Study design: Experimental animal study. Objectives: Locomotion analyses in rat spinal cord contusion injury (SCI) models are widely used for the evaluation of recovery of supraspinal locomotor control. However, many commonly used locomotion tests are inadequate to test for spinal cord integrity as they assess motor function that can be highly mediated through below-level propriospinal pattern-generating circuitry, independently of below-level perception. Here we report a behavioral motor test that is more sensitive for spinal cord integrity, even 6 weeks after injury: the backward locomotion rotating rod. Setting: University of California -San Diego. Methods: A modified rotating rod test was run in reverse. The rod diameter was increased and thin rubber lining was added. As a reference, we included commonly used motor tests: BBB score, catwalk gait analysis, motor-evoked potentials, single frame analyses, a forward rotating rod test and the 551 inclined ladder test. Results: Unlike commonly used motor tests, the backward locomotion rotating rod test significantly discriminates between both sham-operated (falling latency: 20.4 s s.d. ± 4.5) vs mild SCI animals, and mild vs moderate SCI animals (differences between each group at acute, subacute and chronic phases: X6 s, Pp0.01). Moderate SCI animals were practically unable to make even slight backward hindpaw movements. The backward locomotion ability in the chronic phase correlates best with BBB locomotor scores from the acute phase. Conclusion: Our data show that backward locomotion is a highly sensitive and quick test to discriminate between sham, mild and moderate SCI, even after 6 weeks. Backward locomotion testing may improve the translational value of experimental results for the clinic.