Average the betas of runs before RSA

Summary of what happened:

Hi all!
I’m trying to use the decoding_template_similarity to get the RSM of my fmri betas. I’d like to average the betas of 5 runs before calculating the similarity. However it did not work. In the template it says:

% There are two outputs that may make sense: Use 'other' if you want one
% similarity estimate per condition per run, and use 'other_average' if you
% want to average betas across runs before calculating the similarity.
cfg.results.output = 'other';

Accordingly, I set
cfg.results.output = 'other_average'
However, I still got a 60*60 (12 conditions * 5 runs) matrix in the results. Can you point out to me where was wrong?
Thanks so much!

Command used (and if a helper script was used, a link to the helper script or the command generated):

        cfg = decoding_defaults();
        cfg.analysis = 'roi' ;
        cfg.files.mask = ROIdir{theroi};
        cfg.results.overwrite = 1;
        cfg.results.dir = resultDir;
        labelnames = Condis;
        % since the labels are arbitrary, we will set them randomly to -1 and 1
        labels(1:2:length(labelnames)) = -1;
        labels(2:2:length(labelnames)) =  1;
        % set everything to similarity analysis (for available options as model parameters, check decoding_software/pattern_similarity/pattern_similarity.m)
        cfg.decoding.software = 'similarity';
        cfg.decoding.method = 'classification';
        cfg.decoding.train.classification.model_parameters = 'pearson'; % this is pearson correlation
        cfg.results.output = 'other_average';
        cfg.scale.method = 'min0max1';
        cfg.scale.estimation = 'all'; % scaling across all data is equivalent to no scaling (i.e. will yield the same results), it only changes the data range which allows libsvm to compute faster
        % The following function extracts all beta names and corresponding run
        % numbers from the SPM.mat
        regressor_names = design_from_spm(betaDir);
        % Extract all information for the cfg.files structure (labels will be [1 -1] )
        cfg = decoding_describe_data(cfg,labelnames,labels,regressor_names,betaDir);
        % This creates a design in which all data is used to calculate the similarity
        cfg.design = make_design_similarity(cfg);
        results = decoding(cfg);


Environment (Docker, Singularity, custom installation):

Data formatted according to a validatable standard? Please provide the output of the validator:

Relevant log outputs (up to 20 lines):

Screenshots / relevant information: