According to current studies, 18% of all patients with serious brain diseases in intensive care units suffer from undetected non-convulsive epileptic seizures, which lead to a significantly increased mortality or to permanent neurological disabilities. They are not clinically observable and can only be detected in a reliable way through long-term measurements of the EEG. In the clinical practice however long-term measurements are only available in exceptional cases with this patient group. The reason for that is above all the time-consuming visual analysis of the huge amount of EEG signals, which additionally necessitates the expertise of an EEG analysis experienced physician. The deficient EEG monitoring of these patients presents hence a clinically relevant and to a large extent unresolved problem.
A computer based, continuous and automatic evaluation of the EEG signalsprovides a significant contribution here. It renders possible the quick and automatic alarm in case of an epileptic seizure, so that an expert can be consulted immediately. At the AIT highly reliably algorithms for seizure detection for patients with serious brain diseases are developed in close cooperation with the leading hospitals.