In this study we directly tested the hypothesis that isometric strength training increases voluntary drive to muscles. In addition, it was attempted to replicate the findings of an earlier study that showed imagined training increases voluntary strength as much as actual training, as this finding provides key support for the hypothesis that training increases voluntary drive (Yue and Cole 1992). Fifty-four subjects were randomly allocated to groups that performed 8 weeks of isometric training of the elbow flexor muscles, imagined isometric training, or a control task involving the lower limbs. Voluntary isometric strength and activation of the elbow flexor muscles were measured before and after training. Voluntary activation was measured with a sensitive form of twitch interpolation. Training, imagined training and control groups increased voluntary isometric elbow flexor strength by means of 17.8% (±3.1 SEM), 6.8% (+2.6) and 6.5% (±3.0), respectively. The training group increased in strength significantly more than imagined training and control groups (P = 0.01 for both comparisons), but the small difference between imagined training and control groups was not significant (P = 0.31). Prior to training, voluntary activation of all subjects was high (96.2 ± 0.5%). This did not change significantly with training and there were no significant differences between groups. These data challenge the hypothesis that training of the elbow flexor muscles increases isometric strength by inducing adaptations of the central nervous system, because they show that training does not increase voluntary activation and imagined training does not increase strength.