Motor resonance is the modulation of M1 corticospinal excitability induced by observation of others' actions. Recent brain imaging studies have revealed that viewing videos of grasping actions led to a differential activation of the ventral premotor cortex depending on whether the entire person is viewed versus only their disembodied hand. Here we used transcranial magnetic stimulation (TMS) to examine motor evoked potentials (MEPs) in the first dorsal interosseous (FDI) and abductor digiti minimi (ADM) during observation of videos or static images in which a whole person or merely the hand was seen reaching and grasping a peanut (precision grip) or an apple (whole hand grasp). Participants were presented with six visual conditions in which visual stimuli (video vs static image), view (whole person vs hand) and grasp (precision grip vs whole hand grasp) were varied in a 2 × 2 × 2 factorial design. Observing videos, but not static images, of a hand grasping different objects resulted in a grasp-specific interaction, such that FDI and ADM MEPs were differentially modulated depending on the type of grasp being observed (precision grip vs whole hand grasp). This interaction was present when observing the hand acting, but not when observing the whole person acting. Additional experiments revealed that these results were unlikely to be due to the relative size of the hand being observed. Our results suggest that observation of videos rather than static images is critical for motor resonance. Importantly, observing the whole person performing the action abolished the grasp-specific effect, which could be due to a variety of PMv inputs converging on M1.