I wonder about the need of MVC record if I'm going to register EMG signal during 60 m race walking trials. Is there a better method to normalize muscle activity while performing such dynamic movements?
one more way to do it is normalise each muscle activity to the maximum value you have obtained for that respective muscle in that trial( in case you do the analysis without concatenation.) bur you perform concatenation, normalise w.r.t to max value of all trials, in that way you make sure your signal value is [0-1].
"Is there any need to record MVC ...?" evokes a clear response: Yes! To have a commonly accepted reference for the individual muscle situation allows comparison to other research.
"If there is a better method to obtain a reference?" is a more difficult question, because the usually applied method to determine MVC is not dynamic. Why not using a series of 5 MVC measurements (periodically up-down) within 10 s, and doing this before and after your experiment. Then, you obtain a measure for the periodic load, too.