OBJECTIVE: The purpose of this study is to evaluate the relationship between cognitive function and findings of evoked potential study in chronic renal failure patients. METHOD: Thirty chronic renal failure patients with cognitive dysfunction were recruited, whose mini-mental state xamination (MMSE) scores were less than 24 points. According to the underlying diseases of chronic renal failure, we categorized hirty patients into diabetic group (11 patients) and non-diabetic group (19 patients), and the control-group was composed of 15 normal volunteers. Somatosensory evoked potential (SEP) on stimulating median and posterior tibial nerves, and cortical and spinal conduction time of the motor pathways were valuated. RESULTS: In tibial nerve SEP, N22-P38 interpeak latencies (IPL) were 18.1 4.2 msec in the patient group and 15.7 1.9 msec in the control group, respectively. In MEP, cortico-lumbar central motor conduction times (CMCT) were 19.5 2.7 and 16.5 3.0 msec, respectively. CMCT were prolonged in patients than controls (p<0.05). There was significant correlation between serum creatinine concentration and N22-P38 IPL (r=0.64, p<0.05), but, there were no correlations between the underlying diseases of chronic renal failure, duration of disease, MMSE score and cortico-lumbar CMCT, N22-P38 IPL (p>0.05). CONCLUSION: Evoked potentials will be helpful in evaluating the patients with cognitive dysfunction in chronic renal failure.