% POP_REJCHANSPEC - reject artifacts channels in an EEG dataset using
% channel spectrum. The average spectrum for all selected
% is computed and a threshold is applied.
%
% Usage:
% >> pop_rejchanspec( INEEG ) % pop-up interactive window mode
% >> [OUTEEG, indelec] = pop_rejchanspec( INEEG, 'key', 'val');
%
% Inputs:
% INEEG - input EEGLAB dataset
%
% Optional inputs:
% 'freqlims' - [min max] frequency limits. May also be an array where
% each row defines a different set of limits. Default is
% 35 to the Niquist frequency of the data.
% 'stdthresh' - [max] positive threshold in terms of standard deviation.
% Default is 5.
% 'absthresh' - [max] positive threshold in terms of spectrum units
% (overrides the option above).
% 'averef' - ['on'|'off'] 'on' computes average reference before
% applying threshold. Default is 'off'.
% 'plothist' - ['on'|'off'] 'on' plot the histogram of values along
% with the threshold.
% 'plotchans' - ['on'|'off'] 'on' plot the channels scrollplot with
% selected channels for rejection in red. Allow selected
% channels rejection with the 'REJECT' button.
% 'elec' - [integer array] only include specific channels. Default
% is to use all channels.
% 'specdata' - [fload array] use this array containing the precomputed
% spectrum instead of computing the spectrum. Default is
% empty.
% 'specfreqs' - [fload array] frequency array for precomputed spectrum
% above.
% 'verbose' - ['on'|'off'] display information. Default is 'off'.
%
% Outputs:
% OUTEEG - output dataset with updated joint probability array
% indelec - indices of rejected electrodes
% specdata - data spectrum for the selected channels
% specfreqs - frequency array for spectrum above
%
% Author: Arnaud Delorme, CERCO, UPS/CNRS, 2008-
% Copyright (C) 2008 Arnaud Delorme, CERCO, UPS/CNRS
%
% This file is part of EEGLAB, see http://www.eeglab.org
% for the documentation and details.
%
% Redistribution and use in source and binary forms, with or without
% modification, are permitted provided that the following conditions are met:
%
% 1. Redistributions of source code must retain the above copyright notice,
% this list of conditions and the following disclaimer.
%
% 2. Redistributions in binary form must reproduce the above copyright notice,
% this list of conditions and the following disclaimer in the documentation
% and/or other materials provided with the distribution.
%
% THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
% AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
% IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
% ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
% LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
% CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
% SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
% INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
% CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
% ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
% THE POSSIBILITY OF SUCH DAMAGE.
function [EEG allrmchan specdata specfreqs com] = pop_rejchanspec(EEG, varargin)
if nargin < 1
help pop_rejchanspec;
return;
end
allrmchan = [];
specdata = [];
specfreqs = [];
com = '';
if nargin < 2
uilist = { { 'style' 'text' 'string' 'Electrode (number(s); Ex: 2 4 5)' } ...
{ 'style' 'edit' 'string' ['1:' int2str(EEG.nbchan)] } ...
{ 'style' 'text' 'string' 'Frequency limits [min max]' } ...
{ 'style' 'edit' 'string' [ '35 ' int2str(floor(EEG.srate/2)) ] } ...
{ 'style' 'text' 'string' 'Standard dev. threshold limits [max]' } ...
{ 'style' 'edit' 'string' '5' } ...
{ 'style' 'text' 'string' 'OR absolute threshold limit [min max]' } ...
{ 'style' 'edit' 'string' '' } ...
{ 'style' 'text' 'string' 'Compute average reference first (check=on)' } ...
{ 'style' 'checkbox' 'string' '' 'value' 0 } { } ...
{ 'style' 'text' 'string' 'Plot histogram of power values (check=on)' } ...
{ 'style' 'checkbox' 'string' '' 'value' 0 } { } ...
{ 'style' 'text' 'string' 'Plot channels scrollplot (check=on)' } ...
{ 'style' 'checkbox' 'string' '' 'value' 0 } { } ...
};
geom = { [2 1] [2 1] [2 1] [2 1] [2 0.3 0.7] [2 0.3 0.7] [2 0.3 0.7] };
result = inputgui( 'uilist', uilist, 'geometry', geom, 'title', 'Reject channel using spectrum -- pop_rejchanspec()', ...
