This code will extract nback-features from the data. line 7,8 loads correct response indices. Modifications are from line 40-42, for MATLAB2016a version compatibility.

nback_features_test.m

clear all;
clc;

% loading Preprocessed_2s, which contains data
load('Preprocessed_response_2s.mat')
% loading correct respose indices 
load('etc\response_fld0.05_correct.mat');
% load('etc\response_correct_fld0.05.mat');

S=[];
for s = 1:length(Preprocessed_2s)
%     if any(s==[3,4,12,23])
%         continue;
%     end
    % subjects exluded
    if any(s ==[3,4,12,19,22,23,33,34])
            continue;
    end
    % extracting n-back features from the remaing subjects data
    fprintf('Subject %d: Name: %s\n',s,Preprocessed_2s(s).subject) 
    s1 = Preprocessed_2s(s).data;
    s1 = double(s1);
    F2 = [];
    for j = 1:size(s1,3)
        F1 = [];
        %fprintf('[Info]: epoch %d/%d\n',j,size(s1,3));
        for i = 1:size(s1,1)
            x = s1(i,:,j);
            f = nback_features(x);
            F1 = [F1;f];
        end
        F1 = F1';
        F1ravel = F1(:);
        F2 = [F2;F1ravel'];
    end
%NOTE:      
%     F2= (F2-min(F2)) ./ (max(F2)-min(F2)); %This is throwing error for me
%     in 2016a
% Modified: 
    num = F2-min(min(F2));
    den = max(max(F2))-min(min(F2));
    F2 = num./den;
    S = [S;F2];
    disp('#########################')
end
S=S(idx,:);