Cerebral Functional Connectivity Analysis Based on Scalp EEG in Epilepsy Patients

Zhongjiang Sun,Gang Wang,Kuo Li,Zhonglin Zhang,Gang Bao

DOI: https://doi.org/10.1109/bmei.2014.7002786

2014-01-01

Abstract:The pathological status of epilepsy can be revealed by the cerebral functional connectivity analysis. The majority of investigations in this field focused on the functional magnetic resonance imaging (fMRI). In consideration of the high temporal resolution of electroencephalogram (EEG) signals, a causality analysis based on the partial directed coherence (PDC) algorithms was applied to explore the cerebral functional connectivity of EEG signals from the perspective of the direction and the intensity of information flow. First of all, the multivariate autoregressive (MVAR) model was established for a moving analysis window. The PDC intensity was then calculated. According to the pathological features of epileptic seizure, the outflow information was regarded as the features of the cerebral functional connectivity based on scalp EEG. At last, the brain electrical activity mapping (BEAM) of outflow information was obtained to compare the physiological features corresponding to different moving windows. The results demonstrated that the intensity and the direction of information flow would be changed before and after the epilepsy seizure onset. The intensity of outflow information was obviously enhanced during seizure onsets and the strongest area was focused on seizure onset zone, which conforms to the clinical prediction and has clear physiological significance. The good intuitive nature and accuracy of the experimental results can provide the basis for the subsequent investigations of seizure detection based on the EEG signals.