Data Analysis Techniques in Average Evoked Potential Research

From the proceedings of a conference sponsored by the National Aeronautics and Space Administration and the American Institute for Biological Sciences. The conference was held in San Francisco in September 1968 to discuss current problems in the study of average evoked potential. The purpose of most studies of the average evoked potential (AEP) is to determine the extent to which the complex waveform of the AEP varies with the parameters of the stimulation, the state of the subject, and the recording site. Data analysis techniques in AEP research should, therefore, provide for reliable, objective, and easy-to use methods for measuring and specifying differences between any two AEP. Statistical analysis provides us with a body of tools whose purpose is to allow the investigator to evaluate and judge differences of this nature. However, there are two major difficulties in applying classical statistical analysis to AEP data, namely, the fact that the AEP is a multidimensional observation and that more often than not the format in which the data are available is that of the graphic output of an averaging device. The multidimensionality of the AEP makes it insufficient to state that two AEPs are different ; this statement must. always be supplemented by a more detailed specification of the difference. Thus, two AEPs might have identical waveshapes but different amplitudes; the difference might be limited to a small segment of the AEP, or the two AEPs might be virtually identical except for a major difference in one or more components. The differences might be caused by changes in the general waveform, or they might be caused by shifts in the latency of specific components. For these reasons, general statements that two AEP waveforms are different are relatively devoid of meaning