Abstracts

Rationale: Evaluation of language activity of the brain is important for planning epilepsy surgery. Magnetoencephalography (MEG), a non-invasive neuroimaging procedure, is used to map verbal function by using language tasks. However, these MEG data may be distorted by movement-related artifacts. Recently, an algorithm for movement compensation (MC) using continuous head position monitoring has been introduced and used in combination with the temporally-extended signal space separation method (tSSS). The purpose of this study is to assess the feasibility of using MC for evaluating the language activity recorded on MEG. Methods: Seven intractable epilepsy patients (male: 4, female: 3, age: 4-36) were retrospectively studied. MEG was recorded with a 306-channel whole-head system at a sampling rate of 1000Hz. In all patients, high-resolution 3T anatomical MRI data were acquired with magnetization-prepared rapid acquisition gradient-echo (MPRAGE). During MEG recording, patients performed one of four language tasks: (1) Visual word presentation/semantic decision task (n=4), (2) Visual word presentation /word recognition task (n=1), (3) Auditory word presentation/semantic decision task (n=1), (4) Auditory word presentation/word recognition task (n=1). In these tasks, we presented 160 English words visually or auditorily with an interstimulus interval of 2000ms and 3000ms, respectively. The MEG data was preprocessed by using MC coupled with tSSS. We analyzed the original and preprocessed data separately. MEG data epochs from 500ms pre-stimulus and 1000ms post-stimulus period were averaged. In each patient, we calculated spatiotemporal source distribution of the averaged data by using dynamic statistical parametric maps (dSPMs). DSPMs were mapped on the MRI-derived cortical surface reconstructed by Freesurfer. We selected the following regions of interest (ROIs) based on a priori knowledge of Wernicke and Broca areas ; superior temporal, middle temporal, supramarginal, inferior parietal cortices, opecular and triangular parts of the inferior frontal lobe on both hemispheres. The source waveforms were extracted from each vertex of these ROIs within a time window of 250-550ms post stimuli. The maximum amplitude values were obtained within each ROI. We evaluated the source maps of the original and preprocessed data visually, and the maximum values were compared between these two data sets in each patient. Results: In visual inspection of source distribution maps, five of seven patients (71%) showed language activation more clearly in the preprocessed data than the original data. Five patients demonstrated smaller maximum values observed in 8-12 ROIs in the sources obtained from the preprocessed data. In other two patients, the maximum values tended to be larger in the preprocessed data. Conclusions: Preprocessing with the movement compensation algorithm may improve the quality of spatiotemporal source maps obtained from language MEG activity. Smaller activation values obtained from the preprocessed data suggest that motion-related artifacts are reduced by this procedure.