Objectives : Peripheral neurodegeneration occurs in diabetes mellitus (DM), both sensory and motor nerve, but we don't know exactly if DM affects central nerve pathway for all studies. Electrophysiologic study is one of the most important diagnostic tools for diabetic neuropathy. Electroneurography and electromyography are usually used, but evoked potentials (EP) is more sensitive to small nerve fiber damages and useful for central nerve evaluation in addition to peripheral nerves. Most diabetic neuropathy studies by EP have been performed with somatosensory evoked potentials (SSEP). In contrast, the objective of this study is to investigate if DM targets central motor neurons by assessing the relation between fasting blood sugar (FBS) and motor evoked potentials (MEP) latency.Methods : We inspected the medical records of 34 patients who had MEP tests during admitting days. The latency from cervical portion to abductor pollicis brevis was used as peripheral motor conduction time (PMCT), and the latency from vertex to cervical portion was used as central motor conduction time (CMCT). Then, they were correlated to FBS using correlation analysis. Results : There was a significant linear relation between FBS and PMCT (Pearson’s correlation coefficient r=0.487, p<0.01), but a poor linear relation between FBS and CMCT (Pearson’s correlation coefficient r=-0.197, p>0.05).Conclusions : This study suggests that prolonged latencies of MEP in DM may be due to peripheral neuropathy rather than dysfunction of central motor pathway, therefore the clinical use of MEP to diabetic neuropathy has to be divided segmentally.