The onset of cortical activity evoked by a visual stimulus, the early latency (fL), can be measured by the use of multifocal visual evoked potential (mfVEP). In this study we investigated differences in amplitude and fL between pattern-reversal and flash mfVEP.The mfVEP was recorded from 20 subjects using flash and pattern-reversal stimuli. The recordings were done using a two-dimensional Laplacian derivation. After the signals from all 60 stimulus fields were squared and averaged, the sudden rise from baseline activity marked the beginning of evoked cortical activity, which defined the fL.The fL was 48.1+/-1.0 ms for pattern-reversal mfVEP and 52.2+/-1.3 ms for flash mfVEP. The maximum normalized amplitude was 22.5+/-2.8 for pattern-reversal mfVEP and 8.9+/-1.5 for flash mfVEP. Analysis of variance showed that the stimulus type had a significant effect on the fL (p< or =0.01) and on the maximum amplitude (p< or =0.01) of the evoked cortical potentials.The fL can be estimated more accurately with potentials of high amplitude. Because the maximum amplitude for pattern-reversal mfVEP was three times as high as for flash mfVEP, and because the fL was longer for flash mfVEP than for pattern-reversal mfVEP, the results of this study do not imply that flash mfVEP should be used for evaluating fL.