Muscle electromyographic activity normalized to maximal muscle activity, not to M max , better represents voluntary activation

In human applied physiology studies, the amplitude of recorded muscle electromyographic activity (EMG) is often normalized to maximal EMG recorded during a maximal voluntary contraction. When maximal contractions cannot be reliably obtained (e.g. in people with muscle paralysis, anterior cruciate ligament injury, or arthritis), EMG is sometimes normalized to the maximal compound muscle action potiential evoked by stimulation, the M max . However, it is not known how these two methods of normalization affect the conclusions and comparability of studies. To address this limitation, we investigated the relationship between voluntary muscle activation and EMG normalized either to maximal EMG or to M max . Twenty-five able-bodied adults performed voluntary isometric ankle plantarflexion contractions to a range of percentages of maximal voluntary torque. Ankle torque, plantarflexor muscle EMG, and voluntary muscle activation measured by twitch interpolation were recorded. EMG recorded at each contraction intensity was normalized to maximal EMG or to M max for each plantarflexor muscle, and the relationship between the two normalization approaches quantified. A slope >1 indicated EMG amplitude normalized to maximal EMG (vertical axis) was greater than EMG normalized to M max (horizontal axis). Mean estimates of the slopes were large and had moderate precision: soleus 8.7 (95% CI 6.9 to 11.0), medial gastrocnemius 13.4 (10.5 to 17.0), lateral gastrocnemius 11.4 (9.4 to 14.0). This indicates EMG normalized to M max is approximately eleven times smaller than EMG normalized to maximal EMG. Normalization to maximal EMG gave closer approximations to the level of voluntary muscle activation assessed by twitch interpolation.