To investigate the effect of initial conditions on the modulation of motor unit discharge during fast voluntary contractions, we compared ballistic isometric contractions of the ankle dorsiflexor muscles that were produced from either a resting state or superimposed on a sustained contraction. The torque of the dorsiflexors and the surface and intramuscular EMGs from the tibialis anterior were recorded. The results showed that the performance of a ballistic contraction from a sustained contraction ( approximately 25% maximal voluntary contraction (MVC)) had a negative effect on the maximal rate of torque development. Although the electromechanical delay was shortened, the EMG activity during the ballistic contraction was less synchronized. These observations were associated with a significant decline in the average discharge rate of single motor units (89.8 +/- 3.8 versus 115 +/- 5.8 Hz) and in the percentage of units (6.2 versus 15.5% of the whole sample) that exhibited double discharges at brief intervals (= 5 ms). High-threshold units that were not recruited during the sustained contraction displayed the same activation pattern, which indicates that the mechanisms responsible for the decline in discharge rate were not restricted to previously activated units, but appear to influence the entire motor unit pool. When a premotor silent period (SP) was observed at the transition from the sustained muscular activity to the ballistic contraction (19% of the trials), these adjustments in motor unit activity were not present, and the ballistic contractions were similar to those performed from a resting state. Together, these results indicate that initial conditions can influence the capacity for motor unit discharge rate and hence the performance of a fast voluntary contraction.