Abstracts

Rationale: Magnetoencephalogrephy (MEG) is useful tool for evaluation of language lateralization in patients with epilepsy. Since language activities are very small on MEG, the data is often distorted by artifacts caused by muscle and eye movements. Recently, temporally-extended signal space separation (tSSS) has been introduced for removing these artifacts. We assessed the usefulness of preprocessing of language MEG data, comparing with the result of WADA test in patients of epilepsy. Methods: Ten patients with epilepsy had MEG for language lateralization in presurgical evaluation. They also had a WADA test, which showed left language predominance in all patients. MEG data was recorded with a 306-channel whole-head system. The sampling rate was 600Hz. In all patients, we obtained anatomical MRI (magnetization-prepared rapid acquisition gradient-echo: MPRAGE) with a high-resolution 3T scanner. Patients performed a semantic language task for the paradigm of language testing. Serious of words were visually presented on the screen one at a time, and the task was to decide the word is representing abstract or concrete entity. For preprocessing, we processed MEG data with tSSS (Taulu et al., 2004) .In this method, the recored magnetic signals are decomposed into separate components representing both the neuromagnetic signals arising from inside a volume enclosed by the sensor array and any external interference signals arising from outside of the array. For determining language lateralization, we averaged the MEG data by using the timing of word presentation as a trigger. Dynamic statistical parametric maps of language activity were obtained based on a distributed source model (Dale et al., 2000). These maps were projected on the cortical surface derived from each patient s MRI. The activation between 250ms and 550ms after the trigger was analyzed. We calculated laterality index (LI) for each patient by using activation in superior temporal, middle temporal, supramarginal and inferior parietal cortices on both hemispheres. The LI was obtained by LI = (L-R)/(L R), where L and R is the number of unit dipoles with an F value higher than the threshold value in these cortical areas of left and right hemisphere, respectively. Language predominance was determined based on the LI as follows; ?0.1:left, 0.1>LI>-0.1; bilateral, ?-0.1; Right. We calculated LI from the original MEG data, and tSSS-processed data separately, and compared these result with WADA test. Results: LI derived from the original data showed the language predominance on the left, bilateral and the right in six, two and two patients, respectively. The LI ranged from -0.51 to 0.63. In contrast, LI of all patients showed left language lateralization. LI of the patients ranged from 0.10 to 0.37. Conclusions: In our patients, LI derived from tSSS-processed data was more consistent with the language lateralization determined by WADA test than LI obtained from the non-processed data. Preprocessing of MEG data with tSSS may be highly useful for obtaining reliable language lateralization by using MEG.