Table 1.

An overview of the features used in this study, grouped by type. See the Supplementary Methods for a detailed description of all non-trivial features.

Label	Short Description	Type
F1	Signal Skewness	EEG Time Domain
F2	Signal Kurtosis
F3	Zero Crossing Rate
F4	Hjorth Mobility
F5	Hjorth Complexity
F6	75th percentile
F7	Channel Correlation
F8	EMG power	EMG proxy
F9	Minimal EMG Power
F10	Relative EMG burst amplitude
F11	Slow Eye Movement power	EOG proxy
F12	Rapid Eye Movement power
F13- F16	Relative power in α, β, θ, δ-bands	EEG Frequency Domain
F17-F23	Power-ratios: δ/θ, θ/α, α/β, β/γ, (θ + δ)/(α + β)
F24	Spectral edge frequency
F25	Median power frequency
F26	Mean spectral edge frequency difference
F27	Peak power Frequency
F28	Spectral Entropy
F29	Spindle probability	Sleep event proxies
F30	Frequency stationarity
F31	Lowest adj. frequency similarity
F32	Maximum B-spline transform
F33	Longest sleep spindle
F34-F38	Power mean in γ, α, β, θ, δ bands	CWT based features
F39-F43	Power variance in γ, α, β, θ, δ bands
F44-F48	Power skewness in γ, α, β, θ, δ bands
F49-F53	Power kurtosis in γ, α, β, θ, δ bands
F54-F58	Power entropy in γ, α, β, θ, δ bands
F59-F63	Duration of the activation in γ, α, β, θ, δ bands
F64-F68	75 percentile in γ, α, β, θ, δ bands
F69-F73	Relative power in γ, α, β, θ, δ bands
F74-F77	Power ratios: δ/θ, θ/α, α/β, β/γ
F78-F82	Multi scale entropy 1–5	Non-linear features
F83	Lempel-Ziv complexity