Skip to content
Switch branches/tags
Go to file
12 contributors

Users who have contributed to this file

@robertoostenveld @schoffelen @roemervandermeij @Spaak @haegens @jimher @Richter2552 @dmatthes1982 @NirOfir @matsvanes @demaddis @raghav0706
function [cfg] = ft_singleplotTFR(cfg, data)
% FT_SINGLEPLOTTFR plots the time-frequency representation of power of a
% single channel or the average over multiple channels.
% Use as
% ft_singleplotTFR(cfg,data)
% The input freq structure should be a a time-frequency representation of
% power or coherence that was computed using the FT_FREQANALYSIS function.
% The configuration can have the following parameters:
% cfg.parameter = field to be plotted on z-axis, e.g. 'powspctrm' (default depends on data.dimord)
% cfg.maskparameter = field in the data to be used for masking of data, can be logical (e.g. significant data points) or numerical (e.g. t-values).
% (not possible for mean over multiple channels, or when input contains multiple subjects
% or trials)
% cfg.maskstyle = style used to masking, 'opacity', 'saturation', or 'outline' (default = 'opacity')
% 'outline' can only be used with a logical cfg.maskparameter
% use 'saturation' or 'outline' when saving to vector-format (like *.eps) to avoid all sorts of image-problems
% cfg.maskalpha = alpha value between 0 (transparent) and 1 (opaque) used for masking areas dictated by cfg.maskparameter (default = 1)
% (will be ignored in case of numeric cfg.maskparameter or if cfg.maskstyle = 'outline')
% cfg.masknans = 'yes' or 'no' (default = 'yes')
% cfg.xlim = 'maxmin' or [xmin xmax] (default = 'maxmin')
% cfg.ylim = 'maxmin' or [ymin ymax] (default = 'maxmin')
% cfg.zlim = plotting limits for color dimension, 'maxmin', 'maxabs', 'zeromax', 'minzero', or [zmin zmax] (default = 'maxmin')
% cfg.baseline = 'yes', 'no' or [time1 time2] (default = 'no'), see FT_FREQBASELINE
% cfg.baselinetype = 'absolute', 'relative', 'relchange', 'normchange', 'db' or 'zscore' (default = 'absolute')
% cfg.trials = 'all' or a selection given as a 1xN vector (default = 'all')
% = Nx1 cell-array with selection of channels (default = 'all'),
% see FT_CHANNELSELECTION for details
% cfg.title = string, title of plot
% cfg.refchannel = name of reference channel for visualising connectivity, can be 'gui'
% cfg.fontsize = font size of title (default = 8)
% cfg.hotkeys = enables hotkeys (leftarrow/rightarrow/uparrow/downarrow/pageup/pagedown/m) for dynamic zoom and translation (ctrl+) of the axes and color limits
% cfg.colormap = string, or Nx3 matrix, see FT_COLORMAP
% cfg.colorbar = 'yes', 'no' (default = 'yes')
% cfg.colorbartext = string indicating the text next to colorbar
% cfg.interactive = interactive plot 'yes' or 'no' (default = 'yes')
% In a interactive plot you can select areas and produce a new
% interactive plot when a selected area is clicked. Multiple areas
% can be selected by holding down the SHIFT key.
% cfg.renderer = 'painters', 'zbuffer', ' opengl' or 'none' (default = [])
% cfg.directionality = '', 'inflow' or 'outflow' specifies for
% connectivity measures whether the inflow into a
% node, or the outflow from a node is plotted. The
% (default) behavior of this option depends on the dimor
% of the input data (see below).
% The following options for the scaling of the EEG, EOG, ECG, EMG, MEG and NIRS channels
% is optional and can be used to bring the absolute numbers of the different
% channel types in the same range (e.g. fT and uV). The channel types are determined
% from the input data using FT_CHANNELSELECTION.
