Reproducibility of the visual evoked potential using a light-emitting diode stimulator

The intersubject and intrasubject reproducibility of the VEP was evaluated using two different methods of producing pattern reversal--a mirror/projector system and a light-emitting diode (LED) system. Intersubject reproducibility was determined in 100 normal subjects (50 males, 50 females). Ten subjects were studied on ten different occasions over 11 months to establish intrasubject reproducibility. The two methods gave comparable results. It is concluded that the mirror/projector system has no advantages over the LED system, which is cheaper, more robust and more convenient. Most of the intersubject variability was found to be due to subject variables and it is concluded that technical improvements are unlikely to reduce further the variability of the test. A variability in the latency of P100 of up to 7.7 ms was recorded in the serial studies on the ten subjects, indicating that in longitudinal studies on patients, changes in latency must exceed 9-10 ms to be significant. With both stimulus systems, there were significant sex-related differences in latency (P50 and P100 but not N70) and amplitude (N70-P100). The extent of the difference was such that the upper limit of normal latency for P100 (2.5-3 SD from the mean) was 4.2-4.7 ms longer for males than females-a value which exceeds 1 SD. Separate control values for males and females are advisable whichever method is used to produce pattern reversal.