integrateDown

 Ystd = integrateDown(Y,intLength)
 
 This function integrates the data down. You pass it two arguments:
   Y - MxN matrix of data. The function operates on each column.
   intLength - Px1 vector of bin lengths, in terms of number of samples
 For each column in Y, this function bins the data into bins whose length
 is specified in intLength. It calculates the standard deviation of the
 binned data. It does this for each bin length specified in intLength. It
 returns the matrix:
   Ystd - PxN matrix. Row k corresponds to the standard deviation of the
   data when binned by intLength(k) samples. This is given for each
   column.

 ogk, 15 June 2011

0001 function Ystd = integrateDown(Y,intLength)
0002 % Ystd = integrateDown(Y,intLength)
0003 %
0004 % This function integrates the data down. You pass it two arguments:
0005 %   Y - MxN matrix of data. The function operates on each column.
0006 %   intLength - Px1 vector of bin lengths, in terms of number of samples
0007 % For each column in Y, this function bins the data into bins whose length
0008 % is specified in intLength. It calculates the standard deviation of the
0009 % binned data. It does this for each bin length specified in intLength. It
0010 % returns the matrix:
0011 %   Ystd - PxN matrix. Row k corresponds to the standard deviation of the
0012 %   data when binned by intLength(k) samples. This is given for each
0013 %   column.
0014 %
0015 % ogk, 15 June 2011
0016 
0017 % Y is a matrix. We want to apply this calculation to each column of Y.
0018 Nsamples = size(Y,1);
0019 intLength = intLength(:); % just in case it's a row vector instead of a column vector
0020 Ystd = zeros(size(intLength,1),size(Y,2));
0021 
0022 % For each entry in intLength, we want to bin that many samples together,
0023 % then calculate the standard deviation of the binned data. This standard
0024 % deviation data is returned in Ystd.
0025 
0026 for k=1:length(intLength)
0027     Nbin = intLength(k);
0028     if Nbin <= Nsamples/2
0029         Nselect = Nbin*floor(Nsamples/Nbin); % select only these samples, throwing away the rest
0030         % Now use the power of MATLAB!
0031         Ystd(k,:) = std(reshape(mean(reshape(Y(1:Nselect,:),Nbin,Nselect/Nbin,size(Y,2)),1),Nselect/Nbin,size(Y,2),1),0,1);
0032     else
0033         % Cannot calculate a standard deviation, return NaN's
0034         disp('Error in function: integrateDown(Y,intLength).')
0035         fprintf('You have asked for bins of length %d, but there are only %d samples.\n',Nbin,Nsamples)
0036         disp('This makes it impossible to calculate a standard deviation. Sorry. Assigning NaNs.')
0037         Ystd(k,:) = nan;
0038     end
0039 end
0040 
0041 end