In this study, we attempted to better delineate the changes in corticospinal excitability that accompany perceptual to motor transformations when people are asked to observe, image or imitate actions. Motor evoked potentials (MEP) from transcranial magnetic stimulation were recorded in the first dorsal interosseous (FDI) muscle of the dominant hand (15 right, 4 left) in five different conditions: (1) passive observation; (2) observation to imitate; (3) imagery; (4) imitation; and (5) counting backwards mentally. MEPs were also recorded at rest at the beginning and at the end of the session to establish baseline (BL) values. For the observation conditions, participants (n=19, 18-38 years) watched video sequences (5s) of hand actions performed by a model with the right arm (passive observation: scissors; observation to imitate: OK sign). Active imitation produced the greatest MEP facilitation compared to baseline, followed by the two observation conditions and the imagery conditions, which all produced similar levels of facilitation (post hoc comparisons). Mental counting produced some facilitation, but this effect was inconsistent. Baseline MEPs remained stable at the end of the session. A further comparison between right-handers (n=15) and left-handers (n=4) revealed no difference in the pattern of modulation across conditions. The similarity found between observation and imagery of hand actions in terms of corticospinal facilitation is interpreted in the light of the motor-simulation theory of Jeannerod [Neuroimage 14 (2001)], which proposes that perceiving actions involves neural simulation of the same action by the observer, thereby explaining the parallel between actions observed and actions imaged at the representational level.