time   calls  line
                  1 function y = nanstd(varargin)
                  2 %NANSTD Standard deviation, ignoring NaNs.
                  3 %   Y = NANSTD(X) returns the sample standard deviation of the values in X,
                  4 %   treating NaNs as missing values.  For a vector input, Y is the standard
                  5 %   deviation of the non-NaN elements of X.  For a matrix input, Y is a row
                  6 %   vector containing the standard deviation of the non-NaN elements in
                  7 %   each column of X. For N-D arrays, NANSTD operates along the first
                  8 %   non-singleton dimension of X.
                  9 %
                 10 %   NANSTD normalizes Y by (N-1), where N is the sample size.  This is the
                 11 %   square root of an unbiased estimator of the variance of the population
                 12 %   from which X is drawn, as long as X consists of independent, identically
                 13 %   distributed samples and data are missing at random.
                 14 %
                 15 %   Y = NANSTD(X,1) normalizes by N and produces the square root of the
                 16 %   second moment of the sample about its mean.  NANSTD(X,0) is the same as
                 17 %   NANSTD(X).
                 18 %
                 19 %   Y = NANSTD(X,FLAG,DIM) takes the standard deviation along dimension
                 20 %   DIM of X.
                 21 %
                 22 %   See also STD, NANVAR, NANMEAN, NANMEDIAN, NANMIN, NANMAX, NANSUM.
                 23 
                 24 %   Copyright 1993-2006 The MathWorks, Inc.
                 25 %   $Revision: 1.1.8.1 $  $Date: 2010/03/16 00:15:53 $
                 26 
                 27 % Call nanvar(x,flag,dim) with as many inputs as needed
  0.01       1   28 y = sqrt(nanvar(varargin{:}));