Proposal

Surface Electromyography:

Change in Frequency Content with Muscle Fatigue

Investigator:	February 16, 2002	Project Advisor:
Eric Alan Beltt	February 16, 2002	Professor William K. Durfee

Project Summary:

The objective of this experiment is to confirm that the frequency spectrum of an EMG signal shifts to lower frequencies as a muscle, under voluntary contraction, fatigues. A second objective is to gain engineering experience through the design and construction of the EMG apparatus. The proposed project will consist of three distinct stages: design, programming, and measurement.

The first stage of the experiment will be the resolution of the relatively small EMG signal from background noise, which is not a trivial task. Several things need to be considered in order to build a working device. First, the device must be constructed so that internal noise does not obscure the signal. Second, the electrodes must be designed to minimize ambient noise, including signals from other parts of the body. Lastly, frequencies outside the usable range of the EMG signal, 20-500 Hz, need to be filtered out.

The second stage of the experiment will consist of designing a computer program to collect data in the time domain, transform it to the frequency domain, and analyze it. The programming stage could be simple or more involved, depending on whether I choose to learn MATLAB, or use a data acquisition program I am already familiar with, such as LabWindows or LabVIEW. This is an important aspect of the project because it will allow me to tailor it to the time constraints as I go along.

The last stage of the experiment, which will probably be the simplest, will be the actual measurement. It will consist of a subject holding an appropriately sized barbell so that the bicep is fatigued in approximately 2 minutes. The subject will hold the weight with forearm parallel to the ground, and elbow at 90°. I plan to take ½ second epochs of data from the bicep every second. The measurement is another place in the project that can be tailored to the time constraints because the number of subjects can be varied. In the end, I expect to get a single graph with several data-series that each show the mean frequency content of the EMG signal from flexion to fatigue.

Background:

Surface Electromyography is a non-invasive technique for detecting electrical signals given off during muscle contraction. A typical EMG apparatus consists of an electrode stage, an amplification stage, a filtering stage, and a data acquisition system. Resolving the small EMG signal is not a trivial task, so care must be taken in the design of the apparatus. There are several different sources of noise that obscure the signal. The first source of noise is simply inherent in all electrical devices. It can be minimized by careful circuit design. The second source is ambient noise. Power lines are one source, as are the many different electrical signals traveling through the human body. Differential amplification and careful electrode design are the primary ways to reduce ambient noise. The last source of noise is due to the quasi-random nature of the EMG signal, which means that only certain parts of the signal are useful. Specifically, the usable section of the EMG signal is that in the range of 20-500Hz. Frequencies outside this range need to be filtered out. For more details, consult the white-papers at http://www.delsys.com/library/tutorials.htm.

It has been shown that the frequency content of an EMG signal shifts to lower frequencies as fatigue occurs. In the field of Surface Electromyography, however, that fact is not quite common knowledge. This makes it an ideal student research topic and confirming this fact will be the objective of this experiment. The primary focus of the project, though, will be to gain engineering experience through the design and construction of the EMG apparatus.

Experimental Program:

This experiment will be conducted in three distinct stages. The first stage, which will be the most involved, will be the design and construction of the EMG apparatus. The second stage will be programming the software I will use to collect data. The final stage, which I expect to be the easiest, will be to actually sample an EMG signal and plot the mean frequency as a function of time.

The design stage of the experiment will be the most complex because of the reasons listed in the background section. Resolving the EMG signal is not a trivial task. I do believe, however, that it is within the realm of reason to expect that I will be successful. Most of the design process will draw upon skills that I acquired last semester, including analog filtering, differential amplification, and noise reduction. The device I plan to build will follow the basic layout of figure 1, shown in solid lines.

I expect the cost of the components I will need to total less than $10.00, with one exception. I plan to have printed circuit boards made in order to reduce electrical noise. These will cost $59.00 for a set of three, which I will order from ExpressPCB (http://www.expresspcb.com).

I plan to use an existing computer data acquisition system for the programming stage of the experiment. The complexity of this stage depends on whether I choose learn MATLAB, or I use a data acquisition program I am already familiar with, such as LabWindows or LabVIEW. I have left this undecided because it allows me the flexibility to tailor this experiment to fit the time constraints. Regardless of what platform I choose to work with, the task will be the same. I will need to design a program to acquire data in the time domain, transform it to the frequency domain, and calculate the mean frequency.

The final stage of the experiment will be to actually measure an EMG signal. I plan to conduct the experiment by having a subject hold a barbell with forearm parallel to the ground, and elbow at 90°, until the bicep is fatigued. I will take ½ second epochs of data every second, and the barbell will be sized so that the experiment lasts approximately two minutes. Then I will find the mean frequency for each epoch of data, and plot them on a single graph. I expect the graph to show that the mean frequency decreases with time, and I expect it to look similar to figure 2 (Ref. 3). Note that this is data for dynamic muscle contractions, where I will be studying isometric muscle contractions.