Muscle fatigue or neuromuscular disease may result in central activation failure during maximal voluntary contractions (MVCs). Superimposition of an electrically stimulated contraction during an MVC has been used to detect central activation failure. To determine the most sensitive means of quantitating central activation failure using this technique, we compared the increment in isometric force from single, double, and high-frequency trains (50 Hz, 500 or 1000 ms) of stimuli of the peroneal nerve imposed during three separate MVCs of the dorsiflexor muscles. Completeness of activation was quantitated with the central activation ratio (CAR) = MVC/(MVC + stimulated force). Comparisons were made of the CARs of three groups of subjects during the three stimulation conditions: 7 healthy subjects, 13 patients with amyotrophic lateral sclerosis, and 5 healthy subjects after fatiguing exercise. For all three groups, the CAR was significantly lower during the train of stimuli condition (means = 0.76-0.89) compared with either the single or double stimuli conditions (means = 0.96-1.00). The results suggest that a superimposed high-frequency train of stimuli is a more sensitive indicator of central activation failure during isometric MVCs compared with either the superimposed single or double stimuli methods.