Do QEEG topographic map data have a multivariate normal distribution? Implications for tests of individual maps

Area of Science:

• Neuroscience
• Biostatistics
• Medical Imaging

Background:

• Standard statistical methods in neuroscience often assume multivariate normality (MVN) for topographic map data.
• Assessing significant changes in brain maps requires understanding the underlying data distribution.
• The 10/20 electrode configuration is a common standard in electroencephalography (EEG) and magnetoencephalography (MEG).

Purpose of the Study:

• To investigate if topographic map data, when expressed in log power units, adhere to a multivariate normal (MVN) distribution in a healthy population.
• To determine if deviations from MVN distribution significantly affect the reliability of standard multivariate statistical tests for analyzing topographic map deformations.
• To assess the practical implications of non-normality for individual case analysis.

Main Methods:

• Log-transformed power data from 361 healthy adults were analyzed across six frequency bands using 20 electrodes in the 10/20 configuration.
• The multivariate normality of the log-transformed power data was assessed.
• The 'jackknife' resampling method was employed to compute the actual sampling distributions of multivariate tests.
• These empirical distributions were compared to theoretical F-distributions expected for MVN data.

Main Results:

• Log-transformed topographic map power data significantly deviated from the multivariate normal (MVN) distribution.
• The empirical sampling distributions derived from the 'jackknife' method markedly differed from the F-distributions expected under MVN assumptions.
• These findings demonstrate that standard statistical tests may not be appropriate for this type of neurophysiological data.

Conclusions:

• Topographic map data in log power units do not conform to a multivariate normal (MVN) distribution in healthy individuals.
• Deviations from MVN necessitate caution when applying standard multivariate statistical tests to analyze topographic map deformations.
• The study successfully demonstrated significant topographic map deformations in a traumatic brain injury patient using the derived 'jackknife' sampling distributions, highlighting the clinical relevance of accounting for non-normality.