Abstracts

Rationale: Non-invasive electrophysiological source imaging (ESI) provides valuable information about epileptic sources, and has been used to aid pre-surgical planning [1]. It has been suggested that scalp EEG spikes, HFOs and seizures have varied ability to estimate the epileptogenic zone (EZ). However, it remains unclear about how much they differ quantitatively, particularly regarding whether subgroups of these biomarkers provide distinct source imaging capabilities. In this study, we developed a spatial-temporal-spectral source imaging framework (STSI) that delineates and images the EEG signal considering space, time and frequency. Next, we investigated the source imaging capability of spikes, HFO and seizures in a group of focal drug-resistant epilepsy (fDRE) patients, and compared it with the surgical resection as clinical ground truth.


Methods: Components representing the spatial, temporal and spectral features of the EEG signal was fed into an optimization problem to obtain the inverse solution on the cortex. Twelve fDRE patients (ILAE outcome: I-II), with 36 pHFO and pSpike events (spikes with concurrent HFO overriding on it) [2], 68 aHFO events (all HFO events), 110 aSpikes (all spikes) events, and 18 seizure events were included in this study. Patient-specific head models and surgical resection were generated based on pre- and post-operative MRI scans. The evaluation metrics used include localization error (LE), spatial dispersion (SD), precision, sensitivity, and specificity.


Results: The STSI framework could accurately estimate the location and extent of biomarkers with different frequencies, except for aHFO. Figure 1 shows the group results. Averaged metrics for pHFO, pSpikes, aSpikes, seizures and aHFO are as follows, LE: 4.13, 5.54, 9.41, 5.11 20.76 mm; SD: 4.71, 6.50, 11.28, 5.84, 22.71 mm. Notably, within the aHFO group, pHFO in general have good performance, while the performance of the rest of isolated HFO have large variability, with some aHFO have poor performance but the others pointing to the ground truth. This phenomenon might be due to the fact that aHFO consists of both pathological and physiological HFO events.

Conclusions: Our results in this patient cohort suggest that pHFO and seizure were among the most accurate biomarkers, followed by pSpikes and then aSpikes. Our results suggested the feasibility of imaging the location and extent of epileptic sources using EEG spectral contents, and offered quantitative evaluation of the source imaging capabilities of spikes, HFOs and seizures.

References:
1. He B., Sohrabpour A., Brown E., and Liu Z., 2018. Electrophysiological Source Imaging: A Noninvasive Window to Brain Dynamics. Ann Review of Biomedical Engineering, 20, 171-196, 2018.
2. Cai, Z., Sohrabpour, A., Jiang, H., Ye, S., Joseph, B., Brinkmann, B.H., Worrell, G.A. and He, B., 2021. Noninvasive high-frequency oscillations riding spikes delineates epileptogenic sources. PNAS, 118(17), p.e2011130118.






Funding: NIH NS096761 and NS127849.