Mapping language from MEG beta power modulations during auditory and visual naming. Neuroimage 2020 Oct 15;220:117090
Date
07/01/2020Pubmed ID
32593799DOI
10.1016/j.neuroimage.2020.117090Scopus ID
2-s2.0-85086940876 (requires institutional sign-in at Scopus site) 15 CitationsAbstract
Evaluation of language dominance is an essential step prior to epilepsy surgery. There is no consensus on an optimal methodology for determining language dominance using magnetoencephalography (MEG). Oscillatory dynamics are increasingly recognized as being of fundamental importance for brain function and dysfunction. Using task-related beta power modulations in MEG, we developed an analysis framework for localizing and lateralizing areas relevant to language processing in patients with focal epilepsy. We examined MEG responses from 29 patients (age 42 ± 13 years, 15M/14F) during auditory description naming (ADN) and visual picture naming (PN). MEG data were preprocessed using a combination of spatiotemporal filtering, signal thresholding, and ICA decomposition. Beta-band 17-25Hz power decrements were examined at both sensor and source levels. Volumetric grids of anatomical source space were constructed in MNI space at 8 mm isotropic resolution, and beta-band power changes were estimated using the dynamic imaging of coherent sources beamformer technique. A 600 ms temporal-window that ends 100 ms before speech onset was selected for analysis, to focus on later stages of word production such as phonologic selection and motor speech preparation. Cluster-based permutation testing was employed for patient- and group-level statistical inferences. Automated anatomic labeling atlas-driven laterality indices (LIs) were computed for 13 left and right language- and motor speech-related cortical regions. Group localization of ADN and PN consistently revealed significant task-related decrements of beta-power within language-related areas in the frontal, temporal and parietal lobes as well as motor-related regions of precentral/premotor and postcentral/somatomotor gyri. A region-of-interest analysis of ADN and PN suggested a strong correlation of r = 0.74 (p < 0.05, FDR corrected) between the two tasks within the language-related brain regions, with the highest spatial overlap in the prefrontal areas. Laterality indices (LIs) consistently showed left dominance (LI > 0.1) for most individuals (93% and 82% during ADN and PN, respectively), with average LIs of 0.40 ± 0.25 and 0.34 ± 0.20 for ADN and PN, respectively. Source analysis of task-related beta power decrements appears to be a reliable method for lateralizing and localizing brain activations associated with language processing in patients with epilepsy.
Author List
Youssofzadeh V, Stout J, Ustine C, Gross WL, Conant LL, Humphries CJ, Binder JR, Raghavan MAuthors
Jeffrey R. Binder MD Professor in the Neurology department at Medical College of WisconsinWilliam Gross MD, PhD Associate Professor in the Anesthesiology department at Medical College of Wisconsin
Manoj Raghavan MD, PhD Professor in the Neurology department at Medical College of Wisconsin
Vahab Youssofzadeh PhD Assistant Professor in the Neurology department at Medical College of Wisconsin
MESH terms used to index this publication - Major topics in bold
AdultBrain
Brain Mapping
Brain Waves
Epilepsies, Partial
Female
Functional Laterality
Humans
Language
Magnetoencephalography
Male
Middle Aged
Speech
