Background: Spinal cord ischaemia animal models were established by selective ligation of the lumbar artery in a craniocaudal direction between the renal artery and the aortic bifurcation. Transcranial electrical stimulation motorevoked potentials were measured to enable their use in future studies on spinal cord ischaemia protection. Methods: Thirty-three New Zealand rabbits were randomly divided into 6 groups. Transcranial electrical stimulation motor-evoked potentials were recorded before vascular ligation, 30 min after vascular ligation, and 2 days after vascular ligation. Motor functions were assessed after surgery and 2 days after vascular ligation. The specimens were taken 2 days after ligation for histopathologic observation.Results: With increased numbers of ligations, a transient extension of the latency became clear, but there were no significant differences in the statistical analysis. Analysis of variance after ligation at the same time in each group and t tests before and after ligation (P > 0.05) were not significant. One or 2 ligations did not cause spinal cord ischaemic damage. There were no significant differences before and after ligation for the amplitude (P > 0.05). With increased numbers of ligations, the amplitude before and after ligation was gradually reduced in the 3-5 ligation groups (P < 0.05).Conclusions: Ligation of segmental spinal cord vessels on 1 or 2 levels did not cause ischaemic damage. Spinal cord ischaemia was observed after 3, 4, or 5 ligations. The amplitude was more sensitive to spinal cord ischaemia than latency. Spinal cord function can be predicted by early changes in the amplitude.