Estimating vigilance from EEG using manifold clustering guided by instantaneous lapse rate

摘要: Vigilance decrement happens in prolonged and monotonous tasks such as driving, therefore efficient estimation of vigilance using machine learning becomes a growing research field road safety. However, the ground truth level is often unknown. To address brain states with unknown truth, we proposed an unsupervised manifold clustering method guided by task performance, namely instantaneous lapse rate, without directly any artificially labels, electroencephalogram (EEG) data source. The algorithm utilizes information from both structure which especially suitable for applications truth. Future directions include advanced algorithms to increase robustness towards high nonlinearity EEG feature space embedded space, well allowing mapping multiple clusters one level.