Possible preprocessing issue

[donalbertico]

Hi fellow researchers, I've been working on reproducing SI decoding attempts and found your work, one of few that have thought about replicability.

I have found that in your preprocess steps in prerpo_DFICA_for_words.mat file :

words_data=EEG_word.(words_list{tt});
downsampleEEG=[];

      for pp=1:size(words_data,1)
         original_signal=words_data(pp,:);
         Signal_L=filter(low_pass_fliter,1,original_signal);
         Signal_LH=filter(high_pass_filter,1,Signal_L);
         downsampleData=downsample(Signal_LH,sampling_rate);
         downsampleEEG(pp,:)=downsampleData;
      end
  

You do filter and downsample the 6 channels concatenated signal which should distort it and produce a different shape than expected. I wondered if this is an error on the published code only as the preprocessed data seems to have a correct shape.