View Full Version : SUMMARY: EMG electrodes: sizes, shapes and frequency responses

Oyvind Stavdahl
05-07-1995, 10:53 PM

2. April I posted a request for information on the relationships
between surface EMG electrode geometry and resulting frequency

Enclosed you will find my original request (rather long, unfortunately)
and an edited summary of the resonses I received. Thanks to Dr. Dan
Antonelli and Dr. Peter Kyberd who contributed!
(Peter Kyberd's contribution was accidentally deleted, and is therefore
absent in the summary below.)

Oyvind Stavdahl

****** ORIGINAL REQUEST ******

Hello, BIOMCH-L!

I'd like to add the following information - and question - to the
recent EMG electrode discussion/info. exchange:

We just bought a couple of preamplifier EMG electrodes (MYO 115 - EMG
Research Electrodes) from Liberty:

Liberty Mutual Research Center
Prosthetics Group
71 Frankland Road
Hopkinton, MA 01748
Phone: (508) 435-9061
Fax: (508) 435-8369
Orders: (800) 437-0024

and we're about to buy some more in near future if they turn out to be
OK. One of the reasons for our choice was the following publication:

Hogan, N. & Mann, R. W., Myoelectric Signal Processing: Optimal
Estimation Applied to Electromyography - Part II: Experimental
Demonstration of Optimal Myoprocessor Performance. IEEE Transactions on
Biomedical Engineering, Vol. BME-27, No. 7, July 1980.

Here, Hogan and Mann predict and demonstrate that the distance between
the electrodes (that is, a differential electrode pair) influences the
bandwidth of the differential signal, with reduced interelectrode
spacing resulting in a higher signal bandwidth (this will for certain
myoprocessor algorithms, such as root-mean-square, result in higher
signal-to-noise ratio). They also consider the size and shape of each
electrode, which is also shown to influence the spectrum of the
differential EMG observed. It seems to me that Liberty's engineers have
been reading Hogans article....

(A second reason for our choice was the small size of the
electrode/preamp. unit: 26.5x17.2x4.8 mm., similar to those used for
controlling the Boston elbow prosthesis.)

>From the recent discussion I learn that Andrew W. (Drew) Smith at U. of
Toronto uses Motion Control's electrodes, which I know have a
relatively large distance between it's electrodes, and the electrodes
themselves are also quite large. Furtermore, their ground electrode is
placed right between the differetial pair, as opposed to Liberty's (and
several other suppliers') which require an "extra" ground electrode
with arbitrary placement.

Can anyone provide expert opinions or personal experience on the

1. The importance and implications of placing the ground electrode
BETWEEN the differential pair. Does this influence the observed
differential EMG signal in any significant (positive and/or negative)
way? (The only positive result I can think of is a possibly improved
Common Mode Rejection.)

2. Larger electrodes cover a larger area and thus measure the EMG of a
larger number of motor units, while they also apparently low-pass
filter the signals (according to Hogan&Mann). Does anyone have opinions
or experience regarding the tradeoffs that have to be made with respect
to this? Does any formal or informal standard exist with respect to
electrode geometry? (What were your specific reasons for choosing the
Motion Control electrodes, Andrew?)

As usual, I will post a summary (edited) of the responces.

By the way, I have no commercial interests in either of the products
mentioned above.

Best regards,
Oyvind Stavdahl


From: Dr. Dan Antonelli, Research Director
Motion Laboratory, Childrens Hospital Los Angeles.

Placement of the reference electrode between the two active electrodes
results in such a small impedance between each active electrode and the
reference electrode that small differences become a larger percentage
of the impedance, resulting in a larger imbalance and degrading the
common mode rejection of the amplifier. Thus, placing the reference
electrode far away from the active electrodes Ref. Hillbiber circa

Wider spacing results in a longer time for the
depolarization/repolarization wave to pass between the two active
electrodes and, thus, you have a smaller bandwith with lower frequency
signals. Coversely, closer spacing will result in higher frequencies
because of the shorter wave transition time between electrodes.

Electrode size affects the signal by dictating the volume from
which the signal is obtained. The practical size is dictated by things
like the availability of double sided tape attachment collars to
practical considerations like the size of the muscles onto which the
electrodes are to be attached without encountering crosstalk from other

Childrens Hospital, Los Angeles uses surface electrodes from Motion
Control and B & L Engineering. When the units are used with the Motion
Control electrodes, the addition of the reference electrode required
for the B & L Electrode improves the CMRR of the Motion Control
electrode. Refer to the BIOMCH-L Yellow Pages for the address of B & L

The last factor that should be mentioned is the method by which the
signal is processed. Raw EMG signals should be normalized by some
basis such as maximal volutional effort. This will make the
utilization of EMG insensitive to day to day variations in placement
and/or electrode size and spacing.