% cfg.eegscale = number, scaling to apply to the EEG channels prior to display
% cfg.eogscale = number, scaling to apply to the EOG channels prior to display
% cfg.ecgscale = number, scaling to apply to the ECG channels prior to display
% cfg.emgscale = number, scaling to apply to the EMG channels prior to display
% cfg.megscale = number, scaling to apply to the MEG channels prior to display
% cfg.gradscale = number, scaling to apply to the MEG gradiometer channels prior to display (in addition to the cfg.megscale factor)
% cfg.magscale = number, scaling to apply to the MEG magnetometer channels prior to display (in addition to the cfg.megscale factor)
% cfg.nirsscale = number, scaling to apply to the NIRS channels prior to display
% cfg.mychanscale = number, scaling to apply to the channels specified in cfg.mychan
% cfg.mychan = Nx1 cell-array with selection of channels
% cfg.chanscale = Nx1 vector with scaling factors, one per channel specified in
% For the plotting of directional connectivity data the cfg.directionality
% option determines what is plotted. The default value and the supported
% functionality depend on the dimord of the input data. If the input data
% is of dimord 'chan_chan_XXX', the value of directionality determines
% whether, given the reference channel(s), the columns (inflow), or rows
% (outflow) are selected for plotting. In this situation the default is
% 'inflow'. Note that for undirected measures, inflow and outflow should
% give the same output. If the input data is of dimord 'chancmb_XXX', the
% value of directionality determines whether the rows in data.labelcmb are
% selected. With 'inflow' the rows are selected if the refchannel(s) occur in
% the right column, with 'outflow' the rows are selected if the
% refchannel(s) occur in the left column of the labelcmb-field. Default in
% this case is '', which means that all rows are selected in which the
% refchannel(s) occur. This is to robustly support linearly indexed
% undirected connectivity metrics. In the situation where undirected
% connectivity measures are linearly indexed, specifying 'inflow' or
% 'outflow' can result in unexpected behavior.
% Copyright (C) 2005-2017, F.C. Donders Centre
% This file is part of FieldTrip, see
% for the documentation and details.
% FieldTrip is free software: you can redistribute it and/or modify
% it under the terms of the GNU General Public License as published by
% the Free Software Foundation, either version 3 of the License, or
% (at your option) any later version.
% FieldTrip is distributed in the hope that it will be useful,
% but WITHOUT ANY WARRANTY; without even the implied warranty of
% GNU General Public License for more details.
% You should have received a copy of the GNU General Public License
% along with FieldTrip. If not, see <>.
% $Id$
% DEVELOPERS NOTE: This code is organized in a similar fashion for multiplot/singleplot/topoplot
% and for ER/TFR and should remain consistent over those 6 functions.
% Section 1: general cfg handling that is independent from the data
% Section 2: data handling, this also includes converting bivariate (chan_chan and chancmb) into univariate data
% Section 3: select the data to be plotted and determine min/max range
% Section 4: do the actual plotting
