Comparing SVM and convolutional networks for epileptic seizure prediction from intracranial EEG

Recent research suggests that electrophysiological changes develop minutes to hours before the actual clinical onset in focal epileptic seizures. Seizure prediction is a major field of neurological research, enabled by statistical analysis methods applied to features derived from intracranial Electroencephalographic (EEG) recordings of brain activity. However, no reliable seizure prediction method is ready for clinical applications. In this study, we use modern machine learning techniques to predict seizures from a number of features proposed in the literature. We concentrate on aggregated features that encode the relationship between pairs of EEG channels, such as cross-correlation, nonlinear interdependence, difference of Lyapunov exponents and wavelet analysis-based synchrony such as phase locking. We compare L1-regularized logistic regression, convolutional networks, and support vector machines. Results are reported on the standard Freiburg EEG dataset which contains data from 21 patients suffering from medically intractable focal epilepsy. For each patient, at least one method predicts 100% of the seizures on average 60 minutes before the onset, with no false alarm. Possible future applications include implantable devices capable of warning the patient of an upcoming seizure as well as implanted drug-delivery devices.

[1]  L. Bottou.  Stochastic Gradient Learning in Neural Networks , 1991 .

[2]  P. Rajna,et al.  Hungarian multicentre epidemiologic study of the warning and initial symptoms (prodrome, aura) of epileptic seizures , 1997, Seizure.

[3]  Yoshua Bengio,et al.  Gradient-based learning applied to document recognition , 1998, Proc. IEEE.

[4]  P. Grassberger,et al.  A robust method for detecting interdependences: application to intracranially recorded EEG , 1999, chao-dyn/9907013.

[5]  P. Kwan,et al.  Early identification of refractory epilepsy. , 2000, The New England journal of medicine.

[6]  J. Martinerie,et al.  Comparison of Hilbert transform and wavelet methods for the analysis of neuronal synchrony , 2001, Journal of Neuroscience Methods.

[7]  J. Martinerie,et al.  Toward a Neurodynamical Understanding of Ictogenesis , 2003, Epilepsia.

[8]  Klaus Lehnertz,et al.  Testing the null hypothesis of the nonexistence of a preseizure state. , 2003, Physical review. E, Statistical, nonlinear, and soft matter physics.

[9]  A. Schulze-Bonhage,et al.  How well can epileptic seizures be predicted? An evaluation of a nonlinear method. , 2003, Brain : a journal of neurology.

[10]  A. Schulze-Bonhage,et al.  The seizure prediction characteristic: a general framework to assess and compare seizure prediction methods , 2003, Epilepsy & Behavior.

[11]  Brian Litt,et al.  Epileptic seizure prediction using hybrid feature selection over multiple intracranial EEG electrode contacts: a report of four patients , 2003, IEEE Transactions on Biomedical Engineering.

[12]  Corinna Cortes,et al.  Support-Vector Networks , 1995, Machine Learning.

[13]  Arnaud Delorme,et al.  EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis , 2004, Journal of Neuroscience Methods.

[14]  A. Schulze-Bonhage,et al.  Comparison of three nonlinear seizure prediction methods by means of the seizure prediction characteristic , 2004 .

[15]  K. Tsakalis,et al.  Long-term prospective on-line real-time seizure prediction , 2005, Clinical Neurophysiology.

[16]  S. Schiff,et al.  Multivariate linear discrimination of seizures , 2005, Clinical Neurophysiology.

[17]  A. Kraskov,et al.  On the predictability of epileptic seizures , 2005, Clinical Neurophysiology.

[18]  K. Lehnertz,et al.  The First International Collaborative Workshop on Seizure Prediction: summary and data description , 2005, Clinical Neurophysiology.

[19]  J. Martinerie,et al.  Preictal state identification by synchronization changes in long-term intracranial EEG recordings , 2005, Clinical Neurophysiology.

[20]  Andreas Schulze-Bonhage,et al.  Testing statistical significance of multivariate time series analysis techniques for epileptic seizure prediction. , 2006, Chaos.

[21]  F. Mormann,et al.  Seizure anticipation: from algorithms to clinical practice , 2006, Current opinion in neurology.

[22]  A. Schulze-Bonhage,et al.  Do False Predictions of Seizures Depend on the State of Vigilance? A Report from Two Seizure‐Prediction Methods and Proposed Remedies , 2006, Epilepsia.

[23]  A. Schulze-Bonhage,et al.  Seizure anticipation by patients with focal and generalized epilepsy: A multicentre assessment of premonitory symptoms , 2006, Epilepsy Research.

[24]  Yann LeCun,et al.  Time-Delay Neural Networks and Independent Component Analysis for EEG-Based Prediction of Epileptic Seizures Propagation , 2007, AAAI.

[25]  F. Mormann,et al.  State-of-the-Art of Seizure Prediction , 2007, Journal of clinical neurophysiology : official publication of the American Electroencephalographic Society.