Abstracts

Rationale:
Recently we developed a beta-power desynchrony approach for mapping language areas using task magnetoencephalography (MEG) (Youssofzadeh et al., 2020). Here we extend our proposed approach by examining MEG responses in a group of (n=65) patients enrolled in the Epilepsy Connectome Project who completed a protocol contrasting story comprehension with math problem-solving. 
Method:
The story-math contrast was originally designed for fMRI, to activate areas in the anterior temporal lobe (ATL) that may have a role in conceptual integration and at risk of damage during temporal lobe surgery for medically refractory epilepsy (Binder et al., 2011). The protocol consisted of two runs, each interleaving 7 blocks of the story task and 15 blocks of the math task, with an average block duration of ~30 sec. Data were preprocessed using a combination of spatiotemporal filtering, signal thresholding, and ICA decomposition. Data were analyzed over 300ms time windows, 400 to 1900ms after the onset of story sentences or presentation of a math problem, against 300ms before the onset of the blocks. Activation mapping was conducted based on beta-band power source suppression in a frequency range of 17-25Hz using sources estimated using a DICS beamformer analysis on cortical brain areas. Group source activities were statistically examined using a non-parametric permutation test and a paired t-test analysis of story and math responses. A Harvard-Oxford atlas consisting of 48 cortical regions was used to summarize group source localization.
Results:
Group statistical analysis of story and math tasks revealed significantly greater power desynchronization during the story task in multiple language areas, including left anterior supramarginal gyrus (t = 9.4), left inferior temporal gyrus (t = 6.1), left ATL (t = 5.6), a broad region of the left lateral prefrontal cortex (t = 6.3), and angular gyrus (t = 3.3), as shown in Fig. 1C. In addition, greater desynchronization during the story task occurred in right hemisphere homologs of these regions, including the right prefrontal cortex (t = 9), supramarginal gyrus, and inferior temporal gyrus (t = 7.3), and in pre- and postcentral gyri bilaterally (t = 4.4, t = 5.1).
Conclusion:
Compared to previously reported fMRI results using this contrast, MEG detected much more extensive story task-induced response in the prefrontal cortex bilaterally, and less extensive angular gyrus response. The results further support the use of beta desynchrony as an effective measure for localizing task-induced processing with MEG.
Funding:
:Startup fund by the medical college of Wisconsin