Provide your definition for efficiency. Efficiency is traditionally defined as the ratio of work output to energy input. sEMG provides a measure of neither nor the ratio. The sEMG amplitude (e.g., RMS) might be considered to be proportional to the energy input for a muscular task but you still need a measure of work output to get anywhere close to calculating something related to efficiency.
I suppose you could argue that muscle group force or torque divided by sEMG RMS could provide a measure of neuromuscular efficiency. However, I see at least three problems doing so. First, the force/torque you measure originates from the entire muscle group whereas the EMG you sample is only coming from the tissue directly beneath the electrodes. The depth into the muscle group that you sample is roughly equal to the spacing between the electrodes in an electrode pair; that distance is usually ~1 cm. The second problem is that the relationship between force and sEMG RMS is often not linear and when it is linear, the line often does not pass through the origin. Whenever this happens, using a ratio of the two variables is not advised. Third, you cannot compare neuromuscular efficiencies between subjects or even between muscles in the same subject. I can only see comparing neuromuscular efficiencies within the same muscle group in the same subject, for example, comparing between high- and low-force contractions.
The most important thing is that, the sEMG signal may (can) not reflect the total amount of force a muscle produces. Therefore, measuring sEMG signal to reflect neuromuscular efficiency could be questionable..
I also agree with Dr. Waren, read this books, also the important book The Nerv - Muscle interaction by Vrbova Gerda & Gordon Tessa. Better is using clinical tests together wih spirometry, ergometry etc.
The other perspective from Micheal Joyner work on Endurance efficiency performance "http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2375555/".
Efficiency is measured in terms of oxygen intake to output performance (running or cycling). Since sEMG is impacted by factors such as muscle type, muscle size, level of contraction, force generated, skin thickness, placement of electrodes to name a few, it may be difficult to draw a direct relationship. However, relative measure can give an idea.
You may want to provide inputs to your question - What is your understanding and definition of Neuromuscular efficiency? In what scenarios one can say that it is inefficient? This is a good and interesting topic,
I suggest isometric contractions. Try to find some articles from Paolo Pinelli (1970-ties) italian scientist, from F. Buchtal school, pioneer of EMG kinesiology.