Recent studies have shown that self-generated tactile sensations are perceived as weaker than the same sensations externally generated. This has been linked to a central comparator mechanism that uses efference copy to attenuate the predictable component of sensory inputs arising from one's own actions in order to enhance the salience of external stimuli. To provide a quantitative measure of this attenuation, a force-matching task was developed in which subjects experience a force applied to their finger and are then required to match the perceived force by actively pushing on the finger using their other hand. The attenuation of predictable sensory input results in subjects producing a larger active force than was experienced passively. Here, we have examined the effects of a novel rTMS protocol, theta-burst stimulation (TBS), over primary motor cortex on this attenuation. TBS can alter the excitability of motor cortex to incoming activity. We show that application of a 20s continuous train of TBS, that depresses motor cortex, significantly improves performance in a force-matching task. This suggests that the TBS intervention disturbed the predictive process that uses efference copy signals to attenuate predictable sensory input. A possible explanation for the effect is that TBS has a differential effect on the populations of neurones that generate motor output in M1 than on those neural structures that are involved in generating an efference copy of the motor command.