Abstracts

Rationale: In recent years, advances in connectivity measures for magnetoencephalography (MEG) have provided analytical methods for localizing epileptiform related networks in the absence of appreciable discharges.  However, the utility of such techniques are still being realized for the presurgical evaluation of patients with refractory epilepsy. With the advancement of minimally invasive surgical procedures, such as laser thermal ablation to approach difficult surgical cases with decreased risk and morbidity, the use of advanced non-invasive mapping techniques may feature more prominently in the surgical decision process.  Here we report on a pre-surgical MEG mapping case of a patient with reflexive-motor epilepsy involving the right lower extremity who presented with surface-negative seizures after undergoing prior resection of a region in left mesial cortex.   Methods: The patient was reevaluated in the EMU and underwent Magnetic Source Imaging using an Elekta Triux 306 channel magnetoencephalography machine. Simultaneous EEG was recorded using a standard 10-20 montage. Spontaneous recordings and somatomotor mapping protocols were performed. Additionally, after functional mapping, a second spontaneous recording was attempted while the patient’s right lower extremity was manipulated in such a way as to induce a reflexive motor seizure.  Source analysis on the spontaneous MEG data (4 -70Hz)  was performed using a linearly constrained minimum variance (LCMV) beamformer followed by a spatiotemporal decomposition method to create correlation maps related to seizure activity and motor function.  Results: During EMU evaluation there was one clinical event without a clear EEG correlate. Inter-ictally, surface EEG showed parasagittal slowing and runs of activity around Cz, both greater on the left but also without clinical correlation. During routine spontaneous MEG recording there were no clinical events and no interictal epileptiform discharges observed. Upon stretching the right leg, a motor seizure was induced marked by tonic contraction of the right lower extremity and eventual spasms of the facial muscles. Similar to the clinical event in the EMU, the seizure activity produced only muscle artifact on both MEG and EEG which was exacerbated by the facial muscle involvement.  The first component map of the MEG beamformer decomposition of data during the seizure showed primarily noise localization of the muscle artifact. The second strongest correlation map also localized noise, but additionally a strong peak in the left mesial somotomotor region.  Functional localization of the right lower extremity resulted also in an area of left mesial somotomotor cortex. The patient subsequently underwent depth-electrode recording in that same region -confirming significant reflex-seizure related activity- , followed by motor mapping and thermal laser ablation surgery. The patient was seizure free upon discharge. Long-term post-operative evaluation is pending. Conclusions: This case highlights the potential role of MEG mapping in difficult surgical cases and specifically supports previous findings that source level connectivity methods can uncover significant organized structure in non-invasive MEG/EEG data despite low signal to noise. In our example, the beamformer method in particular was helpful in mitigating artifacts and allowing for the non-invasive localization of seizure related activity from a surface-negative recording.  Funding: none