%% Section 1: general cfg handling that is independent from the data
% these are used by the ft_preamble/ft_postamble function and scripts
ft_revision = '$Id$';
ft_nargin = nargin;
ft_nargout = nargout;
% do the general setup of the function
ft_preamble init
ft_preamble debug
ft_preamble loadvar data
ft_preamble provenance data
ft_preamble trackconfig
% the ft_abort variable is set to true or false in ft_preamble_init
if ft_abort
% check if the input data is valid for this function
data = ft_checkdata(data, 'datatype', 'freq');
% check if the input cfg is valid for this function
cfg = ft_checkconfig(cfg, 'forbidden', {'channels', 'trial'}); % prevent accidental typos, see issue 1729
cfg = ft_checkconfig(cfg, 'unused', {'cohtargetchannel'});
cfg = ft_checkconfig(cfg, 'renamed', {'matrixside', 'directionality'});
cfg = ft_checkconfig(cfg, 'renamedval', {'zlim', 'absmax', 'maxabs'});
cfg = ft_checkconfig(cfg, 'renamedval', {'directionality', 'feedforward', 'outflow'});
cfg = ft_checkconfig(cfg, 'renamedval', {'directionality', 'feedback', 'inflow'});
cfg = ft_checkconfig(cfg, 'renamed', {'channelindex', 'channel'});
cfg = ft_checkconfig(cfg, 'renamed', {'channelname', 'channel'});
cfg = ft_checkconfig(cfg, 'renamed', {'cohrefchannel', 'refchannel'});
cfg = ft_checkconfig(cfg, 'renamed', {'zparam', 'parameter'});
cfg = ft_checkconfig(cfg, 'renamed', {'newfigure', 'figure'});
% Set the defaults
cfg.baseline = ft_getopt(cfg, 'baseline', 'no');
cfg.baselinetype = ft_getopt(cfg, 'baselinetype', 'absolute');
cfg.trials = ft_getopt(cfg, 'trials', 'all', 1);
cfg.xlim = ft_getopt(cfg, 'xlim', 'maxmin');
cfg.ylim = ft_getopt(cfg, 'ylim', 'maxmin');
cfg.zlim = ft_getopt(cfg, 'zlim', 'maxmin');
cfg.fontsize = ft_getopt(cfg, 'fontsize', 8);
cfg.colorbar = ft_getopt(cfg, 'colorbar', 'yes');
cfg.colormap = ft_getopt(cfg, 'colormap', 'default');
cfg.colorbartext = ft_getopt(cfg, 'colorbartext', '');
cfg.interactive = ft_getopt(cfg, 'interactive', 'yes');
cfg.hotkeys = ft_getopt(cfg, 'hotkeys', 'yes');
cfg.maskalpha = ft_getopt(cfg, 'maskalpha', 1);
cfg.maskparameter = ft_getopt(cfg, 'maskparameter', []);
cfg.maskstyle = ft_getopt(cfg, 'maskstyle', 'opacity'); = ft_getopt(cfg, 'channel', 'all');
cfg.title = ft_getopt(cfg, 'title', []);
cfg.masknans = ft_getopt(cfg, 'masknans', 'yes');
cfg.directionality = ft_getopt(cfg, 'directionality', []);
cfg.figurename = ft_getopt(cfg, 'figurename', []);
cfg.parameter = ft_getopt(cfg, 'parameter', 'powspctrm');
cfg.renderer = ft_getopt(cfg, 'renderer', []); % let MATLAB decide on the default
% this is needed for the figure title
if isfield(cfg, 'dataname') && ~isempty(cfg.dataname)
dataname = cfg.dataname;
elseif isfield(cfg, 'inputfile') && ~isempty(cfg.inputfile)
dataname = cfg.inputfile;
elseif nargin>1
dataname = arrayfun(@inputname, 2:nargin, 'UniformOutput', false);
dataname = {};
%% Section 2: data handling, this also includes converting bivariate (chan_chan and chancmb) into univariate data
hastime = isfield(data, 'time');
hasfreq = isfield(data, 'freq');
assert((hastime && hasfreq), 'please use ft_singleplotER for time-only or frequency-only data');
xparam = ft_getopt(cfg, 'xparam', 'time');
yparam = ft_getopt(cfg, 'yparam', 'freq');
