OBJECTIVE: The inter-individual variation in the shape of a visual evoked potential (VEP) is large even for simple stimuli. We compared the inter-individual variation in VEP waveform to the intra-individual stability. METHODS: We recorded VEP with checkerboard stimulation in 10 women aged 19-29years in two sessions. We determined the latencies and the peak-to-peak amplitude of N75 and P100. As a new approach in VEP research, we regressed VEP waveforms pairwise onto each other and calculated a t-value between all sessions. The maximal t-value was taken to indicate recognition for all 19 comparisons performed with one session. The recognition rate was cross-validated in a generalized linear regression model (GLM). RESULTS: The number of sessions correctly matched to the correct subject was 19 of 20 (true positives) leading to a sensitivity of 95.0% with confidence interval [75.1% 99.9%] for the method. The number of true negatives was 359 of 360 leading to a specificity of 99.7% [98.5% 100.0%]. CONCLUSIONS: The VEP waveform shows high intra-individual stability compared to the inter-individual variation in healthy women. SIGNIFICANCE: With a new statistical approach the effect of external factors on the VEP waveform can now be contrasted against the normal variability over time in longitudinal studies.