Accurate functional outcome measures are essential in assessing therapeutic interventions after experimental spinal cord injury (SCI). We examined the hypothesis that serial recording of somatosensory (SSEP) and myoelectric motor evoked potentials (mMEPs) would provide complementary information to standard methods of behavioral analysis in a rat model of SCI and would allow objective discrimination of functional recovery in sensory and motor tracts. Clip compression injury of varying severity (sham, 23 g, 34 g, 56 g) and transections were performed at T1 in adult rats. SSEPs were recorded from the right sensorimotor cortex (SMC) after stimulation of the contralateral hind paw; mMEPs were recorded from the paraspinal, quadriceps, and the tibialis anterior muscles after anodal stimulation of the SMC. The inclined plane and Tarlov techniques were used to assess clinical neurological function. All outcome measures were assessed weekly prior to and up to 6 weeks following injury. Changes in clinical neurological function as assessed by the inclined plane and Tarlov methods varied with increasing injury severity (R = -0.72 and R = -0.73, respectively). SSEP latency was strongly correlated with injury severity (R = 0.92) and with clinical behavioral scores (R = -0.93 for inclined plane). The tibialis anterior mMEP correlated significantly, though weakly, with changes in inclined plane (R = 0.49) and Tarlov scores (R = 0.41). Although the mMEPs were sensitive to the presence of SCI, these recordings did not discriminate between severities of injury. We conclude that serial recording of SSEPs but not myoelectric MEPs correlates closely with the extent and temporal course of clinical neurological recovery after graded SCI in the rat.