%% Rest data preprocessing step..
%%%%%%%%%%%%%%% About Markers %%%%%%%%%%%%%%%%%%%%
% S 1 =  Warning Cue
% S 2 =  GO Cue
% get the EEGLAB and have Brain Vision import addon   
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Ref is Pz and Fs = 500 Hz
clear; clc; close all;

cwd = pwd; 
addpath(cwd);

datalocation= [cwd '\ALL_DATA\RAW_DATA\'];   % Data are here

%%% load subject and Behaviour info from the excel sheet
[~, ~, SubjInfo] = xlsread([cwd '\Patients_MS.xlsx'],'AllPedaling');

savelocation= [cwd '\ALL_DATA\ALL_Processed_Data\']; % Data go here
cd(savelocation);
 
% add eeglab to the path
p = genpath('eeglab13_4_4b');  addpath(p); 

%%
for nn = 1:size(SubjInfo,1)
     
    type = SubjInfo{nn,1};
    subj = SubjInfo{nn, 2};
     
    disp([type,num2str(subj)]);
    EEG = pop_loadbv(datalocation,[type,num2str(subj),'.vhdr'], [], []);
     
    % Edit the event code 
    del = 0; 
    for b = 1:size(EEG.event,2)
        if strcmp(EEG.event(b-del).type, 'boundary')
           EEG.event(b-del) = []; del = del+1;
        end
    end
    % get the time difference: S1 ---> S2
    TimeDiff = (diff([EEG.event.latency])/EEG.srate)'; TimeDiff = ([0; TimeDiff]);    
    for b = 1:size(EEG.event,2);  EEG.event(b).TimeDiff = TimeDiff(b); end
     
    % Epoch data on the basis of GO cue
    EEG = pop_epoch( EEG, { 'S  2' }, [-1 3], 'epochinfo', 'yes');
    EEG = eeg_checkset( EEG );
    
    %% Remove X,Y,Z accelerometer channels
    if EEG.nbchan>63
       EEG.X=squeeze(EEG.data(64,:,:));
       EEG.Y=squeeze(EEG.data(65,:,:));
       EEG.Z=squeeze(EEG.data(66,:,:));
       EEG.data=EEG.data(1:63,:,:);
       EEG.nbchan=63;
       EEG.chanlocs(66)=[]; EEG.chanlocs(65)=[]; EEG.chanlocs(64)=[];
    end
     
    % Fix BV-specific issue - - - only needed for APPLE
    for ai=1:size(EEG.urevent,2), EEG.urevent(ai).bvtime=EEG.urevent(ai).bvmknum; end
    for ai=1:size(EEG.event,2), EEG.event(ai).bvtime=EEG.event(ai).bvmknum; end
    for ai=1:size(EEG.epoch,2), EEG.epoch(ai).eventbvtime=EEG.epoch(ai).eventbvmknum; end
 
    % Add Pz which is Ref
    EEG = pop_chanedit(EEG,'append',63,'changefield',{64 'labels' 'Pz'});
    % EEG = pop_chanedit(EEG,'lookup', locpath); %%%% if loading channels location then use this
 
    % Re-Ref to Average Ref and recover Pz
    EEG = pop_reref(EEG,[],'refloc',struct('labels',{'Pz'},'type',{''},'theta',{180},'radius',{0.2535},'X',{-60.7385},'Y',{-7.4383e-15},'Z',{59.4629},...
    'sph_theta',{-180},'sph_phi',{44.3920},'sph_radius',{85},'urchan',{64},'ref',{''}),'keepref','on');
 
     
    % Remove 1=Fp1 and 31=Fp2: Eye blink and artifacts; 26=FT10, 10=TP9 and 20=TP10: Ear muscle artifacts: Since Pz has been reconstructed from the total
    EEG.data = EEG.data([2:9, 11:19, 21:25, 27:30, 32:64],:,:);
    EEG.nbchan=59;
    EEG.chanlocs(31)=[];  EEG.chanlocs(26)=[];  EEG.chanlocs(20)=[];  EEG.chanlocs(10)=[];  EEG.chanlocs(1)=[];  % Have to be in this order! It will delete channel locs
       
    % Should probably re-ref to average again now that the contaminated channels are gone
    EEG = pop_reref(EEG,[]);
    % Remove mean
    EEG = pop_rmbase(EEG,[],[]);
 
 
    %% ----------------------
    % Setup APPLE to interp chans, reject epochs, & ID bad ICs.  Output will be Avg ref'd and ICA'd.
    subno = subj;
    VEOG = 0;
    session = 1;
    TASK = ['Ped_',type];
     
    EEG_chans=1:EEG.nbchan;
    Do_ICA=1;
    Reref = 'FCz'; % Re-Ref to FCz   [THIS IS FOR FASTER, which is a part of APPLE]
    chanidx = find(strcmpi(Reref,{EEG.chanlocs.labels}));
    ref_chan=chanidx; 
     
    EEG = pop_reref(EEG,ref_chan,'keepref','on');
 
    % Run APPLE (will re-ref data to avg ref)
    [EEG,bad_chans,bad_epochs,bad_ICAs]=APPLE_ActiveCap_v2(EEG,EEG_chans,ref_chan,Do_ICA,subno,VEOG,session,TASK);
 
    %%% Save
    save([savelocation,type,num2str(subj),'_Ped_Processed.mat'],'EEG','bad_chans','bad_epochs','bad_ICAs');
     
    %%% for loop
    clearvars -except datalocation savelocation SubjInfo
         
end