Sleep onset estimator: evaluation of parameters

The electroencephalographic (EEG) alterations during the human sleep onset (falling asleep period) has been evaluated by several studies in the past. However, the analysis part has been limited due to standard signal processing methods. This paper has attempted to evaluate a number of advanced parameters for improved sleep onset estimation, such as EEG non-parametric coherence, power frequency and spectral band power. These parameters can be utilised in an on-line algorithm design for neurofeedback applications.