spm_mip.m

function mip = spm_mip(Z,XYZ,M,units)
% SPM Maximum Intensity Projection
% FORMAT mip = spm_mip(Z,XYZ,M,units)
% Z       - vector point list of SPM values for MIP
% XYZ     - matrix of coordinates of points (mip coordinates)
% M       - voxels - > mip matrix or size of voxels (mm)
% units   - defining space     [default {'mm' 'mm' 'mm'}]
%
% mip     - maximum intensity projection
%           if no output, the mip is displayed in current figure.
%__________________________________________________________________________
%
% If the data are 2 dimensional [DIM(3) = 1] the projection is simply an
% image, otherwise:
%
% spm_mip creates and displays a maximum intensity projection of a point
% list of voxel values (Z) and their location (XYZ) in three orthogonal
% views of the brain. It is assumed voxel locations conform to the space
% defined in the atlas of Talairach and Tournoux (1988); unless the third
% dimension is time.
%
% This routine loads a mip outline from MIP.mat. This is an image with
% contours and grids defining the space of Talairach & Tournoux (1988).
% mip95 corresponds to the Talairach atlas, mip96 to the MNI templates.
% The outline and grid are superimposed at intensity 0.4.
%
% A customised mip outline can be used instead of the default.
%
% A default colormap of 64 levels is assumed. The pointlist image is
% scaled to fit in the interval [1/9,1]*64 for display. Flat images
% are scaled to 1*64.
%
% If M is not specified, it is assumed the XYZ locations are 
% in Talairach mm.
%__________________________________________________________________________
% Copyright (C) 1996-2013 Wellcome Trust Centre for Neuroimaging

% Karl Friston
% $Id: spm_mip.m 7654 2019-08-25 20:09:35Z karl $

%-Get units and grid scaling
%--------------------------------------------------------------------------
try, units; catch, units = {'mm' 'mm' 'mm'}; end
try, M;     catch, M = 1; end
Grid = 0.4;

%-Transpose locations if necessary
%--------------------------------------------------------------------------
if size(XYZ,1) ~= 3, XYZ = XYZ';                end
if size(Z,1)   ~= 1, Z   = Z';                  end
if size(Z,2)   == 1, Z   = ones(1,size(XYZ,2)); end
if size(M,1)   == 1, M   = speye(4,4)*M;        end

%-Scale & offset point list values to fit in [1/(1+Scal),1]
%--------------------------------------------------------------------------
Z    = Z - min(Z);
mx   = max(Z);
Scal = 8;
if isempty(mx)
    Z = [];
elseif isfinite(mx) && mx
    Z = (1 + Scal*Z/mx)/(Scal + 1);
else
    Z = ones(1,length(Z));
end

%-Display format
%==========================================================================
% load various grids, DXYZ, CXYZ, scale (see spm_mip_ui and spm_project)
load(char(spm_get_defaults('stats.results.mipmat')));

%-Single slice case
%--------------------------------------------------------------------------
if isempty(units{3}) && ~strcmp(units{2},'mm')

    %-2d case: Time-Frequency or Frequency-Frequency
    %----------------------------------------------------------------------
    mip = 4*grid_trans;
      
elseif isempty(units{3})
    
    %-2d case
    %----------------------------------------------------------------------
    mip = 4*grid_trans + mask_trans;
        
elseif strcmp(units{3},'ms') || strcmp(units{3},'Hz')
    
    %-3d case: Space-time
    %----------------------------------------------------------------------
    mip = 4*grid_time + mask_trans;

else
    %-3d case: Space
    %----------------------------------------------------------------------
    mip = 4*grid_all + mask_all;
end

%-Create maximum intensity projection
%--------------------------------------------------------------------------
mip  = mip/max(mip(:));
c    = [0 0 0 ;
        0 0 1 ;
        0 1 0 ;
        0 1 1 ;
        1 0 0 ;
        1 0 1 ; 
        1 1 0 ; 
        1 1 1 ] - 0.5;
c    = c*M(1:3,1:3);
dim  = [(max(c) - min(c)) size(mip)];
d    = spm_project(Z,round(XYZ),dim,DXYZ,CXYZ);
%if true
    mip  = max(d,Grid*mip);
    mip  = rot90((1 - mip)*64);
%else
%    mip  = rot90((1 - mip)*64);
%    d    = rot90((1 - d)*64);
%    i    = repmat(any(d~=64,3),[1 1 3]);
%    mip  = ind2rgb(ceil(mip),gray(64));
%    c    = hot(128);
%    d    = ind2rgb(ceil(d),c(end:-1:65,:));
%    mip(i) = d(i);
%end

%-And display it
%--------------------------------------------------------------------------
if ~nargout
    image(mip); axis tight; axis off;
end