In experimental stroke research one of the key questions is whether animals models truly reflect human disease. One important reason for modelling stroke is to produce tissue changes that can be used to validate the correlation between the results of advanced radiological imaging and pathology. The aim of this study was to compare two minimally invasive rat models of focal ischaemic stroke as to their ability to simulate clinical stroke reproducibly. In one model a focal infarct was induced using the photochemical Rose Bengal technique. The second model was based on transvascular occlusion of the middle cerebral artery using a cervical approach. In all animals we performed MRI at different times with two different paramagnetic contrast agents. The time course of blood-brain-barrier disruption in the Rose Bengal model differed entirely from that in human brain ischaemia. The experimental stroke showed marked contrast enhancement in the first hour after the onset of ischaemia. On the other hand, the MRI changes in the suture occlusion model were very similar to the changes observed in human brain ischaemia: no early disruption of the blood-brain-barrier and increased T2-signal 4-6 h after the onset of stroke. We hope that the suture occlusion model will be used for research not only by basic and clinical scientists but also by radiologists, who, using modern imaging methods, may greatly influence the diagnosis as well as the management of this devastating disease.