DICS example

✉️ Mailing List Archive (read-only)
1

Hi All,
I am trying to implement DICS. I read the paper mentioned here
http://martinos.org/mne/stable/auto_examples/inverse/plot_dics_source_power.html#example-inverse-plot-dics-source-power-py

The paper illustrates four strategies for selecting a reference region. I fail to see the selection of reference region in the example plot_dics_source_power.py<http://martinos.org/mne/stable/_downloads/plot_dics_source_power.py>

Can someone explain how to implement dics in mne-python?

Thank you,

Chiran Doshi, MS
MEG system technologist
Department of Neurology
Boston Childrens Hospital
Phone: 781-216-1136
Fax: 781-216-1172

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20131218/daf7adce/attachment.html

2

Hi Chiran,

We only implemented estimation of source power (i.e. equation [5] in
Gross et al.) using the DICS approach, not estimation of source
connectivity (e.g. coherence, equation [1]), which I'm guessing you're
after?

You are right that the scripts don't perform selection of a reference
region, because that is only necessary to calculate the connectivity
measures. Selecting a reference could be done easily just by choosing
the source with maximal power. However, calculating coherence to the
reference region at every other region would require a bit more digging
into the code of dics_source_power (i.e. line 392 corresponds to
equation [5] in Gross et al. with the difference that the spatial
filters are calculated for the same source, i.e. r_1 == r_2).

I'd be happy to support you in implementing the coherence calculation if
you're interested.

Best,
Roman