% An example to get the BCI competition IV datasets 2a, 2b is the same

% Data from: http://www.bbci.de/competition/iv/

% using open-source toolbox Biosig on MATLAB: http://biosig.sourceforge.net/

% Just an example, you should change as you need.

function data = getData(subject_index)

subject_index = 6; % 1-9

%% T data

session_type = 'T'; % T and E

dir_1 = ['D:\MI\BCICIV_2a_gdf\A0',num2str(subject_index),session_type,'.gdf']; % set your path of the downloaded data

[s, HDR] = sload(dir_1);

% Label 

% label = HDR.Classlabel;

labeldir_1 = ['D:\MI\true_labels\A0',num2str(subject_index),session_type,'.mat'];

load(labeldir_1);

label_1 = classlabel;

% construct sample - data Section 1000*22*288

Pos = HDR.EVENT.POS; % use POS to get trials

% Dur = HDR.EVENT.DUR;

Typ = HDR.EVENT.TYP;

k = 0;

data_1 = zeros(1000,22,288);

for j = 1:length(Typ)

    if  Typ(j) == 768

        k = k+1;

        data_1(:,:,k) = s((Pos(j)+500):(Pos(j)+1499),1:22);

    end

end

% wipe off NaN

data_1(isnan(data_1)) = 0;

% E data

session_type = 'E';

dir_2 = ['D:\Lab\MI\BCICIV_2a_gdf\A0',num2str(subject_index),session_type,'.gdf'];

% dir = 'D:\Lab\MI\BCICIV_2a_gdf\A01E.gdf';

[s, HDR] = sload(dir_2);

% Label 

% label = HDR.Classlabel;

labeldir_2 = ['D:\Lab\MI\true_labels\A0',num2str(subject_index),session_type,'.mat'];

load(labeldir_2);

label_2 = classlabel;

% construct sample - data Section 1000*22*288

Pos = HDR.EVENT.POS;

% Dur = HDR.EVENT.DUR;

Typ = HDR.EVENT.TYP;

k = 0;

data_2 = zeros(1000,22,288);

for j = 1:length(Typ)

    if  Typ(j) == 768

        k = k+1;

        data_2(:,:,k) = s((Pos(j)+500):(Pos(j)+1499),1:22);

    end

end

% wipe off NaN

data_2(isnan(data_2)) = 0;

%% preprocessing

% option - band-pass filter

fc = 250; % sampling rate

Wl = 4; Wh = 40; % pass band

Wn = [Wl*2 Wh*2]/fc;

[b,a]=cheby2(6,60,Wn);

% a better filter for 4-40 Hz band-pass

% fc = 250;

% Wl = 4; Wh = 40; 

% Wn = [Wl*2 Wh*2]/fc;

% [b,a]=cheby2(8,20,Wn);

for j = 1:288

    data_1(:,:,j) = filtfilt(b,a,data_1(:,:,j));

    data_2(:,:,j) = filtfilt(b,a,data_2(:,:,j));

end

% option - a simple standardization

%{

eeg_mean = mean(data,3);

eeg_std = std(data,1,3); 

fb_data = (data-eeg_mean)./eeg_std;

%}

%% Save the data to a mat file 

data = data_1;

label = label_1;

% label = t_label + 1;

saveDir = ['D:\MI\standard_2a_data\A0',num2str(subject_index),'T.mat'];

save(saveDir,'data','label');

data = data_2;

label = label_2;

saveDir = ['D:\MI\standard_2a_data\A0',num2str(subject_index),'E.mat'];

save(saveDir,'data','label');

end