Hello Mr Johnson and all in the biomedical user group. I constructed an EMG amplifier similar to Beirut's and got the following oscilloscope output. EMG measurements were taken from my flexor digitorum superficialis forearm muscle. Is the quality of the signal sufficient for further processing in labview? I will upload the schematic as soon as possible. Should I add a notch filter using a UAF42 chip for better results? Im planning to transmit this signal wirelessly using arduino and xbee tranceivers to an NI SCC-68 connector block for further processing.
This signal looks great - good steady baseline, no clipping, nice sharp high-frequency content. It would be great to digitize this signal now in LabVIEW (use 1kHz sampling minimum) and try some EMG processing. A simple way to go is run the waveform through the LabVIEW RMS function. Another simple approach is to rectify the signal (this is just taking the absolute value in LabVIEW) and passing the result through a low-pass filter with a relatively low cutoff frequency to effectively average or smooth the output to a basic trend line. There are also some nice EMG processing VI's in the Biomedical Starter Kit download, as well as a ready-to-run logger for acquiring the data to disk in the TDMS format (compact binary format with easy File I/O support in LabVIEW).
I am happy to see your circuit works.
Until now i haven't made any adjustments in my project since i had a lot of exams for the university, but after 1 week from now i have to solder my board and test it to see the results.
Jazlan did you use the same circuit i posted in your project,or have you made any adjustments to the one i posted earlier ?, and if so please share your improved circuit.
Thank you Mr Steve Johnson for your valueble feedback. I previously constructed an op-amp 'envelope detector' circuit which consists of a precision half wave rectifier circuit and a low pass filter with very low cut off frequency (around 0.3Hz). I tested various combinations of resistor and capacitor values determine the low pass filter cut off frequency value that gives smoothest EMG signal output. It was a bit troublesome to test many different resistor/ capacitor combinations and furthermore different muscles require a different low pass filter cut off frequency value in order to get a smooth signal.. Thank you for pointing out that these signal processing techniques can be implemented using Labview functions and the biomedical start up kit. Varying the low pass filter cut off frequency using software will be more convenient instead of varying the cut off frequency electronically.
I have attached the circuit schematic (Beirut) together with a short experiment description. I used an instrumentation amplifer with right leg drive since im planning to modify the same circuit to measure ECG later on. I did a small experiment to try and relate hand grip force to EMG from the flexor digitorum superficialis forearm muscle. I gripped an exercise hand gripper at low, intermediate and strong hand grips and observed how the frequency of the signal measured from the oscilloscope changes with increased hand grip strength. Each hand grip level (low,intermediate and strong) was held for 10 seconds and the frequency of the signal was observed.
Hand Grip Strength
Hello Mr. Steve and fellow bomedical user group members.
I have access to a National Instruments USB 6008 DAQ and a laptop installed with Labview 8.6. The National Instruments USB 6008 DAQ has 12 bit resolution and 10 KS/s maximum sampling rate. Mr Steve mentioned earlier that a suitable sampling rate for EMG signals should be atleast 1000Hz. Since the NI USB 6008 can sample 10000 samples per second, I think I can sample the EMG signal without aliasing but im still not whether I can use the NI USB 6008 for this application. I plan to aquire 2 channel EMG signals from my forearm muscles and one more signal from a hand dynamometer sensor, input these signals into the NI USB 6008, sample each of these signals with a sampling rate of around 5000 samples/sec and do some signal processing in Labview 8.6 (using the biomedical startup kit). Can I accomplish these tasks by using the NI USB 6008?
The advantage of using the NI USB 6008 is that I can connect it to a laptop and power the laptop with batteries only while taking measurements from the subject. This is a student project and im working under the supervision of a professor but i feel more comfortable not having any connections to the 120V AC wall power supply while taking measurements from subjects.
I tested the circuit and i got the results shown in the attached files.
Are the signals i attached good enough to be used for processing or should i have to make any more adjustments?
Sorry for the late reply Beirut. Im glad to see you have inputted your EMG amplifier output into labview. In IMG_0330 these EMG signals are measured when your muscles are at rest right? Where as IMG_0325 is the EMG signal when you’rerepeatedly flexing and resting your muscles. I think the signals are alright,the changes when the muscle is at contracted and at rest are visible.
What is the gain of the circuit at this stage (IMG_0325)? Which muscle are you measuring EMG from? My guess is that you have passed it through your instrumentation amplifier (gain = 5) and Sallen Keyband pass filter (unity gain). But for the final inverting op amp amplifier perhaps you may want to increase the gain so that the total gain of the EMG amplifier is between 1000-2000. As of now I notice that the amplitude of your signal is around +10mV to -10mV. In my circuit the output voltage of the EMG signal is from -2.5V to 2.5V. I applied an op amp offset circuit to shift -2.5Vto 2.5V to 0V to 5V TTL range.
In ‘with contraction.jpeg’ the signal also looks alright but I think if you want to clearly distinguish between when there is muscle activity and when the muscle is at rest you could increase the timebase from 50ms to 500ms or 1 second. To summarize I think the signal is alright but you could further increase the gain of your EMG amplifer by changing the resistors values in your last inverting op amp stage but do be carefull not to 'overamplify' and cause op amp saturation. And also in order to clearly observe the difference when the muscles are at rest and when the muscles are contracting you could increase the time base to 500ms or 1 sec, both on your oscilloscope and in labview.
By the way how do we measure the signal to noise ratio? Im also a student learning about biosignals =). Feel free to contact me at email@example.com
Sure - the USB-6008 should be fine as long as you have a good low-pass filter on your preamplifier to make sure that the high-freq stuff gets attenuated before the digitizer to avoid any aliasing.
Signal to noise ratio is the power ratio between the signal and the background noise.
But if the signal and the noise are calculated through the same impedance it can be calculated as the ratio of the rms value of the signal to that of the noise (Squared).
This link contains the information needed in order to measure it.