Machine-learning applications for differentiation across states/stages of creative thinking based on time-series and time-frequency features of electroencephalography signals and event-related potentials