% check whether rpt/subj is present and remove if necessary
dimord = getdimord(data, cfg.parameter);
dimtok = tokenize(dimord, '_');
hasrpt = any(ismember(dimtok, {'rpt' 'subj'}));
if ~hasrpt
assert(isequal(cfg.trials, 'all') || isequal(cfg.trials, 1), 'incorrect specification of cfg.trials for data without repetitions');
assert(~isempty(cfg.trials), 'empty specification of cfg.trials for data with repetitions');
% parse
if isfield(cfg, 'channel') && isfield(data, 'label') = ft_channelselection(, data.label);
elseif isfield(cfg, 'channel') && isfield(data, 'labelcmb') = ft_channelselection(, unique(data.labelcmb(:)));
% Apply baseline correction:
if ~strcmp(cfg.baseline, 'no')
tmpcfg = keepfields(cfg, {'baseline', 'baselinetype', 'baselinewindow', 'demean', 'parameter', 'channel'});
% keep mask-parameter if it is set
if ~isempty(cfg.maskparameter)
tempmask = data.(cfg.maskparameter);
data = ft_freqbaseline(tmpcfg, data);
% put mask-parameter back if it is set
if ~isempty(cfg.maskparameter)
data.(cfg.maskparameter) = tempmask;
% channels should NOT be selected and averaged here, since a topoplot might follow in interactive mode
tmpcfg = keepfields(cfg, {'trials', 'showcallinfo', 'trackcallinfo', 'trackconfig', 'trackusage', 'trackdatainfo', 'trackmeminfo', 'tracktimeinfo'});
if hasrpt
tmpcfg.avgoverrpt = 'yes';
tmpcfg.avgoverrpt = 'no';
tmpvar = data;
[data] = ft_selectdata(tmpcfg, data);
% restore the provenance information and put back
tmpchannel =;
[cfg, data] = rollback_provenance(cfg, data); = tmpchannel;
if isfield(tmpvar, cfg.maskparameter) && ~isfield(data, cfg.maskparameter)
% the mask parameter is not present after ft_selectdata, because it is
% not included in all input arguments. Make the same selection and copy
% it over
tmpvar = ft_selectdata(tmpcfg, tmpvar);
data.(cfg.maskparameter) = tmpvar.(cfg.maskparameter);
clear tmpvar tmpcfg dimord dimtok hastime hasfreq hasrpt
% ensure that the preproc specific options are located in the cfg.preproc
% substructure, but also ensure that the field 'refchannel' remains at the
% highest level in the structure. This is a little hack by JM because the field
% refchannel can relate to connectivity or to an EEg reference.
if isfield(cfg, 'refchannel'), refchannelincfg = cfg.refchannel; cfg = rmfield(cfg, 'refchannel'); end
cfg = ft_checkconfig(cfg, 'createsubcfg', {'preproc'});
if exist('refchannelincfg', 'var'), cfg.refchannel = refchannelincfg; end
if ~isempty(cfg.preproc)
% preprocess the data, i.e. apply filtering, baselinecorrection, etc.
fprintf('applying preprocessing options\n');
if ~isfield(cfg.preproc, 'feedback') = cfg.interactive;
data = ft_preprocessing(cfg.preproc, data);
% Handle the bivariate case
dimord = getdimord(data, cfg.parameter);
if startsWith(dimord, 'chan_chan_') || startsWith(dimord, 'chancmb_')
% convert the bivariate data to univariate and call this plotting function again
cfg.originalfunction = 'ft_singleplotTFR';
cfg.trials = 'all'; % trial selection has been taken care off
bivariate_common(cfg, data);
% Apply channel-type specific scaling
fn = fieldnames(cfg);
fn = setdiff(fn, {'skipscale', 'showscale', 'gridscale'}); % these are for the layout and plotting, not for CHANSCALE_COMMON
fn = fn(endsWith(fn, 'scale') | startsWith(fn, 'mychan') | strcmp(fn, 'channel') | strcmp(fn, 'parameter'));
