0.05) during spinal cord monitoring. The values of SEP amplitudes at the 2nd stage was higher in comparison with that of the 1st stage(P0.05). There were 5 patients false-positive(3.2％ ) when the values of SEP at pre-incision was used as baseline, while no false-positive case was found in spine exposure baseline. Conclusion Some variability of latencies and amplitudes during different stages of scoliosis surgery is normal. The study results suggest that to determine the SEP after spine exposure as the reference baseline is more reliable for intraoperative spinal cord monitoring.