IEEE trans: Functional EEG network to identify refractory epilepsy without structural abnormalities

Drug refractorymedsci.

In this study, this paper collected the resting state EEG signals of 14 patients with MCE and 16 patients with RE-no-SA, and evaluated the functional network (linear estimation of non-directional connectivity-frequency domain coherence) and effective network (direction).

Method for extracting SPN characteristics of RE-no-SA and MCE patients from EEG brain segments in resting state

First, a functional and effective network of different frequency bands is constructed from the EEG signal.

The classification performance based on SPN features shows that the difference between RE-no-SA patients and MCE patients lies in the functional and effective networks.

RE-no-SA patients and MCE patients have significant differences in the connection strength of functional and effective networks (p<0.

Theoretically, the SPN features extracted by the first pair of SPN filters have the most distinguishing ability for the two groups of epilepsy patients.

In summary, this article proposes to use the spatial features of fusion of functionality and effective connectivity to distinguish RE-no-SA patients from MCE patients.

Y.