We studied the motor simulation processes involved in concurrent action observation and motor imagery (AO+MI) using motor evoked potentials induced by transcranial magnetic stimulation. During congruent AO+MI, participants were shown videos of a model's hand performing rhythmical finger movements, and they imagined moving the same finger of their own hand in synchrony with the observed finger. During incongruent AO+MI, the imagery task involved a different finger from the observed one. As expected, congruent AO+MI yielded robust facilitatory effects, relative to baseline, only in the effector involved in the task. Incongruent AO+MI produced equally pronounced effects in the effector that was engaged in MI, whilst no corticospinal facilitation was found for the effector corresponding to the observed action. We further replicated that engaging in pure AO without MI does not produce reliable effects. These results do not support the proposal that observed and imagined action are both simulated at the level of the primary motor cortex. Rather, motor imagery alone can sufficiently explain the observed effects in both AO+MI conditions. This bears clear implications for the application of AO+MI procedures in sport and neurorehabilitation.