Combining time series and frequency domain analysis for a automatic seizure detection

Abstract

The detection of epileptic seizures in long-term electroencephalographic (EEG) recordings is a time-consuming and tedious task requiring specially trained medical experts. The EpiScan seizure detection algorithm developed by the Austrian Institute of Technology (AIT) has proven to achieve high detection performance with a robust false alarm rate in the clinical setting. This paper introduces a novel time domain method for detection of epileptic seizure patterns with focus on irregular and distorted rhythmic activity. The method scans the EEG for sequences of similar epileptiform discharges and uses a combination of duration and similarity measure to decide for a seizure. The resulting method was tested on an EEG database with 275 patients including over 22000h of unselected and uncut EEG recording and 623 seizures. Used in combination with the EpiScan algorithm we increased the overall sensitivity from 70% to 73% while reducing the false alarm rate from 0.33 to 0.30 alarms per hour.

Prospective multi-center study of an automatic online seizure detection system for epilepsy monitoring units

“…This method was developed over several years by a team of physicians, mathematicians and medical experts (Schachinger et al, 2006;Perko et al, 2007;Kluge et al, 2009;Hartmann et al, 2011;Fürbass et al, 2012). It is intended to analyze the EEG ad-hoc and to act as an online detection system.…”

“…This will avoid false alarms based on measurement problems. The EEG is then scanned for rhythmic patterns in the time and frequency domain by algorithms called Epileptiform Wave Sequence Analysis (EWS) and Periodic Waveform Analysis (PWA), respectively Fürbass et al, 2012). An energy detector scans for tonic or tonic-clonic seizures with strong muscle artifacts.…”

A retrospective glance at automatic detection of epileptic spike in electroencephalogram

“…It is gratifying to see that a majority of the researches in Table 1 (20 out of 29) have been able to acquire certain amount of patients (over 10) for analysis. Fürbass's work involves the most patients (275) (65). By applying a periodic waveform analysis with focus on irregular and distorted rhythmic activities, Fürbass achieves 83.6% sensitivity and false alarm rate at …”

“…Yet lacking of uniform dataset and evaluation rules makes the comparison of different strategies impossible. In Table 1, the least number of patients involved in study is one, while the largest is 275 (13,65). In three cases, the data is labeled by samples rather than by patients (9,10,19).…”

Automatic optimization of parameters for seizure detection systems

“…The algorithm is based on a frequency domain method called Periodic Waveform Analysis (PWA) and the time domain analysis of epileptiform sequences (EWS) [2]. The system convinces with a very good detection performance and no need of patient dependent parameter adjustment.…”