tmpcfg = keepfields(cfg, fn);
if ~isempty(tmpcfg)
data = chanscale_common(tmpcfg, data);
% remove the scaling fields from the configuration, to prevent them from being called again in interactive mode
% but keep the parameter and channel field
cfg = removefields(cfg, setdiff(fn, {'parameter', 'channel'}));
% do nothing
%% Section 3: select the data to be plotted and determine min/max range
% Take the subselection of channels that is contained in the layout, this is the same in all datasets
[selchan] = match_str(data.label,;
% Get physical min/max range of x, i.e. time
if strcmp(cfg.xlim, 'maxmin')
xmin = min(data.(xparam));
xmax = max(data.(xparam));
xmin = cfg.xlim(1);
xmax = cfg.xlim(2);
% Get the index of the nearest bin
xminindx = nearest(data.(xparam), xmin);
xmaxindx = nearest(data.(xparam), xmax);
xmin = data.(xparam)(xminindx);
xmax = data.(xparam)(xmaxindx);
selx = xminindx:xmaxindx;
xval = data.(xparam)(selx);
% Get physical min/max range of y, i.e. frequency
if strcmp(cfg.ylim, 'maxmin')
ymin = min(data.(yparam));
ymax = max(data.(yparam));
ymin = cfg.ylim(1);
ymax = cfg.ylim(2);
% Get the index of the nearest bin
yminindx = nearest(data.(yparam), ymin);
ymaxindx = nearest(data.(yparam), ymax);
ymin = data.(yparam)(yminindx);
ymax = data.(yparam)(ymaxindx);
sely = yminindx:ymaxindx;
yval = data.(yparam)(sely);
% test if X and Y are linearly spaced (to within 10^-12): % FROM UIMAGE
dx = min(diff(xval)); % smallest interval for X
dy = min(diff(yval)); % smallest interval for Y
evenx = all(abs(diff(xval)/dx-1)<1e-12); % true if X is linearly spaced
eveny = all(abs(diff(yval)/dy-1)<1e-12); % true if Y is linearly spaced
if ~evenx || ~eveny
ft_warning('(one of the) axis is/are not evenly spaced, but plots are made as if axis are linear')
% masking is only possible for evenly spaced axis
if strcmp(cfg.masknans, 'yes') && (~evenx || ~eveny)
ft_warning('(one of the) axis are not evenly spaced -> nans cannot be masked out -> cfg.masknans is set to ''no'';')
cfg.masknans = 'no';
% the usual data is chan_freq_time, but other dimords should also work
dimtok = tokenize(dimord, '_');
datamatrix = data.(cfg.parameter);
[c, ia, ib] = intersect({'chan', yparam, xparam}, dimtok, 'stable');
datamatrix = permute(datamatrix, ib);
datamatrix = datamatrix(selchan, sely, selx);
if ~isempty(cfg.maskparameter)
maskmatrix = data.(cfg.maskparameter)(selchan, sely, selx);
if islogical(maskmatrix) && any(strcmp(cfg.maskstyle, {'saturation', 'opacity'}))
maskmatrix = double(maskmatrix);
maskmatrix(~maskmatrix) = cfg.maskalpha;
elseif isnumeric(maskmatrix)
if strcmp(cfg.maskstyle, 'outline')
ft_error('Outline masking with a numeric cfg.maskparameter is not supported. Please use a logical mask instead.')
if cfg.maskalpha ~= 1
ft_warning('Using field "%s" for masking, cfg.maskalpha is ignored.', cfg.maskparameter)
% scale mask between 0 and 1
minval = min(maskmatrix(:));
maxval = max(maskmatrix(:));
maskmatrix = (maskmatrix - minval) / (maxval-minval);
% create an Nx0x0 matrix
maskmatrix = zeros(length(selchan), 0, 0);
%% Section 4: do the actual plotting
% open a new figure, or add it to the existing one
% note that in general adding a TFR to an existing one does not make sense, since they will overlap
