limo_getclustersum.m

function maxclustersum = limo_getclustersum(f,p,channeighbstructmat,minnbchan,alphav)

% limo_getclustersum finds clusters of significant F values, computes the
% sum of each cluster, and return the maximum sum across all clusters.
%
% FORMAT: maxclustersum = limo_getclustersum(f,p,channeighbstructmat,minnbchan,alpha)
% 
% INPUTS: f = 2D/3D matrix of F values - so if using T values, enter T.^2
%         p = 2D/3D matrix of P values
%         channeighbstructmat = output of limo_ft_neighbourselection
%         minnbchan = minimum number of channels, default = 2, see
%         alpha level, default 0.05
%
% OUTPUTS: maxclustersum = max cluster sum across all clusters of significant F values
%
% See also LIMO_CLUSTER_TEST LIMO_FT_FINDCLUSTER
% -----------------------------
%  Copyright (C) LIMO Team 2014

% v1: GAR, University of Glasgow, June 2010

if nargin<5;alphav=.05;end
if nargin<4;minnbchan=2;end

[posclusterslabelmat,nposclusters] = limo_ft_findcluster(p<=alphav,channeighbstructmat,minnbchan);
% figure;imagesc(posclusterslabelmat)

if nposclusters~=0
    tmp=zeros(1,nposclusters);
    for C = 1:nposclusters % compute sum for each cluster
        tmp(C) = sum( f(posclusterslabelmat==C) );
    end
    maxclustersum = max(tmp(:)); % save max across clusters
else
    maxclustersum = 0;
end