This work introduces a diverse range of techniques to discern wakefulness and sleep states through intracranial electroencephalographic (iEEG) activity. Our analysis provides valuable population-level insights across various brain regions and states. We unify the exploration of brain rhythms and scale-free behavior with machine learning methodologies. Oscillatory and scale-free features serve as inputs for neural network analysis, revealing rich sleep-wake patterns within the low-dimensional latent space of variational autoencoders