Neural network detection of antiepileptic drugs from a single EEG trace

Three systems-a Bayes quadratic classifier (BQ), a self-organizing polynomial network (PN), and a backpropagation neural network (NN)-were created for subject-independent classification into either no-drug or antiepileptic drug categories from a single 2.1-second EEG trace at the Pz/A1+A2 channel. Spectral features were derived from the raw signals, and then preprocessed using singular value decomposition prior to their classification. Recognition rates on an independent data set were 63.5%, 60.4%, and 74.0% for the BQ, PN, and NN respectively.<<ETX>>

[1]  D W King,et al.  Effects of Carbamazepine and Phenytoin on EEG and Memory in Healthy Adults , 1993, Epilepsia.

[2]  E. John,et al.  Neurometrics: computer-assisted differential diagnosis of brain dysfunctions. , 1988, Science.

[3]  Z. Keirn,et al.  A new mode of communication between man and his surroundings , 1990, IEEE Transactions on Biomedical Engineering.

[4]  D.R. Hush,et al.  Progress in supervised neural networks , 1993, IEEE Signal Processing Magazine.

[5]  Roger L. Barron,et al.  Applications of polynomial neural networks to FDIE and reconfigurable flight control , 1990, IEEE Conference on Aerospace and Electronics.