A patient-specific seizure prediction algorithm using time/space-differential ECoG features achieved 86.25% sensitivity and 0.1281 false positives per hour in out-of-sample testing.
Does a patient-specific seizure prediction algorithm using time/space-differential ECoG signals and cost-sensitive SVMs accurately predict seizures in patients with epilepsy?
A patient-specific seizure prediction algorithm using time/space-differential ECoG signals and cost-sensitive SVMs achieved 86.25% sensitivity and low false positive rates.
A patient-specific seizure prediction algorithm is proposed using a classifier to differentiate preictal from interictal ECoG signals. Spectral power of ECoG processed in four different fashions are used as features: raw, time-differential, space-differential, and time/space-differential ECoG. The features are classified using cost-sensitive support vector machines by the double cross-validation methodology. The proposed algorithm has been applied to ECoG recordings of 18 patients in the Freiburg EEG database, totaling 80 seizures and 437-hour-long interictal recordings. Classification with the feature obtained from time/space-differential ECoG demonstrates performance of 86.25% sensitivity and 0.1281 false positives per hour in out-of-sample testing.
Park et al. (Fri,) conducted a other in Seizures (n=18). Patient-specific seizure prediction algorithm using cost-sensitive support vector machine was evaluated on Sensitivity and false positives per hour. A patient-specific seizure prediction algorithm using time/space-differential ECoG features achieved 86.25% sensitivity and 0.1281 false positives per hour in out-of-sample testing.