Two 'removeevoked' options in 'process_cohere1n'

Discussions
,
1

Dear brainstorm community,

I'm working on scouts*scouts coherence in my MEG data. I followed the tutorial (https://neuroimage.usc.edu/brainstorm/Tutorials/CorticomuscularCoherence#Coherence_Scouts_x_Scouts) and then got codes from the pipeline editor in matlab (Please see codes below).

 % Process: Coherence NxN [2023]
        sFileCohNN = bst_process('CallProcess', 'process_cohere1n', sFiles, [], ...
            'timewindow',    [0, 0.4], ...
            'scouts',        {'Volume 13263: Desikan-Killiany', {'Fusiform L', 'Fusiform R', 'Rostral mid front L'}}, ...
            'flatten',       0, ...
            'scouttime',     'before', ...  % before
            'scoutfunc',     'mean', ...  % Mean
            'scoutfuncaft',  'mean', ...  % Mean
            'pcaedit',       struct(...
                 'Method',         'pca', ...
                 'Baseline',       [ ], ...
                 'DataTimeWindow', [ ], ...
                 'RemoveDcOffset', 'file'), ...
            'removeevoked',  0, ...
            'cohmeasure',    'mscohere', ...  % Magnitude-squared coherence
            'tfmeasure',     'morlet', ...  % Morlet wavelets
            'tfedit',        struct(...
                  'Comment',         'ComplexMorlet,1-50Hz', ...
                  'TimeBands',       [], ...
                  'Freqs',           [1:50], ...
                  'MorletFc',        1, ...
                  'MorletFwhmTc',    3, ...
                  'ClusterFuncTime', 'none', ...
                  'Measure',         'none', ...
                  'Output',          'all', ...
                  'RemoveEvoked',    0, ...
                  'SaveKernel',      0), ...
            'timeres',       'none', ...  % None
            'avgwinlength',  1, ...
            'avgwinoverlap', 50, ...
            'outputmode',    'avg');  % across combined files/epochs

However, I noticed that there are two 'removedevoked' in the code. What is the difference between these two options? I'd like to estimate coherence between the fusiform and left middle frontal cortex using the induced power of these two brain regions. Should I change the first 'removeevoked' to 1, the second 'RemoveEvoked', or both?

Best,
Shuting

2

I've tried to analyze the data from one participant using different 'removeevoked' options this morning and found a slight difference (I've added some other brain regions for exploration). Hope to give a reference to someone who has the same question as me...

First 'removeevoked' = 1 & Second 'RemoveEvoked' = 0

first1second0
first1second01919×514 134 KB

First 'removeevoked' = 0 & Second 'RemoveEvoked' = 1
first0second1
first0second11917×517 137 KB

First 'removeevoked' = 1 & Second 'RemoveEvoked' = 1
first1second1
first1second11918×517 136 KB

Any suggestions and comments would be helpful. Thanks in advance!

3

Indeed, once the process is exported there are two fields RemoveEvoked, one belongs to the connectivity (coherence) process and one for the TF decomposition.

  1. RemoveEvoked from coherence, it does remove the average across trials for signals that will be computed.
    In your example, coherence is computed between Scouts, with the Scout function being applied before, thus there will be one time series per Scout per File. Setting RemoveEvoked to 1, will remove the average Scout time series across files from each Scout before computing coherence.

  2. RemoveEvoked from TF options, this option not used and does not have an impact in the computation of connectivity (this can be seen in the shared examples). In fact, it is not shown in the GUI of the TF options (Edit button).

4

Dear Raymundo,

I see. Thanks for your reply!

Best,
Shuting