Hi all,
I got sEMG sginals from different subjects at 30%, 45% and 60% of MVC for a particular muscle. I want to calculate the the RMS values (time parameter) for the data. is it necessary to normalize it?
My point of view is, as the loads were already normalized during exercise (first we determined MVC and then subject was asked to perform at 30%, 45% and 60% of his own MVC), hence we are already getting the normalized values of sEMG signal. The activation level of muscle is already 30%, 45% and 60% for every subject.
while processing on data, should I divide the data (got during submaximal exercise) with the maximum value (got during MVC)? will it not be normalization for the second time, and apparently it have no purpose.
Jawad, as far as i can see youve answered your own question with regard to normalisation at the % of their MVC.
Thanks Ben,
The thing I want is some authentic source supporting my argument.
This is what I ARGUED (could be right, could be wrong). I couldn't find this statement in literature. Nearly all articles I read simply passed the statement, "the data is filtered and normalized prior to further processing". This statement signifies that the data has been normalized "off-line".
If i want to stick with my argument, I must have some source doing so. Otherwise, I'm certainly not at a level to coin something new :P Cheers
Dear Jawad,
Although the EMG-force relationship is often linear (not always) does not means that if a participant execute a contraction at 40% of a MVC the EMG also will be 40% of the maximum EMG. For instance, one participant performing a contraction at 40% (MVC) may show a 30% of activation (sub-maximum EMG as percentage of EMG at MVC). A second participant for the same level of load (40% MVC) may show only 21% of activation. However, you only will be able to detect these differences if you normalise your EMG data.
The most commons references to normalise this type of data is the EMG at the MVC or supra-maximal m-waves.
I hope this helps you. Regards!!
@Marcel, Sir really thankful to you for insight.
As I understand from your reply, EMG activation is different in different subjects at same load. Well, this is what is expected (because everyone have different strategy to use muscles). Then the purpose of normalization is gone!!! isn't it?
I think it is due to this very reason, subjects are taken from the same class (age group, BMI, non-athlete etc etc).
Hello Jawad,
Normalisation is necessary exactly because everyone could present different levels of neuromuscular activation at the same percentage of load (in relation to the maximum MVC).
Using my exemple from my first comment, how can you tell if participant 1 has 30% of neuromuscular activation and the second only 21% if you do not normalised the EMG data? It is just not possible. You only will have the absolute data which does not give you what you need: a fair comparison between participants. The normalisation allows you to quantify (in relation to a reference that should be the same for everyone) how much neuromuscular activation was produced during a given task,
There is a literature review from Burden 2010 (Article How should we normalize electromyograms obtained from health...
) that might help you to understand the importance of normalisation procedure and also few other ways to normalise the EMG data.Kind Regards!!!
Thanks Marcel,
got the point. I was thinking the other way. Thanks for the article and detailed reply.
SEMG is not ideal indicator of neural drive. Just think of phase cancelling on the higher contraction levels. Therefore, I would not claim that 30% of EMG max means 30% of the maximal neural drive. However, you may have both normalized and not normalized EMG and look at their relations, too.
Normalization to MVC is still important because of how widely individuals can vary in the raw units of voltage. For example, you may have one person that had 2800 mV on their MVC, and another that had 750 mV. This is due to different tissue characteristics. Perhaps they have the same neural control strategy for your submaximal task, but their submaximal values, unnormalized, would look drastically different.
- Badges
- Science topic
- Similar topics
- Control Systems Engineering
- Filtering