinson. "We have to train the computer to identify the various squiggle patterns and what they represent," he said.

"It's based on a simple concept," said Atkinson, "that sound can indicate whether something is functioning correctly or whether something is wrong." Atkinson likened listening to the sounds of the body, to a mechanic listening to the sounds a machine makes.

"Things that move, like motors, sound differently based on what they are doing," he said. "In industry, manufacturers can analyze the sound a machine is making to determine how far the machine has progressed in its life cycle and predict when it might experience failure - all based on the sounds the machine makes."

Being able to predict the need for maintenance helps extend the life cycle of machinery," said Atkinson, "it's all about timing." "The body has moving parts just like a machine - the heart, lungs, bones - as a first cut, if we apply the sound test to bones or joints, then we can predict when they are close to failure and replace things like hips and knees before they break or degenerate too far as occurs in arthritis."

"Being able to 'listen' to joints like a shoulder, hip, or knee would also help identify when cartilage is wearing thin so a person could reduce use and delay having joint replacement surgery," he said. According to Atkinson, cartilage is the body's way of allowing two bones to rub against one another without pain.

Despite being dense and hard, bones contain nerves and can feel pain. Cartilage does not have nerve endings, it is the body's way of buffering two bones in knee and hip joints. When cartilage wears out in a joint area, bones rub against one another and become painful. "The maxim in bioengineering is that the body is the perfectly optimized machine - it only puts tissues where they are needed," said Atkinson, referring to cartilage.

Atkinson and Kargus selected shoulder joints for their study because it is typically difficult for surgeons to diagnose problems in the shoulder. "The shoulder is an unstable joint, it is not a complete ball and socket joint like the hip," said Atkinson. Instead of a cupping socket similar to the hip joint, the shoulder "socket" is more like a golf ball resting on a tee.

Image removed. (shoulder art courtesy of the Nicholas Institute of Sports Medicine and Athletic Trauma)

The American Academy of Orthopaedic Surgeons (AAOS) web site describes the shoulder as "a ball-and-socket joint that enables you to raise, twist, bend, and move your arms forward, to the sides and behind you. The head of the upper arm bone (humerus) is the ball and a circular depression (glenoid) in the shoulder bone (scapula) is the socket. A soft-tissue rim (labrum) surrounds and deepens the socket. The head of the upper arm bone is coated with a smooth, durable covering (articular cartilage) and the joint has a thin, inner lining (synovium) for smooth movement. The surrounding muscles and tendons provide stability and support.

Kargus and Atkinson have recruited 20 Kettering students, between the ages of 18 and 24 with no history of shoulder injury or extensive overhand throwing, to establish benchmark sounds for the shoulder when reaching high, behind and in front of the test subject.

During the test an orthopaedic surgeon moved the shoulder joint of each subject using motions known to illicit pain in persons with shoulder injuries. None of the first test group experienced pain, because they were injury free, thus providing a benchmark of pain free range of motion, according to Atkinson. "People normally raise their hands over their heads for only two reasons that are non-sports related," said Atkinson, "for hygiene purposes and to reach for something."

"The subjects in the first group showed no noise on the wavelet analysis, providing data on the sounds of pain free shoulder movement," he said. "There tends to be a 'clunk' if there has been an injury."

The first test group is only one part of the overall research project. Other groups will include individuals with previously diagnosed injuries such as a torn rotator cuff and those who have already had shoulder surgery.

Image removed. Starting with a group that has no history of injury or extensive use of the shoulder joint is part of the medical protocol approved by the InstitutionalReview Board at McLaren Regional Medical Center. Such Boards are designed to ensure that the volunteers' participation is indeed voluntary and the study has scientific merit.

"For the final test group we will 'listen' to the shoulder noises of individuals with undiagnosed shoulder injuries, and using wavelet analysis data, we'll allow the computer to determine what the injury is, based on the 'noise' the shoulder makes," said Atkinson. The computer's analysis will be compared to an MRI to correlate the data and confirm the diagnosis.

Their hope is that the study will establish an accurate and non-invasive method of diagnosing shoulder injuries for use by orthopaedic surgeons. In the meantime, they continue to eavesdrop on the shoulder noises of willing test subjects, building the data base one (un)injured joint at a time.

For more information on shoulder anatomy and injuries, visit the American Academy of Orthopaedic Surgeons (AAOS) web site. For more information on the wavelet analysis research on shoulders at Kettering University, contact Dr. Patrick Atkinson at patkinso@kettering.edu.

Written by Dawn Hibbard
(810) 762-9865
dhibbard@kettering.edu