In recent months I’ve written about speech synthesis, voice recognition, and direct brain control of robotic devices. Most recently I discussed what Fluke has to offer to investigate the brain activity used to control some recently available games.
This month we’ll look into the electrical activity associated with voluntary muscle control prostheses, the devices that help amputees overcome the loss of a hand or arm, for example.
Robot hand with string tendons
How voluntary muscles work to control your hand
It’s amazing, really. In order to maintain the manual dexterity we enjoy, nature has provided an ingenious system of tendons to connect to a system of finger control muscles, called flexors and extensors, which are contained in our forearms. Additional muscles in our forearms allow rotation of the hand and other moves involved in gripping.
In a recent Fluke-funded program at a local children’s museum, I supervised a construction project where the children made robotic hands out of plastic straws, string, a food handler’s glove, and some cardboard and tape. By pulling on the strings coming from the assembled materials, the kids could clearly see how the tendons control their fingers under remote muscle control. I had them feel the muscle action on their underarms as they clenched and relaxed their fists. It’s easy to see how this principle could be adapted to a prosthetic device with actuators pulling the strings, but how would one provide the program control necessary for smooth operation?
Prosthetic hand evolution
Present day capabilities in prosthetic hands are a far cry from the primitive devices portrayed in the story of Peter Pan, where Captain Hook had only a crude hook attached to his forearm. Later improvements used cables operated by mechanical movements of a shoulder to operate a two-finger gripper in place of the hand.
Today, a myoelectric prosthesis can use the electrical signals detected in the skin over residual forearm muscles, or directly in the nerve paths to those muscles, to operate a lifelike battery powered hand for an amputee. Here are a couple of examples:
My experiment for the month
After reading about TouchBionics, I wondered how difficult it would be to locate the signals on the skin of the forearm associated with specific finger movements. I decided to try transcutaneous electrical nerve stimulation (TENS) to see if I could identify some finger control sites on my arm. Note that while this sounds impressive, you can actually purchase such a unit, advertised for muscle relaxation and pain control, for well under $100.00. Note also, that it is highly recommended that these devices only be used under the direction of a medical professional. For my purposes, however, the stimulation is localized to the underside of the forearm, so no electrical currents are being induced around or through the heart, throat, or brain. It didn’t take long to identify sites that could cause a finger to twitch. I imagined myself as the frog working with Luigi Galvani in Italy, as he experimented with electrical stimulation in 1791.
Measuring the nerve signals that cause muscle contraction
As it turns out, the myoelectric signals available on the surface of the forearm are much larger than the electroencephalogram (EEG) signals detected on a person’s scalp. So now I could use a Fluke digital multimeter (DMM), such as the 87V or the 289, to detect the signals. In the case of the 87V set to measure ACmV, the peak signals can be recorded using Peak Min-Max when I clench and unclench my fist.
To make good connections to my arm, I made my own electrode pads using aluminum foil under adhesive bandages as contact points. I also prepared the skin at the measurement sites by scrubbing the target points using a mild saline solution before my crude electrodes were applied.
Sample taken with the Trend display of the Fluke 289 multimeter.
For the 289, I had the option to use the record function to capture activity as long as I limited my hand grip movements to once every two or three seconds. That’s because of the minimum sampling interval of one second available on the meter.
What I discovered using the 289 was that the input sensitivity of the mvDC function is more than adequate on the 50 mV range, but the minimum sample interval of one per second for the record function was not fast enough to reliably capture the peak signals. It was, however, possible to measure the peak signals using Peak Min-Max function.
I used a Fluke 199C ScopeMeter® portable oscilloscope to accomplish the final, and most detailed, recording of muscle activity in the forearm that controls finger movement. I had the option of using either TrendCapture or ScopeRecord™ to document my tests. In this experiment, I verified that the input sensitivity was more than adequate to register the signals, and that the ScopeRecord function would be the best method for collecting the desired information.
One caution here - I did have a problem with 60 Hz noise on my signals while the 199C was plugged into line power for battery charging. it’s best to use the battery powered option for the ScopeMeter® to eliminate possible low level interference from the line power option. That’s one of the real plusses of the tools in the Fluke DMM and ScopeMeter families - operation completely free of ac power systems.
Student projects involving robotic hands
Here’s an example of a student project that closely approximates the children’s museum project using straws and strings. It’s a little more complex but uses the same basic principles.
And, here is a more complex radio controlled robot hand. One can easily imagine myoelectric signals from a sleeve with electrodes on an amputee’s arm controlling this rather than the radio control gloved hand used in this example.
We’d like to hear about your applications of Fluke test tools in bioengineering projects such as these. Let us know if you have been involved in such a project: FlukePlus@fluke.com.