Coherence analysis of language function using the spatial filter technique of the magnetoencephalography

The aim of this study was to propose a coherence using the special filter technique in magnetoencephalography. We reported previously intensive beta activity in the dominant inferior frontal language area during both silent reading and word generation task. However, it was hard to detect the posterior language area in our MEG studies. Therefore, we used the coherence technique to detect the neuronal network for language function. The above-mentioned two tasks were performed with right-handed volunteers using 160 channels of the whole head MEG. We presented the task Japanese character every six minutes and repeated this trial about 80 times. The volunteers repeatedly recollected one word related to the character. After presenting a trigger character, we analyzed 400 ms of data (from 200 ms to 600 ms after presentation of the trigger character) using a special filter technique. We set the seed point on the left inferior frontal gyrus in the MRI and explored the coherence for the alpha (8-13 Hz), beta (13-25 Hz), and gamma (25-50 Hz) bands. The results showed various patterns of the coherence that were not stable and not common in all cases. Alpha, beta, and gamma activities were detected in the ipsilateral front-temporal region, and the beta intensity tended to be higher than the alpha intensity. The neural network was found to be very complicated, and there were many discrepancies in each condition. However, this coherent analysis is useful for confirming the speech dominant side and learning about human cognitive function.