Rationale:
It is of importance to accurately localize epileptogenic zone guiding presurgical planning. We aim to develop noninvasive techniques that can localize epileptogenic zone from noninvasive high density EEG recordings. In this study, we show that a novel electrophysiological source imaging algorithm - fast spatiotemporal iteratively reweighted edge sparsity (FAST-IRES) [1] - is able to accurately estimate the location and extent of epileptic sources from pathological spikes (pSpikes) – interictal spikes overriding high-frequency oscillations (HFOs) [2]. The source imaging analysis results are validated by surgical resection volume in a cohort of focal drug-resistant epilepsy patients who were seizure free at least one year post-op follow up.
Methods:
39 pSpikes from 75-channel scalp EEG were analyzed in a cohort of 14 focal epilepsy patients. All patients received surgical resection and were declared seizure-free (ILAE I) based on at least one year post-op follow up. Surgical resection regions were determined by comparing each patient’s pre-op and post-op MRI. Localization error (LE), spatial dispersion (SD), precision, recall, and specificity of source imaging results were calculated with respect to the surgical resection region. FAST-IRES results were also compared to sLORETA [3] results as a benchmark.
Results:
As shown in Figure 1, the FAST-IRES estimations are concordant with the resection region. The value for each metric is as follows: 4.9 ± 3.2 mm LE, 5.9 ± 5.8 mm SD, 0.57 ± 0.33 precision, 0.59 ± 0.28 recall, and 0.96 ± 0.04 specificity. All values are reported as mean ± standard deviation. For sLORETA, the mean values are 9.2 ± 8.2 mm LE, 16.0 ± 15.2 mm SD, 0.49 ± 0.35 precision, 0.55 ± 0.29 recall, and 0.93 ± 0.06 specificity. FAST-IRES performed significantly better than sLORETA in LE, SD, and specificity.
Conclusions:
The present results demonstrate the merits of noninvasive EEG source localization, and that 1) pSpike is an effective biomarker to be used for localization of epileptic sources; and 2) FAST-IRES provides excellent performance that is highly concordant with successful surgical resection outcome. Further investigation should be performed to compare the performance of source localization from pSpikes vs. other spikes, and investigate the performance in patients with multiple types of spikes.
References:
1 A. Sohrabpour, Z. Cai, S. Ye, B. Brinkmann, G. Worrell, and B. He, “Noninvasive electromagnetic source imaging of spatiotemporally distributed epileptogenic brain sources,”
Nat Commun, 11, 1946, 2020.
2 Z. Cai, A. Sohrabpour, H. Jiang, S. Ye, B. Joseph, B. Brinkmann, G. Worrell, B. He, “Noninvasive high-frequency oscillations riding spikes delineates epileptogenic sources,”
Proceedings of the National Academy of Sciences, 118, e2011130118, 2021.
3 R. D. Pascual-Marqui, “Standardized low-resolution brain electromagnetic tomography (sLORETA): technical details,”
Methods Find Exp Clin Pharmacol, 24 Suppl D, 5–12, 2002.
Funding: NIH R01NS096761, T32 EB029365