Internight reliability and benchmark values for computer analyses of non-rapid eye movement (NREM) and REM EEG in normal young adult and elderly subjects

Objective: To determine the reliability of computer measured non-rapid eye movement (NREM) and REM frequency bands in the 0.3–45 Hz range and to provide benchmark data for these measures in young normal (YN) and elderly normal (EN) subjects (Ss). Methods: Sleep EEG was recorded in 19 YN and 19 EN Ss on 4 non-consecutive baseline nights and simultaneously quantified as fast Fourier transform (FFT) power and 3 zero-cross period-amplitude (PA) measures: integrated amplitude, time in band and average wave amplitude. Results: The shapes of both the FFT and PA spectra differed among Ss but were highly consistent within individuals. Inter-night reliability of the separate frequency bands was correspondingly high. Despite substantial age effects, the reliability of computer-measured sleep EEG in the elderly equaled that of the YN Ss. Within both the YN and EN groups, the shapes of the NREM and REM spectral curves differed significantly. The NREM and REM also differed significantly in the two age groups. Conclusions: Computer-measured sleep EEG is highly reliable across non-consecutive nights in both young and elderly normal Ss. The trait-like stability of these measures suggests they are genetically determined. This possibility is supported by twin study data that show strong heritability for FFT-measured waking EEG. The different shapes of NREM and REM spectra add further evidence that these are fundamentally different states of brain organization. The age differences in spectral shape, along with PA data for wave incidence, demonstrate that age effects on sleep EEG are not caused by changes in skull impedance or other non-cerebral factors.