Artificial neural network for detecting drowsiness from EEG recordings

We describe a novel method for classifying alert vs. drowsy states from one-second long sequences of full spectrum EEG recordings. This method uses time series of inter-hemispeheric and intra-hemispheric cross spectral densities of full spectrum EEG as input to an artificial neural network (ANN) with two discrete outputs: drowsy and alert. The experimental data were collected from 17 subjects. Two experts in EEG interpretation visually inspected the data and provided the necessary expertise for the training of an ANN. After several experiments we selected the learning vector quantization (LVQ) as the most suitable neural network and used the data from 5 subjects for the training. Classification properties of LVQ were validated using the data recorded from the remaining 12 subjects, whose EEG recordings have not been used for the training of the ANN. The statistics were used as a measure of potential applicability of the LVQ: the t-distribution showed that in 95% (confidence interval) of the target group the matching between the human assessment and the network output was 94, 37/spl plusmn/1.95 percent.

[1]  G Pfurtscheller,et al.  EEG Classification by Learning Vector Quantization - EEG-Klassifikation mit Hilfe eines Learning Vector Quantizers , 1992, Biomedizinische Technik. Biomedical engineering.

[2]  P. Achermann,et al.  A new method for detecting state changes in the EEG: exploratory application to sleep data , 1998, Journal of sleep research.

[3]  Andrew C. N. Chen,et al.  Automatic recognition of alertness and drowsiness from EEG by an artificial neural network. , 2002, Medical engineering & physics.

[4]  Teuvo Kohonen,et al.  Learning Vector Quantisation and the Self Organising Map , 1992 .

[5]  B. J. Wilson,et al.  Alertness monitor using neural networks for EEG analysis , 2000, Neural Networks for Signal Processing X. Proceedings of the 2000 IEEE Signal Processing Society Workshop (Cat. No.00TH8501).

[6]  Y. Wada,et al.  Inter- and Intrahemispheric EEG Coherence during Light Drowsiness , 1996, Clinical EEG.

[7]  A. Gevins,et al.  Detecting transient cognitive impairment with EEG pattern recognition methods. , 1999, Aviation, space, and environmental medicine.

[8]  Gert Pfurtscheller,et al.  Automatic differentiation of multichannel EEG signals , 2001, IEEE Transactions on Biomedical Engineering.

[9]  Georg Dorffner,et al.  Using Selforganizing Feature Maps to Classify EEG Coherence Maps , 1993 .

[10]  T. Jung,et al.  Tonic, phasic, and transient EEG correlates of auditory awareness in drowsiness. , 1996, Brain research. Cognitive brain research.

[11]  G Pfurtscheller,et al.  Sleep Classification in Infants Based on Artificial Neural Networks. Schlafklassifikation mit Hilfe neuronaler Netzwerke , 1992, Biomedizinische Technik. Biomedical engineering.

[12]  T. Sejnowski,et al.  Estimating alertness from the EEG power spectrum , 1997, IEEE Transactions on Biomedical Engineering.

[13]  G Pfurtscheller,et al.  Using time-dependent neural networks for EEG classification. , 2000, IEEE transactions on rehabilitation engineering : a publication of the IEEE Engineering in Medicine and Biology Society.

[14]  G. Pfurtscheller,et al.  On-line EEG classification during externally-paced hand movements using a neural network-based classifier. , 1996, Electroencephalography and clinical neurophysiology.

[15]  J.C. Principe,et al.  Sleep staging automaton based on the theory of evidence , 1989, IEEE Transactions on Biomedical Engineering.