The dorsal premotor cortex (PMd) has a role in selection of movements through a densely interconnected network linking the premotor areas and the motor cortex (MCx) of the two hemispheres. Mochizuki et al. (2004) recently described a method using paired pulse transcranial magnetic stimulation to explore the transcallosal projections connecting the PMd with the contralateral primary MCx. They found that the amplitude of EMG responses evoked by a standard TMS pulse to the left MCx could be modulated by applying a conditioning stimulus (CS) to the right PMd some 8-10 ms beforehand. Inhibition occurred if the intensity of the CS was 110% rest motor threshold (RMT). In another study facilitation was induced at a similar interstimulus interval (8 ms) using low intensity CS to the left PMd (CS=80% RMT) (T. Bäumer, F. Bock, G. Koch, R. Lange, J.C. Rothwell, H.R. Siebner & A. Münchau, unpublished observations). We tested how the excitability of these connections changed in the reaction interval of an acoustic choice reaction time (CRT) task in 10 healthy subjects. Subjects squeezed a button with either their left or right hands after a low or high tone respectively. PMd-MCx connectivity at an ISI of 8ms was studied at intervals from 50-200 ms after the reaction signal using two intensities of CS. Unexpectedly, the usual patterns of transcallosal suppression (CS=110% RMT) and facilitation (CS=80% AMT) were absent for most of the reaction period, apart from the time point 75 ms and 100 ms after the reaction signal. A three-factor ANOVA on the data with time course, hand, and intensity showed a significant three-way interaction (F=2.62; P<0.05). Subsequent analyses showed that this was because facilitation (CS=80% RMT) from left PMd to right MCx was prominent at 75 ms when subjects had to move the left hand. Conversely, the inhibitory interaction (CS=110% RMT) from left PMd to right MCx became prominent at 100 ms when subjects had to move the right hand. In a control experiment we did not find any significant changes on either side in the functioning of intracortical circuits (SICI and ICF) in right MCx at 75 or 100 ms. Our results demonstrate that the transcallosal interactions between PMd and contralateral M1 previously observed at rest may play a physiological role in movement planning. The left PMd may either facilitate or inhibit the contralateral motor cortex depending on whether the ipsilateral or the contralateral hand is selected for movement.