Visual and computer-based analysis of 24 h sleep-waking patterns in the dog.

Sleep and waking EEG patterns recorded over 24 h periods in 7 adult beagle dogs were studied both visually and by computer. Online analysis was carried out using a mini-computer. The following quantitative data were computed and listed (and/or stored on disc) at 30 sec intervals: power spectrum analysis of one cortical derivation by means of Fast Fourier Transformation, the spectrum being divided in 4 bands (0.5-3.5, 3.5-7.5, 7.5-13.5, 13.5-25 c/sec); the power contained in the 3.5-7.5 c/sec band of the hippocampal derivation; a spindle detection algorithm empirically defined and based on the power contained in the 10.5-14.5 c/sec band of subepochs of 0.8 sec and the background power; average EMG and EOG amplitudes. Plots were made of the quantitative data (on the y axis) showing evolution with time (on the x axis). Comparisons with the y ordinates enabled a visual estimate of stage to be made. Automatic classification was done by means of minimal distance, using visually selected epochs and calculated parameters as references. Visual and computer-based analyses allowed the following patterns to be defined: wakefulness, transitional stage, light slow wave sleep, deep slow wave sleep and REM sleep. Sleep-waking cycles of 20-30 min alternated during the 24 h period with rapid transitions from one stage to another.