Statistical Features in High-Frequency Bands of Interictal iEEG Work Efficiently in Identifying the Seizure Onset Zone in Patients with Focal Epilepsy

Abstract

The design of a computer-aided system for identifying the seizure onset zone (SOZ) from interictal and ictal electroencephalograms (EEGs) is desired by epileptologists. This study aims to introduce the statistical features of high-frequency components (HFCs) in interictal intracranial electroencephalograms (iEEGs) to identify the possible seizure onset zone (SOZ) channels. It is known that the activity of HFCs in interictal iEEGs, including ripple and fast ripple bands, is associated with epileptic seizures. This paper proposes to decompose multi-channel interictal iEEG signals into a number of subbands. For every 20 s segment, twelve features are computed from each subband. A mutual information (MI)-based method with grid search was applied to select the most prominent bands and features. A gradient-boosting decision tree-based algorithm called LightGBM was used to score each segment of the channels and these were averaged together to achieve a final score for each channel. The possible SOZ channels were localized based on the higher value channels. The experimental results with eleven epilepsy patients were tested to observe the efficiency of the proposed design compared to the state-of-the-art methods.

Keywords: epilepsy; feature extraction; filter bank; seizure; seizure onset zone (SOZ).

Figure 1

The intracranial electroencephalogram (iEEG) signals and their magnetic resonance imaging (MRI) images with 17 electrodes for patients Pt1 and Pt6. The red circles indicate the SOZ electrodes labeled by clinical experts.

Figure 2

The allocation of iEEG data used for training and testing in the proposed method.

Figure 3

Grid parameter search for optimizing features and subbands. The F-score was obtained as a joint combination of features (y-axis) and subbands (x-axis) with maximum MI scores derived from Equations (22) and (24).

Figure 4

The simulated results using the FbFM/Sb/FS method to identify the seizure onset zone (SOZ) channels. For each patient, the color map (left) and average value of scores with channels achieved from each segment using the proposed method (right) are presented. The red color bars indicate the SOZ channels labeled by epileptologists.

Figure 5

MRI scan images with electrode positions for 5 patients (Pt1 to Pt5). The electrodes of SOZ channels are represented by “X”, labeled using clinical expertise, and the circles with color indicate the magnitude of each channel score achieved by the proposed method.

Figure 6

MRI scan images with electrode positions for 6 patients (Pt6 to Pt11). The electrodes of SOZ channels are represented by “X”, labeled using clinical expertise, and the circles with color indicate the magnitude of each channel score achieved by the proposed method.

Figure 7

Average computational time with each feature extraction method for 10 subbands (Left). Average computational time with number of subbands for feature extraction methods (Right).