Superimposed electrical stimulation techniques can be used to detect central activation failure (CAF), that is, incomplete central nervous system recruitment or suboptimal activation of motor units. The purpose of this study was to evaluate the effects of two stimulation parameters on perceived discomfort and torque increment size and variability. Discomfort was evaluated using a visual analog scale (0-100 mm) for pain. The rectus femoris muscle of the dominant leg of 24 young healthy men was stimulated during submaximal (80% maximal) voluntary contractions. The size and variability of torque increments and perceived discomfort were assessed following stimulation with: (1) pulse trains (100 HZ, 150 V, 0.2-ms pulse duration) of different lengths (50 ms and 100 ms); and (2) pulse trains (100 HZ, 100 ms, 150 V) with different pulse durations (0.2 ms and 0.1 ms). Pulse trains of 100 ms generated larger torque increments and produced less variability, but caused more discomfort than pulse trains of 50 ms. Average discomfort ratings for pulse trains of 100 ms were 43.1 mm, and of 50 ms were 53.2 mm. There was no difference in torque increment size or in variability between pulse trains with pulse durations of 0.1 ms and 0.2 ms, whereas discomfort was less for the shorter pulse durations; average discomfort ratings were 53.1 mm and 58.1 mm for pulse durations of 0.1 ms and 0.2 ms, respectively. Thus, the appropriate selection of stimulation parameters can reduce discomfort but maintain the ability to detect CAF.