EMG during Walking : On/Off detection

I would like to define an accurate threshold to detect ON/OFF bursts.

Please find attached my EMG signal (rectified).

I am able to code when the muscle activity exceeds the double or triple SD rang (i.e. the muscle is defined to be

“On” or activated).

Threshold=std(A(1:800,1)).*(2.5)

if A>Threshold

DataON=A>Threshold

DataONOFF=diff(DataON)

BUT : Because single spontaneous spikes can easily exceed the SD range, I need to define

a minimum time (minimum subperiod duration) that the EMG signal has to constantly stay OVER the threshold

to be accepted as “On” (e.g. 50 ms). So I would like to translate something like that :

if A>Threshold (and data over threshold during 50 ms)

[Star Strider]

The 'Threshold' value is really low — about 17.5 — with the peaks in the range of 100 to 2000. Also, the sampling frequency is missing, so it’s not possible to determine the activation duration. I would use a differerent approach to this, however I first need to understand what exactly you want to do.

It would also help to know what release you’re using, and if you have the Signal Processing Toolbox.

[Marie-Caroline]

The sampling frequency is 2000 Hz.

My objective is to define subperiods (each walking step) within my entire trial.

I use Matlab R2023b and have the Signal processing Toolbox.

Many thanks in advance for your help !

Answer 1

[Alex Wu]

Hi Marie-Caroline

You may find sigrangebinmask useful.

Please check this example Label High-Amplitude Regions in EMG Signal, where we use sigrangebinmask function to label regions where the EMG amplitude is greater than 0.25 (threshold) for at least 0.5 seconds (minimum subperiod duration).