Resting-state EEG network markers differentiate people with epilepsy and functional seizures

Background: Distinguishing epilepsy from functional/dissociative seizures (FDS) is an ongoing diagnostic challenge. Using a well-controlled clinical EEG dataset, we run the first diagnostic accuracy study assessing the potential of resting-state EEG network markers to directly discriminate between the two conditions at the time when a diagnosis is suspected and prior to treatment initiation. Methods: The dataset, previously examined in a published study, includes 148 age- and sex-matched individuals with suspected seizure disorder, later diagnosed with non-lesional epilepsy (n=75) or FDS (n=73). Functional network measures in the 6-9 Hz range were extracted from normal-looking, eyes-closed resting-state EEG data acquired while participants were medication-free. Machine learning was implemented to assess their predictive potential; different model configurations were tested to identify the most promising approach for translational implementations. Results: EEG-derived network measures discriminate between conditions at levels significantly above chance (maximum balanced accuracy: 65.8%; SD: 2.9). Their sensitivity to epilepsy (78.4%; SD: 5.8)) is higher than their sensitivity to FDS (53.1%; SD: 4.0). Multiple nonlinear models succeed on the classification problem, but model choice remains a determinant of overall accuracy. Conclusion: We establish evidence for the clinical validity of selected network-based markers to discriminate between a diagnosis of non-lesional epilepsy and FDS prior to treatment initiation, highlighting the measures potential to support post-test probability estimation in the clinic. These measures are more specific to epilepsy than FDS and should not be interpreted as markers of a positive diagnosis of FDS.