LIMO.cache

Checking my last whole space analysis

Whenever you ask for results at a given threshold and method (say TDCE at p<0.05), an analysis is run to threshold data and get the equivalent under the null (i.e. on bootstrapped data if you have computed them). That result is saved in LIMO.cache, which is primarily intended for quickly redisplaying the results. Note that each time you ask for a new result (a different effect, or the same effect but with a different threshold or method), the cache is overwritten. Only the last result is cached.

Cached variables

LIMO.cache.fig is the result of the whole space analysis when clicking 'Image all'.

        name: 'xxx.mat'             % the file you choose to look at
          MCC: 2                    % the method used 1=None, 2=Cluster, 3=TFCE, 4=max
        stats: [32×188 double]      % the t or F values 
    threshold: 0.0500               % the alpha threshold used 
         pval: [32×188 double]      % the (corrected) p-values
         mask: [32×188 double]      % the mask of significant cells (clusters are numbered)
        title: 'XXX t/ F values cluster correction (N clusters)'

LIMO.cache.Courseplot is the result of the plot done at a given channel for ERP or Power at the subject level

       extra: 'Modelled'           % data type plotted (original data, modelled or adjusted)
      average: [4×750 double]      % mean values per regressor
      channel: XX                  % channel selected
    regressor: [1 2 3 4]           % which regressor were plotted
           ci: [4×2×750 double]    % 95% CI of each regressor
        title: 'Modelled ERP at channel XX (XX)'

What can you do with those cached data?

• how many clusters? max(unique(LIMO.cache.fig.mask(:)))
• what is the p-value of cluster N? LIMO.cache.fig.pval(LIMO.cache.fig.mask == N)
• what is the largest p-value? max(LIMO.cache.fig.pval(LIMO.cache.fig.mask ~= 0))
• what is the range of significant F/t values? [min(LIMO.cache.fig.stats(LIMO.cache.fig.mask ~= 0)) max(LIMO.cache.fig.stats(LIMO.cache.fig.mask ~= 0))] etc ..

You can atually replot using limo_display_image(LIMO,LIMO.cache.fig.stats,LIMO.cache.fig.mask,'your own title,0)