A pipeline for exploring time evolving M/EEG networks, based on the paper Measurement of dynamic task related functional networks using MEG (O'Neill et al; NeuroImage 2017) https://doi.org/10.1016/j.neuroimage.2016.08.061
All you need is a relatively up-to-date version of Fieldtrip http://www.fieldtriptoolbox.org/ as this handles all the data preprocessing and source reconstruction for ease.
The code currently assumes you are giving it CTF data and a coregistered MRI file for source reconstruction. For those who want to use other manufacturers data, you will need to edit go_generateAtlasFilters.
Owing to the simultaneous flexibility and rigidity of Fieldtrip, there are some subtle differences between the above publication and this repository, namely:
- Forward modelling is via. Nolte's corrected sphere approach rather than Huang's local spheres approximation.
- Optimal weights are derived from Sekihara's eigenvalue decomposition, rather than explicit search. These changes should not be to any detriment of the results.
Please cite the following paper if this code is of any benefit to yourself.
O’Neill G.C., Tewarie P.K., Colclough G.L., Gascoyne L.E., Hunt B.A.E., Morris P.G., Woolrich M.W., Brookes M.J., (2017) Measurement of dynamic task related functional networks using MEG. NeuroImage 146 pp. 667-78