Identifying Refractory Epilepsy Without Structural Abnormalities by Fusing the Common Spatial Patterns of Functional...

Area of Science:

• Neuroscience
• Medical Imaging
• Epileptology

Background:

• Drug-refractory epilepsy (RE) often lacks visible structural lesions on MRI.
• Differentiating RE patients without structural abnormalities (RE-no-SA) from medically controlled epilepsy (MCE) patients is challenging.
• Accurate diagnosis is crucial to prevent misdiagnosis and ensure appropriate treatment for epilepsy.

Purpose of the Study:

• To develop a reliable method for distinguishing RE-no-SA patients from MCE patients.
• To explore the utility of spatial pattern of network (SPN) features from resting-state EEG.
• To assess the diagnostic performance of functional and effective EEG networks.

Main Methods:

• Resting-state scalp electroencephalogram (EEG) datasets were analyzed.
• Spatial pattern of network (SPN) features were extracted from functional and effective EEG networks.
• Classification performance of SPN features was compared to traditional EEG network properties.

Main Results:

• SPN features significantly outperformed traditional EEG network properties in classifying epilepsy groups.
• SPN features of functional and effective EEG networks achieved 90.00% and 80.00% accuracy, respectively.
• Fused SPN features reached a highest accuracy of 96.67% with 100% sensitivity and 92.86% specificity.

Conclusions:

• Fused functional and effective SPN features from EEG show promise for distinguishing RE-no-SA from MCE patients.
• This approach may offer a new perspective for understanding the neurophysiology of refractory epilepsy.
• SPN features represent a reliable measurement for epilepsy subtyping.