The reliability of P50 suppression as measured by the conditioning/testing ratio is vastly improved by dipole modeling.

Suppression of auditory P50 evoked potential amplitude to the second of a pair of clicks is potentially important in psychiatric research because it has been shown to be abnormal in both schizophrenics and their relatives. However, its clinical utility using the standard single-channel electroencephalographic (EEG) peak picking methodology is under question because of low test-retest reliability. Dipole Components Modeling of the P50 component was attempted as a method for increasing the reliability of the P50 suppression measure. It was hypothesized that this procedure might work because of pooling of noise from the two responses and because of the use of topographic information. Six replications of a P50 suppression paradigm in 12 subjects were analyzed. Reliability using peak picking was 0.27, and was significantly increased to 0.63 using dipole modeling. Dipole modeling was helpful not only for better modeling the P50 when it was present, but also for deciding that there was no P50 response in one subject.