P3 latency determination by global field power in normal subjects.

The usefulness for assessing the peak latency of the P3 component of the cognitive event-related potential by means of global field power (GFP) measurements was investigated in 25 normal individuals. An initial test of the program utilizing the visual evoked potential elicited by checkerboard pattern reversals reproduced the findings originally reported by Lehmann and Skrandies demonstrating that the method is technically sound. Grand averages of the cognitive event-related potential showed on GFP determinations two peaks representing the N1 and P3 components, respectively. Correlation of P3 GFP latency with conventional peak determination showed that it was somewhat more closely related to the peak as seen on average reference recordings than when the linked ears were used as reference. Filtering the data to eliminate the intrusion of frequencies above 7 Hz shifted the curves somewhat to the right and correlation coefficients between GFP and conventional peak determinations decreased. GFP can aid in establishing the P3 peak latency, especially when two peaks are present in conventional recordings, e.g., at PZ and CZ with somewhat different latencies, but they cannot be substituted for conventional peak determinations, since the P3 may either show up only as a notch or not at all on the GFP curve, and the major GFP peak may not be related to the P3 component. The large standard deviation for the P3 component on conventional tracings is not reduced by GFP determinations.