Plot the results

Result representation pipeline

In addition to the graphic user interface, MVPAlab implements different high-level functions to generate highly-customizable graphical representation of the results. Once the decoding analysis is completed and the results files are saved, the graphical representation pipeline runs as follows:

% 1. Initialize and configure plots:
graph = mvpalab_plotinit();

% 2. Load results and and statistics if needed:
load results/time_resolved/acc/result.mat
load results/time_resolved/acc/stats.mat

% 3. Some extra graphic configuration:

% 3.1 Axis limits:
graph.xlim = [-200 1500];
graph.ylim = [.3 .95];

% 3.2 Axes labels and titles:
graph.xlabel = 'Time (ms)';
graph.ylabel = 'Classifier performance';
graph.title = 'Demo plot (no statistical significance)';

% 3.3 Smooth results:
graph.smoothdata = 5; % (1 => no smoothing)

% 3.4 Plot significant clusters (above and below chance):
graph.stats.above = true;
graph.stats.below = true;
graph.sigh = .4;

% 4. Plot results:
figure;
hold on
mvpalab_plotdecoding(graph,cfg,result,stats);

First, the function mvpalab_plotinit() generates and returns a default configuration structure (graph) containing all the required configuration parameter:

graph = mvpalab_plotinit();

Then, the specific result file (and stats if needed) to be plotted should be loaded:

load results/time_resolved/acc/result.mat
load results/time_resolved/acc/stats.mat

Finally, the high-level plotting function returns the graphical representation of the selected result file:

mvpalab_plotdecoding(graph,cfg,result,stats);

The variable stats is optional and contains, among others, the statistically significant clusters. If this variable is not omitted, significant results will be highlighted in the resulting figure.

Several plotting functions are available for different types of analysis:

mvpalab_plotdecoding(graph,cfg,result,stats);
mvpalab_plottempogen(graph,cfg,result,stats);
mvpalab_plotfreqcont(graph,cfg,result,stats);
mvpalab_plotfeatcont(graph,cfg,wvector,result);

• The mvpalab_plotdecoding() function generates time-resolved performance plots.
• The mvpalab_plottempogen() function is used for the graphical representation of temporal generalization matrices.
• The mvpalab_plotslidfilt() function generates the graphical representation for the frequency contribution analysis.
• The mvpalab_plotfeatcont() can generate topological representations and temporal animations of features contribution to the decoding performance.

Extra configuration parameters:

To get the best of the MVPAlab Toolbox plotting capabilities the use of the graphic user interface is highly recommended. This is a fast, flexible and very intuitive manner to design high-quality plots. Even so, the same results can be obtained by hand coding several configuration parameters included in the graph configuration structure.

A complete selection of the most useful configuration parameters and a short explanation is listed below:

% Time-resolved decoding analysis:
% --------------------------------------------------
graph.plotmean   = true; % Plot group average
graph.smoothdata = 5;    % Window size for smooth
graph.stdsem     = true; % Plot STD or SEM
graph.linestyle  = '-';  % Line style
graph.linewidth  = 1;    % Line width

% 2D decoding analysis (TEMPOGEN or FREQCONT):
% --------------------------------------------------
graph.clusterLineColor = [0 0 0]; % Cluster color.
graph.clusterLineWidth = 1;       % Cluster width.
graph.caxis = [.4 .9];            % Color range.

% Feature contribution analysis:
% --------------------------------------------------
graph.weights.type  = 'raw'; % Raw or corrected
graph.weights.anim  = true;  % Animated/static plot
graph.weights.speed = 0.1;   % Animation speed
graph.weights.start = 400;   % Start time (ms) 
graph.weights.end   = 450;   % End time (ms)
graph.weights.sub   = 1;     % Individual subject

% Highlight significant result:
% --------------------------------------------------
graph.sigmode.points = true; % Points/shade plot
graph.stats.above    = true; % Above chance clusters
graph.stats.below    = true; % Below chance clusters
graph.sigh           = 0.4;  % Sig. points height 

% Font, titles, labels and axes limits:
% --------------------------------------------------
graph.fontsize = 14;
graph.title    = 'MVPAlab - default figure';
graph.ylabel   = 'Classifier performance';
graph.xlabel   = 'Time (ms)';
graph.xlim     = [-200 1500];
graph.ylim     = [0 1];

% Individual subject plots:
% --------------------------------------------------
graph.subject = 3; % Subject idx (individual plot)