Page 138 - MEGIN Book Of Abstracts - 2023

Economics, Moscow, Russian Federation Virtual MEG sensors based on beamformer and
independent component analysis can reconstruct
ABSTRACT The reliable identification of the irrita- epileptic activity as measured on simultaneous
tive zone (IZ) is a prerequisite for the correct clinical intracerebral recordings (2022)
evaluation of medically refractory patients affected
by epilepsy. Given the complexity of MEG data, visual Velmurugan, Jayabal; Badier, Jean-Michel; Pizzo,
analysis of epileptiform neurophysiological activity Francesca; Medina Villalon, Samuel; Papageorgakis,
is highly time consuming and might leave clinically Christos; López-Madrona, Victor; Jegou, Aude; Carron,
relevant information undetected. We recorded and ana- Romain; Bartolomei, Fabrice; Bénar, Christian-G
lyzed the interictal activity from seven patients affected
by epilepsy (Vectorview Neuromag), who successfully Aix Marseille Univ, INSERM, INS, Inst Neurosci Syst, Marseille,
underwent epilepsy surgery (Engel > = II). We visu- F-13005, France; Aix Marseille Univ, INSERM, INS, Inst Neuro-
ally marked and localized characteristic epileptiform sci Syst, Marseille, F-13005, France; APHM, Timone Hospital,
activity (VIS). We implemented a two-stage pipeline for Epileptology and Cerebral Rhythmology, Marseille, F-13005,
the detection of interictal spikes and the delineation of France; Aix Marseille Univ, INSERM, INS, Inst Neurosci Syst,
the IZ. First, we detected candidate events from peaky Marseille, F-13005, France; APHM, Timone Hospital, Function-
ICA components, and then clustered events around al and Stereotactic Neurosurgery, Marseille, F-13005, France;
spatio-temporal patterns identified by convolutional Aix Marseille Univ, INSERM, INS, Inst Neurosci Syst, Marseille,
sparse coding. We used the average of clustered events F-13005, France. Electronic address: christian.benar@univ-
to create IZ maps computed at the amplitude peak amu.fr
(PEAK), and at the 50% of the peak ascending slope
(SLOPE). We validated our approach by computing the ABSTRACT The prevailing gold standard for presurgi-
distance of the estimated IZ (VIS, SLOPE and PEAK) cal determination of epileptogenic brain networks
from the border of the surgically resected area (RA). We is intracerebral EEG, a potent yet invasive approach.
identified 25 spatiotemporal patterns mimicking the Magnetoencephalography (MEG) is a state-of-the art
underlying interictal activity (3.6 clusters/patient). Each non-invasive method for investigating epileptiform
cluster was populated on average by 22.1 [15.0-31.0] discharges. However, it is not clear at what level the
spikes. The predicted IZ maps had an average distance precision offered by MEG can reach that of SEEG. Here,
from the resection margin of 8.4 ± 9.3 mm for visual we present a strategy for non-invasively retrieving the
analysis, 12.0 ± 16.5 mm for SLOPE and 22.7 ±. 16.4 mm constituents of the interictal network, with high spatial
for PEAK. The consideration of the source spread at the and temporal precision. Our method is based on MEG
ascending slope provided an IZ closer to RA and resem- and a combination of spatial filtering and independent
bled the analysis of an expert observer. We validated component analysis (ICA). We validated this approach
here the performance of a data-driven approach for the in twelve patients with drug-resistant focal epilepsy,
automated detection of interictal spikes and delinea- thanks to the unprecedented ground truth provided by
tion of the IZ. This computational framework provides simultaneous recordings of MEG and SEEG. A minimum
the basis for reproducible and bias-free analysis of MEG variance adaptive beamformer estimated the source
recordings in epilepsy. time series and ICA was used to further decompose
these time series into network constituents (MEG-ICs),
PloS one (2022), Vol. 17, No. 10 (36282803) (0 citations) each having a time series (virtual electrode) and a
topography (spatial distribution of amplitudes in the
brain). We show that MEG has a considerable sensitivity
of 0.80 and 0.84 and a specificity of 0.93 and 0.91 for
reconstructing deep and superficial sources, respec-
tively, when compared to the ground truth (SEEG). For
each epileptic MEG-IC (n = 131), we found at least one

 ontents  Index 117
C