[Original method of multivariate analysis, adapted for comparison of mean evoked potentials: E2 test]

Area of Science:

• Multivariate statistical analysis
• Biostatistics
• Neuroscience signal processing

Background:

• Comparing paired data series is crucial in various scientific fields.
• Existing methods like Student's t-test and generalized Hotteling T2 test have limitations for specific paired data structures.
• Need for a robust statistical method to analyze complex paired data, such as time-series measurements.

Purpose of the Study:

• To introduce and describe a novel multivariate statistical method, the E2 test.
• To establish the analogy of the E2 test to the generalized Hotteling T2 test for paired series.
• To demonstrate the application of the E2 test in analyzing biological signals, specifically auditory evoked potentials.

Main Methods:

• Development of the E2 test, a multivariate statistical method for comparing two paired series.
• Mathematical formulation establishing the E2 test's analogy to the generalized Hotteling T2 test.
• Application of the E2 test to analyze average auditory evoked potentials recorded from symmetrical scalp points.

Main Results:

• The E2 test provides a robust framework for the multivariate comparison of paired data.
• The epsilon test, a pairing method, is shown to be analogous to Student's t-test.
• Successful application demonstrated in comparing complex biological signal curves (auditory evoked potentials).

Conclusions:

• The E2 test is a valuable statistical tool for analyzing paired data, particularly in signal processing and neuroscience.
• The method offers an advancement over existing techniques for specific types of paired data comparisons.
• The demonstrated application highlights the utility of the E2 test in neurophysiological research.