Several studies have focused upon adaptations within human skeletal muscle in response to electrostimulation (ES) training and have indicated that this modality of training enables the development of maximal force, albeit with a considerable diversity in reported strength gains (for review see 18, 42). It has been reported that ES training, performed under isometric conditions, induced specific isometric torque gains ranging from 0.2 39 to 44% 41 but could also decrease maximal torque production by 2.5% 39. There has been a certain amount of debate concerning the dependency of torque gain specificity upon the mode of training contraction 18. Indeed some investigators have reported that isometric ES training improves muscular torque under dynamic conditions. It has been shown from Torque/Angular Velocity (T/AV) relationships that isometric ES training increases torque for concentric angular velocities ranging from 30 to 1208× s–114 or over a whole range of concentric angular velocities 31. Pichon et al. 38 have shown a significant increase in torque with concentric angular velocities over 1808× s–1 and at 608× s–1 under eccentric conditions after 3 weeks of isometric ES training of the latissimus dorsi (3 sessions per week). This wide range of torque gains after isometric ES training has been ascribed to differences in the mode of stimulation (frequency, pulse duration), training protocols (number and duration of the sessions), methods of testing, and the muscle groups studied 12, 18.

The underlying mechanisms of increase in muscle torque production in response to isometric ES training can result from changes which occur within the central neural drive 13, 29, 33 and/or at peripheral level 11. A common approach used to distinguish central versus peripheral adaptations has been to analyze modifications to the EMG/Isometric Torque relationship obtained before and after training. Indeed, and as suggested by Moritani and de Vries 35, changes in the slope of this relationship could result from peripheral adaptations whereas increases of the maximal myoelectrical activity of the muscle studied resulted from modifications of the central neural drive. More recently it has been shown that increases in maximal torque production with training can not only be attributed to modifications of the central neural drive of the trained muscle but also to a reduction in the level of co-activation of the antagonist muscles 4. This phenomenon has been observed after isometric voluntary training but to our knowledge no study has investigated the effects of isometric ES training on the co-activation levels of the antagonist muscles. In these conditions a reduction in the co-activation level of the antagonist muscle could induce a modification of the slope of the EMG/Isometric Torque relationship. Thus, in order to elucidate this particular process, it appeared necessary to simultaneously evaluate agonist and antagonist myoelectrical muscular ac-