The seizure prediction characteristic: a new terminology and assessment criterion for epileptic seizure prediction methods

Epilepsy is characterized by the sudden occurrence of seizures disturbing the perception or behavior of epileptic patients. Several prediction methods have claimed to be able to predict seizures based on EEG-recordings minutes in advance, which opens up new approaches to treat the disease. However, the term seizure prediction is not unequivocally defined and different assessment criteria for prediction methods exist which impedes the comparison between methods. Moreover, only little attention is paid to the dependency between sensitivity and false prediction rate. We address these shortcomings and introduce a terminology and assessment criterion for seizure prediction methods based on statistical and clinical considerations: the seizure prediction characteristic.