OBJECTIVE: The aim of present study is to obtain the reference values of motor and sensory nerve conduction study (NCS) of the median nerve and to re-assess the values of criteria for abnormal findings in the electrophysiological diagnosis of Carpal tunnel syndrome (CTS). METHOD: Median and ulnar NCS were performed in 62 neurologically healthy adults as controls and 142 patients with CTS. Reference values of the median and ulnar NCS were obtained from the controls and the criteria for the diagnosis of CTS were made. Sensitivities of diagnostic criteria were calculated. RESULTS: The criteria included 7 items: (1) A 5-cm short segment study through carpal tunnel, > or =1.3 msec; (2) median motor latency (8 cm), > or =4.0 msec; (3) median sensory latency (14 cm, onset), > or =3.0 msec; (4) median sensory nerve action potential amplitude (baseline to peak), < or =20 uV; (5) ratio of median to ulnar motor latency, > or =1.5; (6) ratio of median to ulnar sensory latency, > or =1.2; (7) ratio of median to ulnar sensory amplitude, < or =0.6. The order of the highest to lowest sensitivity were as follows: a 5-cm short segment study, median sensory latency, median motor latency, ratio of median to ulnar sensory latency, and ratio of median to ulnar motor latency. Approximately 65% of the patients met all the 7 diagnostic criteria and 95% of the patients met 3 or more. CONCLUSION: The criteria proposed in this study would enhance the diagnostic sensitivity for the CTS.