Noninvasive localization of epileptogenic zones by MEG and EEG data of patients with epilepsy

Epilepsy is a neurological disorder affecting at least 50 million people and characterized by regular seizures. Approximately 30% of patients with epilepsy are not susceptible to pharmacological intervention. In such cases, surgical resection of the epileptogenic zones that produce epileptiform activity is necessary. Non-invasive localization of epileptogenic zones is possible based on the use of such EEG/MEG marker of epilepsy as interconvulsive discharges (interictal spikes) - acute patterns of specific morphology occurring between seizures.


Fig. 1. Kleeva, D., Soghoyan, G., Komoltsev, I., Sinkin, M., & Ossadtchi, A. (2022). Fast parametric curve matching (FPCM) for automatic spike detection. Journal of Neural Engineering, 19(3), 036003.

The entire analysis of EEG/MEG data performed to search for epileptogenic zones can be divided into the following steps:

1. Automatic detection of interconvulsive discharges;
2. Localization of detectable discharges in the cortical source space;
3. Dynamic analysis of the activity of detected epileptogenic clusters.

Within each of the stages we use a specially developed arsenal of tools. In particular, for automatic spike detection, we use the FCPM method (Kleeva et al., 2022), a mimetic algorithm in which the shape of the interstitial discharge is parameterized by two linear segments and a parabola, the resulting morphological model is fitted to the data, and direct detection is based on logical predicates corresponding to the criteria of the visual search for discharges performed by clinicians. An important feature of this algorithm is that, while maintaining its sensitivity to genuine discharges, it is robust to high-amplitude artifacts.


Fig. 2. Kleeva, D., Soghoyan, G., Komoltsev, I., Sinkin, M., & Ossadtchi, A. (2022). Fast parametric curve matching (FPCM) for automatic spike detection. Journal of Neural Engineering, 19(3), 036003.

The detected discharges are then used to localize in source space using dipole matching or distributive inverse modeling methods characterized by focal solutions (e.g., LCMV). The localization results in clusters characterized by different spatial properties and activation peculiarities (Ossadtchi et al., 2004). These data form the basis for the final stage of the analysis, i.e., the evaluation of the patterns of epileptogenic activity propagation between the detected clusters. In particular, this is done by means of hidden Markov models (Ossadtchi et al., 2005). In the current stages of work in this area, we are comparing the characteristics of MEG and EEG data recorded simultaneously in patients with epilepsy. The complementary nature of these noninvasive neuroimaging techniques will potentially expand the informative value of the current approach to analysis.

Published articles on the project:

1) Kleeva, D., Soghoyan, G., Komoltsev, I., Sinkin, M., & Ossadtchi, A. (2022). Fast parametric curve matching (FPCM) for automatic spike detection. Journal of Neural Engineering, 19(3), 036003.

2) Chirkov, V., Kryuchkova, A., Koptelova, A., Stroganova, T., Kuznetsova, A., Kleeva, D., ... & Fedele, T. (2022). Data-driven approach for the delineation of the irritative zone in epilepsy in MEG. Plos one, 17(10), e0275063.

3) Ossadtchi, A., Baillet, S., Mosher, J. C., Thyerlei, D., Sutherling, W., & Leahy, R. M. (2004). Automated interictal spike detection and source localization in magnetoencephalography using independent components analysis and spatio-temporal clustering. Clinical Neurophysiology, 115(3), 508-522.

4) Ossadtchi, A., Mosher, J. C., Sutherling, W. W., Greenblatt, R. E., & Leahy, R. M. (2005). Hidden Markov modelling of spike propagation from interictal MEG data. Physics in Medicine & Biology, 50(14), 3447.