Usefulness of synthetic aperture magnetometry for presurgical evaluation in the epilepsy surgery

1 Introduction Selecting the surgical candidate for medically intractable epilepsy, detection and localization of seizure focus and cerebral functions are essential for diagnosis, for localization of surgical treatment and for prevention of surgical complications. Due to the limitations of noninvasive evaluations, invasive examinations, such as chronic ECoG grid recording, cortical stimulation mapping or amytal test , are often necessary. However, they have disadvantages of invasiveness and sampling error. Recently developed synthetic aperture magnetome-try (SAM) is a novel spatial filtering technique based on the nonlinear constrained minimum-variance beamformer [1][2]. This technique overcomes the nonuniqueness of generalized inverse solutions, such as the minimum norm, and thereby permits ambiguous three-dimensional source mapping. In the present study we have applied two techniques derived from SAM: SAM virtual sensor (VS) method, which enables to measure currento-densitogram of the small portion of the brain with an enhanced sensitivity as if intracranial electrodes were inserted [3] to estimate the epileptogenic area, and SAM statistical method which detects activated cerebral area during task performance [4]. 2 Methods 2.1 Simultaneous recording of EEG and MEG Nine patients with refractory epilepsy, 5 being temporal and 4 extratemporal, were examined (Table 1). A helmet-shaped 64-channel SQUID system (Model 100, CTF Systems) was used for MEG data acquisition. EEG signals based on international 10-20 electrode configuration were processed on the MEG system to compare both signals on the same time axis. Signals of both interictal spikes and sub-clinical electrical seizures were collected. A total of 40 raw data sets, each of them comprising of 10 seconds recording, were culled from 2 hours recording so as to include the epileptic activities on EEG. 2.2 MEG dipole analysis A single equivalent current dipole (ECD) modeling was used to estimate the source of interictal epileptic spike on MEG. ECDs were plotted on MR images scanned with the same fiducial markers that were coordinated with the marker points during MEG recording. 2.3 SAM-VS SAM virtual sensors were set in lattice of 1.0 to 1.5 cm width (8 patients) in and around the cerebral tissue where abnormal activity was detected by EEG or MEG. The signal from each virtual sensor was inspected every millisecond to determine the origin of epileptic discharge and its sequence of spread. The origin of epileptic discharge was classified according to the anatomical aspects and compared with the results of other preoperative examinations including MRI, MRS, PET, SPECT and MEG-ECD, as well as …