open_figure(keepfields(cfg, {'figure', 'position', 'visible', 'renderer', 'figurename', 'title'}));
zval = mean(datamatrix, 1); % over channels
zval = reshape(zval, size(zval,2), size(zval,3));
mask = squeeze(mean(maskmatrix, 1)); % over channels
% Get physical z-axis range (color axis):
if strcmp(cfg.zlim, 'maxmin')
zmin = nanmin(zval(:));
zmax = nanmax(zval(:));
elseif strcmp(cfg.zlim, 'maxabs')
zmin = -nanmax(abs(zval(:)));
zmax = nanmax(abs(zval(:)));
elseif strcmp(cfg.zlim, 'zeromax')
zmin = 0;
zmax = nanmax(zval(:));
elseif strcmp(cfg.zlim, 'minzero')
zmin = nanmin(zval(:));
zmax = 0;
zmin = cfg.zlim(1);
zmax = cfg.zlim(2);
% Draw the data and mask NaN's if requested
if isequal(cfg.masknans, 'yes') && isempty(cfg.maskparameter)
nans_mask = ~isnan(zval);
mask = double(nans_mask);
ft_plot_matrix(xval, yval, zval, 'clim', [zmin zmax], 'tag', 'cip', 'highlightstyle', cfg.maskstyle, 'highlight', mask)
elseif isequal(cfg.masknans, 'yes') && ~isempty(cfg.maskparameter)
nans_mask = ~isnan(zval);
mask = mask .* nans_mask;
mask = double(mask);
ft_plot_matrix(xval, yval, zval, 'clim', [zmin zmax], 'tag', 'cip', 'highlightstyle', cfg.maskstyle, 'highlight', mask)
elseif isequal(cfg.masknans, 'no') && ~isempty(cfg.maskparameter)
mask = double(mask);
ft_plot_matrix(xval, yval, zval, 'clim', [zmin zmax], 'tag', 'cip', 'highlightstyle', cfg.maskstyle, 'highlight', mask)
ft_plot_matrix(xval, yval, zval, 'clim', [zmin zmax], 'tag', 'cip')
% check if the colormap is in the proper format and set it
if ~isequal(cfg.colormap, 'default')
if ischar(cfg.colormap)
cfg.colormap = ft_colormap(cfg.colormap);
elseif iscell(cfg.colormap)
cfg.colormap = ft_colormap(cfg.colormap{:});
elseif isnumeric(cfg.colormap) && size(cfg.colormap,2)~=3
ft_error('colormap must be a Nx3 matrix');
set(gcf, 'colormap', cfg.colormap);
axis xy
if isequal(cfg.colorbar, 'yes')
c = colorbar;
ylabel(c, cfg.colorbartext);
% Set callback to adjust color axis
if strcmp('yes', cfg.hotkeys)
% Attach data and cfg to figure and attach a key listener to the figure
set(gcf, 'KeyPressFcn', {@key_sub, xmin, xmax, ymin, ymax, zmin, zmax})
% Create axis title containing channel name(s) and channel number(s):
if ~isempty(cfg.title)
t = cfg.title;
if length( == 1
t = [char( ' / ' num2str(selchan) ];
t = sprintf('mean(%0s)', join_str(', ',;
title(t, 'fontsize', cfg.fontsize);
% set the figure window title, add channel labels if number is small
if isempty(get(gcf, 'Name'))
if length(selchan) < 5
chans = join_str(', ',;
chans = '<multiple channels>';
if ~isempty(cfg.figurename)
set(gcf, 'name', cfg.figurename);
set(gcf, 'NumberTitle', 'off');
elseif ~isempty(dataname)
set(gcf, 'Name', sprintf('%d: %s: %s (%s)', double(gcf), mfilename, join_str(', ', dataname), chans));
set(gcf, 'NumberTitle', 'off');
set(gcf, 'Name', sprintf('%d: %s (%s)', double(gcf), mfilename, chans));
set(gcf, 'NumberTitle', 'off');
axis tight
% Make the figure interactive
if strcmp(cfg.interactive, 'yes')
% add the cfg/data information to the figure under identifier linked to this axis
ident = ['axh' num2str(round(sum(clock.*1e6)))]; % unique identifier for this axis
set(gca, 'tag',ident);
info = guidata(gcf);
info.(ident).dataname = dataname;
info.(ident).cfg = cfg;
