Spinal cord fMRI is a useful tool for studying spinal mechanisms of pain, hence for analgesic drug development. Its technical feasibility in both humans and rats has been demonstrated. This study investigates the reproducibility, robustness, and spatial accuracy of fMRI of lumbar spinal cord activation due to transcutaneous noxious and non-noxious electrical stimulation of the hindpaw in alpha-chloralose-anesthetized rats. Blood oxygenation level-dependent (BOLD) and blood volume-weighted fMRI data were acquired without and with intravenous injection of ultra small superparamagnetic iron oxide particles (USPIO), respectively, using a gradient echo (GE) echo planar imaging (EPI) technique at 4.7 T. Neuronal activation in the spinal cord induced by noxious stimulation to the hindpaw (2 ms wide, 5 mA amplitude, known to activate C-fibers) can be robustly detected by both fMRI techniques with excellent reproducibility and peaked at the stimulus frequency of 40 Hz. However, both fMRI techniques were not sensitive to neuronal activation in spinal cord induced by non-noxious stimulation (0.3 ms, 1.5 mA, known only to activate A-fibers). Spatially, the fMRI signal extended approximately 5 mm in the longitudinal direction, covering L(3)-L(5) segments. In the cross-sectional direction, the highest signal change of blood volume-weighted fMRI was in the middle of the ipsilateral dorsal horn, which roughly corresponds to laminae V and VI, while the highest signal change of BOLD fMRI was in the ipsilateral dorsal surface. This study demonstrates that spinal cord fMRI can be performed in anesthetized rats reliably and reproducibly offering it as a potential tool for analgesic drug discovery.