Seizure prediction from EEG signal analysis in temporal lobe epilepsy

Epilepsy is a brain disorder characterized by abrupt and recurrent seizures caused by sudden and brief changes in the brain conditions. Affecting approximately one per cent of people worldwide, epilepsy results in an increased chance of accidental injury and death, and a decreased quality of life. Drug therapy is not always effective in controlling the recurrence of seizures, especially with temporal lobe epilepsy (TLE). The toxicities of these drugs and frequent resistance of TLE to drugs greatly decrease quality of life for patients. Therefore, it is important to investigate new techniques for the prediction of impending seizures to facilitate prompt therapy. Dr. Reza Tafreshi’s PhD work in mechanical engineering involved using statistical pattern recognition to detect and diagnose engine faults. Now he is using this knowledge to predict epileptic seizures by employing computer algorithms and analyzing brain electrical activity through scalp EEG recordings. Predicting seizure onset by a few seconds would give patients a chance to remove themselves from dangerous circumstances and allow administration of a short-acting anticonvulsant drug in a dose that would prevent the seizure. This procedure could be employed in conjunction with an advisory system to warn patients of impending seizures, leading to increased safety and better quality of life.