Authors :
Presenting Author: Sofia Kasradze, MD PhD – Institute of Neurology and Neuropsychology
Giorgi Lomidze, MD PhD – European University, Tbilisi, Georgia
Lasha Ephremidze, PhD DMath – Kutaisi International University, Kutaisi, Georgia
Mukesh Dhamala, PhD Prof – Neuroscience Institute, Physics and Astronomy, Math & Statistics, Georgia State University, Atlanta, USA.
Rationale:
Accurate localization of epileptic seizure onset zones (SOZs) and their propagation patterns is crucial for successful epilepsy surgery and other treatments, such as laser ablation, focal stimulation, and gene therapy. This process typically relies on information obtained from invasive electrophysiological recordings, such as intracranial EEGs. Previous research has shown that analyzing information flow patterns, particularly in high-frequency oscillations ( >80 Hz) using parametric and Wilson-algorithm (WL) based nonparametric Granger Causality (GC), is valuable for identifying SOZs. In this study, we analyzed scalp EEG recordings from epilepsy patients using an alternative nonparametric GC, which relies on spectral density matrix factorization based on the Janashia-Lagvilava algorithm (JLA). Our findings demonstrate that this approach can accurately localize SOZs even from noninvasive EEG recordings in two pilot study cases. These preliminary results highlight an exciting potential for noninvasive brain recordings in the diagnosis and treatment planning of epilepsy patients.
Goal of our study was to assess the effectiveness of JLA-based matrix factorization in nonparametric Granger causality for identifying seizure onset zones from ictal EEG recordings in drug-resistant epilepsy patients.Methods: Two regions (
X and
Y) of interest in pairs across different time epochs were isolated in twelve people referred for presurgical evaluation. To apply the nonparametric Granger causality (GC) estimation approach to the EEG recordings from these regions, the cross-power spectral density matrix
S(ƒ) was first constructed by the multitaper method and then subsequently factorized. The factorization
S(ƒ)=
H(ƒ)
Ʃ H*(ƒ), gave a transfer function
H(ƒ) and a noise covariances matrix
Ʃ needed for GC estimations. The GC estimations
IY→X (ƒ) and IX→Y (ƒ) were computed for high frequency values ( >80 Hz) by the standard formula using
H(ƒ) and
Ʃ obtained from
S(ƒ) by the JLA algorithm. These estimates were used to confirm the visually suspected seizure onset region and its propagation. In all cases, written consent was obtained from patients for the use of their data in the study.
Results:
JLA-based spectral factorization in Granger causality applied to scalp EEGs successfully identified seizure onset zones (SOZs) and their propagation patterns, aligning with positive outcomes (Engel I) in five epilepsy surgery cases.Conclusions:
The JLA-based spectral factorization in Granger causality has the potential not only for accurately localizing SOZs to aid in diagnosis and treatment but also for broader applications in uncovering information flow patterns in neuroimaging and computational neuroscience.
Funding:
The study is ongoing in a frame of the Shota Rustaveli National Science Foundation of Georgia (#FR-23-18000).