Abstracts

Rationale: Interictal EEG spikes are highly correlated with seizure onset in surgical candidates for temporal lobe epilepsy (TLE) The additional clinical information from MEG in TLE is said to be limited due to difficulties with identifying deeper sources with MEG and the high sensitivity of other imaging modalities for detecting temporal lobe dysfunction.Dipole analysis has been used most frequently in previous series. This method is limited by need for sufficiently large signal to noise ratio (SNR) and confounded when there are multiple spike generators. Beamforming may compensate for some of these difficulties. We used synthetic aperture magnetometry to image the excess kurtosis (SAM(g2)) associated with spike generation in 7 cases of suspected TLE where standard testing for focus localization showed discordant or atypical findings.Methods: Spontaneous MEG and EEG activity was recorded with a 275 channel CTF MEG 2005 system for ~1 hour at 600 Hz with a bandpass of 1-70 Hz. All MEG data were coregistered with each subject s structural MRI. SAM(g2) was used to localize the putative interictal spike generators (20 70 Hz) and to reconstruct the source time series (i.e. virtual depth electrodes, at 3-70 Hz) at all local maxima with g2?2. Spike times were marked in the source series and compared to the activity visible in the simultaneous EEG. This study was conducted in compliance with IRB15854. Results: SAM(g2)was able to identify abnormalities in 6 of 7 patients. The findings included 1) Mesial temporal spike localization not defined on EEG, 2) Unitemporal MEG spikes ipsilateral to seizure onset in case with contralateral sclerosis and bitemporal EEG spikes, 3) Spikes at posterior edge of resection in case of recurrent seizures after surgery for unitemporal epilepsy 4)Anteroinferior temporal lobe MEG discharges in case with olfactory auras and nondiagnostic EEG, 5) MEG spikes contralateral to side of suspected seizure onset and most frequent EEG spikes 6)Multiple areas in ipsilateral,contralateral temporal and extratemporal areas suggesting propagation in case of neocortical temporal epilepsy where EEG suggested rapid bilateral synchrony but not clearly evident.Conclusions: SAM(g2) can be used to localize and visualize the kurtotic spiking activity associated with temporal lobe epilepsy. This may occur in deeper sources of mesial temporal epilepsy but may explain how some bitemporal EEG spikes appear in suspected unilateral TLE. Beamforming and SAM(g2) maps and source series reconstructions can illustrate spike propagation through brain networks. Further study is needed to correlate the utility of these findings with clinical semiology and invasive recordings.