% LOC_SUBSETS - Separate channels into maximally evenly-spaced subsets.
% This is achieved by exchanging channels between subsets so as to
% increase the sum of average of distances within each channel subset.
% Usage:
% >> subset = loc_subsets(chanlocs, nchans); % select an evenly spaced nchans
% >> [subsets subidx pos] = loc_subsets(chanlocs, nchans, plotobj, plotchans, keepchans);
%
% Inputs:
%
% chanlocs - EEGLAB dataset channel locations structure (e.g., EEG.chanlocs)
%
% nchans - (1,N) matrix containing number of channels that should be in each
% of N channel subsets respectively. if total number of channels in
% all subsets is less than the number of EEG channels in the chanlocs
% structure, N+1 subsets are created. The last subset includes the
% remaining channels.
%
% Optional inputs:
%
% plotobj - (true|false|1|0) plot the time course of the objective function
% {default: false|0)
% plotchans - (true|false|1|0) make 3-D plots showing the channel locations of
% the subsets {default: false|0)
% keepchans - (cell array) channels that has to be kept in each set and
% not undergo optimization. You can use this option to make
% sure certain channels will be assigned to each set. For
% example, to keep channels 1:10 to the first subset and
% channels 20:30 to the second, use keepchans = {[1:10], [20:30]}.
% To only keeps channels 1:5 in the first set, use
% keepchans = {1:5}. {default: {}}
%
% Outputs:
%
% subsets - {1,N} or {1,N+1} cell array containing channel indices for each
% requested channel subset.
% subidx - (1, EEG.nbchans) vector giving the index of the subset associated
% with each channel.
% pos - (3,N) matrix, columns containing the cartesian (x,y,z) channel
% positions plotted.
% Example:
%
% % Create three sub-montages of a 256-channel montage containing 32, 60, and 100
% % channels respectively.The 64 remaining channels will be put into a fourth subset.
% % Also visualize time course of optimization and the channel subset head locations.
%
% >> subset = loc_subsets(EEG.chanlocs, [32 60 100], true, true);
%
% Author: Nima Bigdely Shamlo, SCCN/INC/UCSD, 2007
% Copyright (C) 2007 Nima Bigdely Shamlo, SCCN/INC/UCSD, nima@sccn.ucsd.edu
%
% 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.
% if plotOptimization|plotSubsets in line 130 removed by nima 3/6/2007
% line 133, removed num2str removed by nima 3/6/2007
function [subset idx pos] = loc_subsets(chanlocs, numberOfChannelsInSubset, plotOptimization, plotSubsets, mandatoryChannelsForSet);
if sum(numberOfChannelsInSubset)> length(chanlocs)
error('Total channels in requested subsets larger than number of EEG channels.');
end
if min(numberOfChannelsInSubset) < 2
error('Number of channels in the requested subsets must be >= 2.');
end
rand('state',0);
if nargin < 5
mandatoryChannelsForSet = {};
end
if nargin < 4
plotSubsets = false;
end
if nargin < 3
plotOptimization = false;
end
pos=[cell2mat({chanlocs.X}); cell2mat({chanlocs.Y}); cell2mat({chanlocs.Z});];
dist = squareform(pdist(pos'));
nChans = length(chanlocs);
idx = ones(nChans,1);
setId = cell(1, length(numberOfChannelsInSubset)); % cell array containing channels in each set
remainingChannels = 1:nChans; % channels that are to be assigned to subsets
% channels that have to stay in their original subset (as in
% mandatoryChannelsForSet) and should not be re-assigned
allMandatoryChannels = cell2mat(mandatoryChannelsForSet);
% assign requested mandatory channels to subset
for i=1:length(mandatoryChannelsForSet)
setId{i} = mandatoryChannelsForSet{i};
remainingChannels(mandatoryChannelsForSet{i}) = NaN; % flag with Nan so they can be deleted later, this is to keep indexing simple
end
remainingChannels(isnan(remainingChannels)) = [];
r = remainingChannels(randperm(length(remainingChannels)));
% randomly assign remaining channels to subsets
for i=1:length(numberOfChannelsInSubset)
numberOfChannelsTobeAddedToSubset = numberOfChannelsInSubset(i) - length(setId{i})
setId{i} = [setId{i} r(1:numberOfChannelsTobeAddedToSubset)];
r(1:numberOfChannelsTobeAddedToSubset) = [];
end
if length(r) > 0
setId{length(numberOfChannelsInSubset) + 1} = r; % last set gets remaining channels
end
if plotOptimization|plotSubsets
fprintf(['Creating total of ' num2str(length(setId)) ' channel subsets:\n']);
end
if plotOptimization
figure;
xp = floor(sqrt(length(setId)));
yp = ceil(length(setId)/xp);
end
counter = 1;
exchangeHappened = true;
while exchangeHappened
exchangeHappened = false;
for set1 = 1:length(setId)
for set2 = (set1 + 1):length(setId)
for i = 1:length(setId{set1})
for j = 1:length(setId{set2})
chan(1) = setId{set1}(i);
chan(2) = setId{set2}(j);
if cost_of_exchanging_channels(chan,[set1 set2], setId, dist) < 0 && ~any(ismember(chan, allMandatoryChannels))
setId{set1}(find(setId{set1} == chan(1))) = chan(2);
setId{set2}(find(setId{set2} == chan(2))) = chan(1);
sumDistances(counter) = 0;
for s = 1:length(setId)
sumDistances(counter) = sumDistances(counter) ...
+ (sum(sum(dist(setId{s},setId{s}))) / length(setId{s}));
end
if plotOptimization
plot(1:counter,sumDistances,'-b');
xlabel('number of exchanges');
ylabel('sum mean distances within each channel subset');
drawnow;
else
if mod(counter, 20) ==0
fprintf('number of exchanges = %d\nsum of mean distances = %g\n',...
counter, sumDistances(counter));
end
end
counter = counter + 1;
exchangeHappened = true;
end
end
end
end
end
end
for set = 1:length(setId)
idx(setId{set}) = set;
% legendTitle{set} = ['subset ' num2str(set)];
end
subset = setId;
if plotSubsets
FIG_OFFSET = 40;
fpos = get(gcf,'position');
figure('position',[fpos(1)+FIG_OFFSET,fpos(2)-FIG_OFFSET,fpos(3),fpos(4)]);
scatter3(pos(1,:), pos(2,:), pos(3,:),100,idx,'fill');
axis equal;
%legend(legendTitle); it does not work propr
th=title('Channel Subsets');
%set(th,'fontsize',14)
end
if length(r)>0 && (plotOptimization|plotSubsets)
fprintf('The last subset returned contains the %d unused channels.\n',...
length(r));
end
function cost = cost_of_exchanging_channels(chan, betweenSets, setId, dist);
mirrorChan(1) = 2;
mirrorChan(2) = 1;
for i=1:2
newSetId{betweenSets(i)} = setId{betweenSets(i)};
newSetId{betweenSets(i)}(find(newSetId{betweenSets(i)} == chan(i))) = chan(mirrorChan(i));
end
cost = 0;
for i=betweenSets
distSumBefore{i} = sum(sum(dist(setId{i},setId{i})));
distSumAfter{i} = sum(sum(dist(newSetId{i},newSetId{i})));
cost = cost + (distSumBefore{i} - distSumAfter{i}) / length(setId{i});
end
%cost = (distSumAfter{1} > distSumBefore{1}) && (distSumAfter{2} > distSumBefore{2});
Datasets
Models
