Hi. I successfully performed AEC connectivity analysis on source-based HD-EEGs, getting similar mean connectivity values to other literature. However when I try to perform wPLI connectivity analysis on the same HD-EEGs (but focusing on a selection of scouts relevant to the DMN rather than all 68 dessikan-killarney nodes) I keep getting values of all 1 or NaN, rather than a range between 0 and 1 as expected. I'm not sure if there's something else I need to do prior to the connectivity step to process the data differently, as I assume these values aren't a true representation of the wPLI. Any advice would be very helpful, thanks.
@uqlhaine, please provide more details on the parameters that are being used in the wPLI computation.
Hi, thanks for answering (sorry for my delay - I work part-time). For each subject I have EEGLAB channels (approximately 110), OpenMEEG BEM, No noise modeling, and approximately 120 2 second epochs of preprocessed EEG. I then compute head model, and then compute sources (I chose Kernel only: one per file, then dSPM, diagonal noise covariance) to generate a dSPM file for each EEG epoch. I then click on the brain tab and run to extract values for each dSPM file: I select All file time window, and choose scouts (Desikan-Killiany) relevant to the default mode network (L&R isthmus cingulate, L&R poster cingulate, L&R anterior cingulate cortex, L&R parahippocampus, L&R rostral middle frontal), Scout function 'mean', concatenate time. I then run connectivity, choosing Phase locking value N: All file, wPLI, Hilbert transform (alpha and beta frequencies), full time resolution, output across combined epochs. This generates a matrix as I'd expect, but the values are all 1, which makes me think I'm doing something wrong. I have read the tutorial on this, and I've tried a few variations with extracting values prior to the connectivity step, but keep getting 1s. Thanks again for any input.
I think it's probably because I'm computing dSPM and this loses the phase information?