In Deep Brain Stimulation surgery, the efficiency of the procedure heavily relies on the accuracy of the placement of the stimulating electrode. Meanwhile, the effectiveness of the placement is difficult due to brain shifts occurring during and after the procedure. We propose an approach to overcome the limitations of current planning software that ignores brain shift. In particular, we consider the motion of vascular structures in order to reduce risks of dissecting a vessel during the procedure. Facing the difficulty to produce an exact brain shift prediction, we propose to build a brain shift aware risk map which embeds the vascular motion risk. This risk map is extrapolated using simulation from clinical studies that provide statistics on the displacement of anatomical landmarks during the procedure. Risk maps can be directly integrated into automatic path planning algorithms to better predict optimal electrode trajectories. The method relies on a physics-based simulation that takes into account brain deformation, electrode placement, cerebrospinal fluid, and vascular motion. The goal is to reproduce the spread of brain shift situations that are noted in clinical studies. Preliminary results show that it is possible to compute safe electrode trajectories even in case of brain shift and yet optimal regarding the placement within the targeted area.