Need to create a 3 dimensional matrix of subjects, block numbers and 6 sub condition values within each block.

Hi, Until now I have worked only with 2D matrices which resemble spreadsheets. Now I have data with one more dimension. I have 19 subjects who each did 8 blocks of an experiment. In each block they got different numbers of task 1 to 6 correct.

Here is the script so far:

clear 
close all

subno={'s_01','s_03',  's_05', 's_06', 's_09', 's_11',  's_13',  's_16', 's_18', 's_20',  's_22','s_23',  's_26','s_28',  's_31',  's_33',  's_34','s_35', 's_37',};
blockno={'_01', '_02','_03', '_04', '_05', '_06','_07', '_08'};

for k=1:length(subno)

    %defining paths
MAIN = '/data/janani/EoA/logfiles';
PATHIN=[MAIN, '/',subno{k}, '/'];
PATHOUT=[MAIN, '/BEHanalysis/'];

for j=1:length (blockno)
textfile= ['jad1_',subno{k}, blockno{j}, '.txt']; 
cd (PATHIN);
list1=dir('*.txt');

fname=[PATHIN, textfile];
fid=fopen(fname);

    for i=1:72
    [M1(i,:)]=fscanf(fid,'%i', [1 6]);
    end
%We write the following loops to extract the correct responses into corr_H0, 1 and 2
%the below loop checks if there is 0, 1 and 2 in both the nuhBIG Hs (2) and resp column (4)

   count0=1;
   for p = 1:size (M1, 1)
      if (M1 (p, 2)==0 && M1 (p, 4)==0)
         corr_H0(count0)=p
         count0=count0+1;
      end
   end

   count1=1;
   for q = 1:size (M1, 1)
      if (M1 (q, 2)==1 && M1 (q, 4)==1)
         corr_H1(count1)=q
         count1=count1+1;
      end
   end

   count2=1;
   for r = 1:size (M1, 1)
      if (M1 (r, 2)==2 && M1 (r, 4)==2)
         corr_H2(count2)=r
         count2=count2+1;
      end
   end
    %the below loop checks if there is 0, 1 and 2 in both the nuhSmall Hs (3) and resp column (4)

   count0s=1;
   for p = 1:size (M1, 1)
      if (M1 (p, 3)==0 && M1 (p, 4)==0)
         corr_H0(count0s)=p
         count0s=count0s+1;
      end
   end

   count1s=1;
   for q = 1:size (M1, 1)
       if (M1 (q, 3)==1 && M1 (q, 4)==1)
          corr_H1(count1s)=q
          count1s=count1s+1;
       end
    end

   count2s=1;
   for r = 1:size (M1, 1)
      if (M1 (r, 3)==2 && M1 (r, 4)==2)
         corr_H2(count2s)=r
         count2s=count2s+1;
      end
   end

c= [count0, count1, count2, count0s, count1s, count2s]; %is the array of the 6 values that need to join the j and k values.
%I tried this:
SBC=(k, j, c);
%but it doesn´t give me what I want.

end      
end

Just before the end of the loop I would need to input count0, count1, count2, count0s, count1s, and count2s into the matrix which also indicates k (the subject number) and j (which indicates block number) into the matrix before their values are overwritten.

Any help would be appreciated. Thanks! Jan

Answer 1

[Jarrod Rivituso]

Is there a reason you want to store things in a 3-dimensional array? Seems you might be better off, at least at first, keeping things in a structure. For clarity, that would mean replacing the line

subject(k).block(j).count0 = count0;
subject(k).block(j).count1 = count1;
subject(k).block(j).count2 = count2;
subject(k).block(j).count0s = count0s;
subject(k).block(j).count1s = count1s;
subject(k).block(j).count2s = count2s;

To me, this is a bit clearer. The 3-D array approach would be something like

sbc(k,j,1:6) = c;

If you go that route, don't forget to preallocate!

Best of luck!

[Jan]

The reason was to be able to export it into a spreadsheet and put it through statistical analysis for which I use dlmwrite because I am working with Linux and despite changes, xlswrite doesn´t work here. Do you happen to know how to put a path to the dlmwrite?

Thanks for your help!