A multivariate comparison between two EEG analysis techniques: period analysis and fast Fourier transform

The present study investigated the statistical correspondence between two frequently used methods of analysing tonic EEG activity: (1) period analysis and (2) Fast Fourier Transforms (FFTs). For replication purposes, independent statistical analyses were carried out on the EEG of sleeping subjects and of awake subjects performing a cognitive task. From the results of two canonical correlations, the first two canonical variates were sufficient to account for 97% (awake group) and 99% (sleeping group) of the variance between the separate linear composites of all FFT and period analysis variables. Linear regressions of the 3 dependent measures generated by period analysis onto each respective FFT measure gave high and significant multiple correlations for all frequencies. Zero-order correlations were also gathered and discussed. It is concluded that period analytic and FFT analysed EEG share similar types of information and that period analysis can be used more often in EEG research.