Abstracts

Rationale: Previous studies have shown that magnetoencephalography (MEG) using conventional magnetic shielded room (MSR) is able to detect interictal epileptic activity in 30 up to 80 % of patients with mesial temporal epilepsy. The aim of this study is to determine if MEG performed in a light weight magnetic shield (LW-MSR) based on a new concept provides sufficient signal to noise ratio (SNR) to reliably detect and localize the magnetic correlate of mesial temporal epileptic activity. Methods: Twenty-six patients (17 women; age range: 7-54 years) with a presumed location of the epileptogenic zone in the mesial temporal region based on non-invasive presurgical evaluation were prospectively studied. Four patients had normal structural MRI and 22 had symptomatic epilepsy related to hippocampal sclerosis (14 patients), tumor (6 patients), post-traumatic encephalomalacia (1 patient), or focal cortical dysplasia (1 patient). MEG investigations were performed with the 306-channel Elekta Neuromag system into a MaxShield LW-MSR (Elekta Neuromag Oy). The LW-MSR combined three interference suppression methods: 1) a passive light-weight MSR comprising a single shell of interleaved mu-metal/aluminum layers, 2) six internal active feedback compensation coils that use the regular sensor array of the MEG system as a zero indicator to deliver a cancellation field for attenuating the interference, and 3) the signal space separation method to reduce the residual magnetic noise. In all patients, spontaneous magnetic activity (eyes-closed rest, lying position) was continuously recorded during one hour and then visually inspected for epileptic events. Equivalent current dipoles (ECD, g/% >80%) corresponding to epileptic events were fitted in the patients’ spherical head model at the onset and the peak of the epileptic spikes and then superimposed on their co-registered MRI. Results: Interictal epileptic activity was found in 18 out 26 patients (69 %) in the affected temporal lobe. Anterior temporal, mesial temporal and posterior temporal ECDs were found in 16 patients (89%), 7 patients (39%) and 3 patients (17%), respectively. Interestingly, in 3 patients, ECDs fitted at spike-onset were localized in the hippocampus while at the peak of the spike, they had an anterior temporal location. In 3 out of 8 patients with normal MEG, EEG-video monitoring did not show any interictal epileptic abnormalities. Conclusions: This study shows that MEG using LW-MSR provides adequate attenuation of magnetic interference and good SNR to allow reliable detection and localization of single epileptic abnormalities on continuous MEG data in patients with presumed mesial temporal epilepsy. Moreover, our results demonstrate that MEG using LW-MSR provides sufficient SNR to detect mesial temporal epileptic sources in 39% of the patients. They also suggest that anterior temporal ECDs may have a source in the mesial temporal structures that is probably not picked up by the MEG sensors in some cases due to the weakness against background brain activity, the orientation or the depth of these sources.