'helpcom', 'pophelp(''pop_rejchan'')');
if isempty(result), return; end
options = { 'elec' eval( [ '[' result{1} ']' ] ) 'stdthresh' str2num(result{3}) 'freqlims' str2num(result{2}) };
if ~isempty(result{4})
options = { options{:} 'absthresh' str2num(result{4}) };
end
if result{5},
options = { options{:} 'averef', 'on' };
end
if result{6},
options = { options{:} 'plothist', 'on' };
end
% Begin: Added by Romain on 22 July 2010
if result{7},
options = { options{:} 'plotchans', 'on' };
end
% End: Added by Romain on 22 July 2010
else
options = varargin;
end
% decode options
% --------------
opt = finputcheck( options, { 'averef' 'string' { 'on';'off' } 'off';
'plothist' 'string' { 'on';'off' } 'off';
'plotchans' 'string' { 'on';'off' } 'off';
'verbose' 'string' { 'on';'off' } 'off';
'elec' 'integer' [] [1:EEG.nbchan];
'freqlims' 'real' [] [35 EEG.srate/2];
'specdata' 'real' [] [];
'specfreqs' 'real' [] [];
'absthresh' 'real' [] [];
'stdthresh' 'real' [] 5 }, 'pop_rejchanspec');
if ischar(opt), error(opt); end
% compute average reference if necessary
if strcmpi(opt.averef, 'on')
NEWEEG = pop_reref(EEG, [], 'exclude', setdiff([1:EEG.nbchan], opt.elec));
else NEWEEG = EEG;
end
if isempty(opt.specdata)
[tmpspecdata specfreqs] = pop_spectopo(NEWEEG, 1, [], 'EEG' , 'percent', 100, 'freqrange',[0 EEG.srate/2], 'plot', 'off');
% add back 0 channels
devStd = std(EEG.data(:,:), [], 2);
if any(devStd == 0)
goodchan = find(devStd ~= 0);
specdata = zeros(length(opt.elec), size(tmpspecdata,2));
specdata(goodchan,:) = tmpspecdata;
else
specdata = tmpspecdata;
end
else
specdata = opt.specdata;
specfreqs = opt.specfreqs;
end
if size(opt.stdthresh,1) == 1 && size(opt.freqlims,1) > 1
opt.stdthresh = ones(length(opt.stdthresh), size(opt.freqlims,1))*opt.stdthresh;
end
allrmchan = [];
for index = 1:size(opt.freqlims,1)
% select frequencies, compute median and std then reject channels
% ---------------------------------------------------------------
[tmp fbeg] = min(abs(specfreqs - opt.freqlims(index,1)));
[tmp fend] = min(abs(specfreqs - opt.freqlims(index,2)));
selectedspec = mean(specdata(opt.elec, fbeg:fend), 2);
if ~isempty(opt.absthresh)
rmchan = find(selectedspec <= opt.absthresh(1) | selectedspec >= opt.absthresh(2));
else
m = median(selectedspec);
s = std( selectedspec);
nbTresh = size(opt.stdthresh);
if length(opt.stdthresh) > 1
rmchan = find(selectedspec <= m+s*opt.stdthresh(index,1) | selectedspec >= m+s*opt.stdthresh(index,2));
else
rmchan = find(selectedspec > m+s*opt.stdthresh(index));
end
end
% print out results
% -----------------
if isempty(rmchan)
textout = sprintf('Range %2.1f-%2.1f Hz: no channel removed\n', opt.freqlims(index,1), opt.freqlims(index,2));
else textout = sprintf('Range %2.1f-%2.1f Hz: channels %s removed\n', opt.freqlims(index,1), opt.freqlims(index,2), int2str(opt.elec(rmchan')));
end
fprintf(textout);
if strcmpi(opt.verbose, 'on')
for inde = 1:length(opt.elec)
if ismember(inde, rmchan)
fprintf('Elec %s power: %1.2f *\n', EEG.chanlocs(opt.elec(inde)).labels, selectedspec(inde));
else fprintf('Elec %s power: %1.2f\n', EEG.chanlocs(opt.elec(inde)).labels , selectedspec(inde));
end
end
end
allrmchan = [ allrmchan rmchan' ];
% plot histogram
% --------------
if strcmpi(opt.plothist, 'on')
figure; hist(selectedspec);
hold on; yl = ylim;
if ~isempty(opt.absthresh)
plot([opt.absthresh(1) opt.absthresh(1)], yl, 'r');
plot([opt.absthresh(2) opt.absthresh(2)], yl, 'r');
else
if length(opt.stdthresh) > 1
threshold1 = m+s*opt.stdthresh(index,1);
threshold2 = m+s*opt.stdthresh(index,2);
plot([m m], yl, 'g');
plot([threshold1 threshold1], yl, 'r');
plot([threshold2 threshold2], yl, 'r');
else
threshold = m+s*opt.stdthresh(index,1);
plot([threshold threshold], yl, 'r');
end
end
title(textout);
end
end
allrmchan = unique_bc(allrmchan);
com = sprintf('EEG = pop_rejchan(EEG, %s);', vararg2str(options));
if strcmpi(opt.plotchans, 'on')
tmpcom = [ 'EEGTMP = pop_select(EEG, ''nochannel'', [' num2str(opt.elec(allrmchan)) ']);' ];
tmpcom = [ tmpcom ...
'LASTCOM = ' vararg2str(com) ';' ...
'[ALLEEG EEG CURRENTSET tmpcom] = pop_newset(ALLEEG, EEGTMP, CURRENTSET);' ...
' if ~isempty(tmpcom),' ...
' EEG = eegh(LASTCOM, EEG);' ...
' eegh(tmpcom);' ...
' eeglab(''redraw'');' ...
' end; clear EEGTMP tmpcom;' ];
colors = cell(1,length(opt.elec)); colors(:) = { 'k' };
colors(allrmchan) = { 'r' }; colors = colors(end:-1:1);
fprintf('%d electrodes labeled for rejection\n', length(find(allrmchan)));
tmpchanlocs = EEG.chanlocs;
if ~isempty(EEG.chanlocs), tmplocs = EEG.chanlocs(opt.elec); tmpelec = { tmpchanlocs(opt.elec).labels }';
else tmplocs = []; tmpelec = mattocell([1:EEG.nbchan]');
end
eegplot(EEG.data(opt.elec,:,:), 'srate', EEG.srate, 'title', 'Scroll component activities -- eegplot()', ...
'limits', [EEG.xmin EEG.xmax]*1000, 'color', colors, 'eloc_file', tmplocs, 'command', tmpcom);
else
EEG = pop_select(EEG, 'nochannel', opt.elec(allrmchan));
end
if nargin < 2
allrmchan = sprintf('EEG = pop_rejchanspec(EEG, %s);', vararg2str(options));
end
Datasets
Models
