Biology Becomes Electric
Gauging Muscle Dysfunction
Researchers in the NeuroMuscular Research Center (NMRC) have developed an electronic device that measures electrical signals from contracting muscle and may also help to diagnose diseases early. Known as an electromyographic (EMG) signal detector, it enables researchers and clinicians to assess the extent of neuromuscular injury or disability and monitor the progress of rehabilitative therapies by tracking changes in muscle fiber control. The EMG detector acts as an early warning system that may ultimately allow physicians to slow or halt diseases of the central nervous system, such as ALS and Parkinson's, before overt symptoms arise.
A visualization of the code used by the brain to control muscle fibers. Each train of impulses represents sequential contractions of a particular muscle fiber during a contraction that generated the force profile of the white trace. The electrical pulses detected within the muscle to characterize each train are shown on the left.
That's the vision of biomedical engineer Carlo De Luca, who directs the NMRC and heads a 14-year-old spin-off company called Delsys. To realize this vision, De Luca and his NMRC colleagues—physical therapist Serge H. Roy and electrical engineer S. Hamid Nawab, along with four graduate students—are working to improve the accuracy of EMG signal processing algorithms for both clinical and research applications.
The EMG signal consists of a superposition of electrical pulses which propagate along the muscle fiber every time it contracts. As more muscle fibers are stimulated by the brain, more force is generated and the amplitude of the EMG signal increases. “EMG signals have small amplitude but radiate and reach the surface of the muscle,” says De Luca. “If you apply special sensors on or below the surface of the skin, it's possible to detect the sum of all these pulses and to determine whether the muscle is firing in a normal or abnormal fashion, but it's a noisy, complex signal.”
The firing behavior of healthy muscle fibers follows well-established rules, some of which were discovered by researchers at the NMRC. In several investigations, De Luca has shown that firing behavior remains unchanged even during fatiguing tasks or extended bed rest, but is altered in elderly subjects, acute cerebellar stroke patients, and returning Space Shuttle astronauts.
Biomedical engineer Carlo De Luca inserts a special needle sensor containing four electrodes near the tip. The sensor, which was developed to test astronauts returning from Space Shuttle missions, detects an electromyographic (EMG) signal, which is then processed with a complex artificial intelligence algorithm to yield the data shown above.
For the past three decades, he and his colleagues have developed and refined two types of sensors and algorithms to decompose the EMG signal into individual pulses, which reflect how individual muscle fibers are controlled within the muscle to perform particular functions. Intended for clinical purposes, the first sensor is attached to a needle that gets inserted into muscle, and has achieved 85 percent accuracy. The second sensor, developed by Delsys, is used for motor control studies. It works on the surface of the skin and has achieved 65 percent accuracy.
Over the next five years, De Luca aims to boost those figures to 95 percent and 85 percent, respectively. “We have shown this level of accuracy is sufficient to provide a good estimate of parameters needed to extract information that describes how the muscle is controlled,” says De Luca.
Developing and commercializing De Luca's NMRC research, Delsys provides electromyography equipment, physiological sensors, and data analysis programs for educational, research, and medical markets across the globe. The company serves clients in 55 countries and attributes its rapid growth to funding from the Small Business Innovation Research program, sponsored by the Small Business Administration (SBA). Delsys received the SBA Tibbetts Award in 2006.
For more information, see http://nmrc.bu.edu.