info.(ident).data = data;
guidata(gcf, info);
set(gcf, 'WindowButtonUpFcn', {@ft_select_range, 'multiple', false, 'callback', {@select_topoplotTFR}, 'event', 'WindowButtonUpFcn'});
set(gcf, 'WindowButtonDownFcn', {@ft_select_range, 'multiple', false, 'callback', {@select_topoplotTFR}, 'event', 'WindowButtonDownFcn'});
set(gcf, 'WindowButtonMotionFcn', {@ft_select_range, 'multiple', false, 'callback', {@select_topoplotTFR}, 'event', 'WindowButtonMotionFcn'});
% do the general cleanup and bookkeeping at the end of the function
ft_postamble debug
ft_postamble trackconfig
ft_postamble previous data
ft_postamble provenance
ft_postamble savefig
% add a menu to the figure, but only if the current figure does not have subplots
menu_fieldtrip(gcf, cfg, false);
if ~ft_nargout
% don't return anything
clear cfg
% SUBFUNCTION which is called after selecting a time range
function select_topoplotTFR(range, varargin)
% fetch cfg/data based on axis indentifier given as tag
ident = get(gca, 'tag');
info = guidata(gcf);
cfg = info.(ident).cfg;
data = info.(ident).data;
if ~isempty(range)
cfg = removefields(cfg, 'inputfile'); % the reading has already been done and varargin contains the data
cfg = removefields(cfg, 'showlabels'); % this is not allowed in topoplotER
cfg.trials = 'all'; % trial selection has already been taken care of
cfg.baseline = 'no'; % make sure the next function does not apply a baseline correction again = 'all'; % make sure the topo displays all channels, not just the ones in this singleplot
cfg.comment = 'auto';
cfg.dataname = info.(ident).dataname; % put data name in here, this cannot be resolved by other means
cfg.xlim = range(1:2);
cfg.ylim = range(3:4);
fprintf('selected cfg.xlim = [%f %f]\n', cfg.xlim(1), cfg.xlim(2));
fprintf('selected cfg.ylim = [%f %f]\n', cfg.ylim(1), cfg.ylim(2));
% ensure that the new figure appears at the same position
cfg.figure = 'yes';
cfg.position = get(gcf, 'Position');
ft_topoplotTFR(cfg, data);
% SUBFUNCTION which handles hot keys in the current plot
function key_sub(handle, eventdata, varargin)
xlimits = xlim;
ylimits = ylim;
climits = caxis;
incr_x = abs(xlimits(2) - xlimits(1)) /10;
incr_y = abs(ylimits(2) - ylimits(1)) /10;
incr_c = abs(climits(2) - climits(1)) /10;
if length(eventdata.Modifier) == 1 && strcmp(eventdata.Modifier{:}, 'control')
% TRANSLATE by 10%
switch eventdata.Key
case 'pageup'
caxis([min(caxis)+incr_c max(caxis)+incr_c]);
case 'pagedown'
caxis([min(caxis)-incr_c max(caxis)-incr_c]);
case 'leftarrow'
xlim([xlimits(1)+incr_x xlimits(2)+incr_x])
case 'rightarrow'
xlim([xlimits(1)-incr_x xlimits(2)-incr_x])
case 'uparrow'
ylim([ylimits(1)-incr_y ylimits(2)-incr_y])
case 'downarrow'
ylim([ylimits(1)+incr_y ylimits(2)+incr_y])
end % switch
% ZOOM by 10%
switch eventdata.Key
case 'pageup'
caxis([min(caxis)-incr_c max(caxis)+incr_c]);
case 'pagedown'
caxis([min(caxis)+incr_c max(caxis)-incr_c]);
case 'leftarrow'
xlim([xlimits(1)-incr_x xlimits(2)+incr_x])
case 'rightarrow'
xlim([xlimits(1)+incr_x xlimits(2)-incr_x])
case 'uparrow'
ylim([ylimits(1)-incr_y ylimits(2)+incr_y])
case 'downarrow'
ylim([ylimits(1)+incr_y ylimits(2)-incr_y])
case 'm'
xlim([varargin{1} varargin{2}])
ylim([varargin{3} varargin{4}])
caxis([varargin{5} varargin{6}]);
end % switch
end % if