Matlab for neuroscientists Cap 3 project

Question

Gonzalo Lerner 2016-3-19

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/274448-matlab-for-neuroscientists-cap-3-project

评论： Gonzalo Lerner 2016-3-20

Hi there, i'm new to matlab and don't have any experience with programming.

I'm reading Matlab for neuroscientists and on chapter number 3 "Visual pop out", they ask you to replicate Treisman's experiment.

I couldn't do this by myself so i found a script on the internet that works and i'm trying to understand it.

I have a problem with one of the lines, it plots something but the way it's written doesn't sound familiar to me at all. "Help on selection" on matlab didn't help either.

I pasted the entire script allthough the line is 23

The line which i'm talking about is this:

g=Treisman(n, 'f', target);

%main script
start=figure;
text(0.2, 0.7, 'Press any key to begin the 1st block');
text(0.05, 0.5, 'For the trials press ''t'' if there''s a target and ''n'' if there isn''t a target');
pause;
close;
right = 0;
wrong = 0;
trial_block=30;
%1st block pop-out with 4 objects; we do 30 trials;
n=4;
t_target4pop=[];
t_no_target4pop=[];
z=randperm(trial_block);
for i=1:trial_block
  if(z(i) <=(trial_block/2))
      target='y';
  else
      target='n';
  end;
g=Treisman(n, 'f', target);
tic;
pause;
temp_t=toc;
%t=[t, temp_t];
temp_ans=get(g, 'CurrentCharacter');
%ans=[ans, temp_ans];
close;
  if((strcmp(target, 'y') && strcmp(temp_ans, 't'))) 
  t_target4pop=[t_target4pop, temp_t];
  right = right + 1;
  elseif((strcmp(target, 'n') && strcmp(temp_ans, 'n')))
  t_no_target4pop=[t_no_target4pop, temp_t];
  right = right + 1;
  else
  wrong = wrong + 1;
  end;
end;
t_targ4pop_mean=mean(t_target4pop);
t_no_targ4pop_mean=mean(t_no_target4pop);
%Next block
next=figure;
text(0.2, 0.7, 'Press any key to begin the next block');
text(0.05, 0.5, 'For the trials press ''t'' if there''s a target and ''n'' if there isn''t a target');
pause;
close;
right8 = 0;
wrong8 = 0;
%trial_block=30;
%2nd block pop-out with 8 objects; we do 30 trials;
n=8;
t_target8pop=[];
t_no_target8pop=[];
z=randperm(trial_block);
for i=1:trial_block
  if(z(i) <=(trial_block/2))
      target='y';
  else
      target='n';
  end;
g=Treisman(n, 'f', target);
tic;
pause;
temp_t=toc;
%t=[t, temp_t];
temp_ans=get(g, 'CurrentCharacter');
%ans=[ans, temp_ans];
close;
  if((strcmp(target, 'y') && strcmp(temp_ans, 't'))) 
  t_target8pop=[t_target8pop, temp_t];
  right8 = right8 + 1;
  elseif((strcmp(target, 'n') && strcmp(temp_ans, 'n')))
  t_no_target8pop=[t_no_target8pop, temp_t];
  right8 = right8 + 1;
  else
  wrong8 = wrong8 + 1;
  end;
end;
t_targ8pop_mean=mean(t_target8pop);
t_no_targ8pop_mean=mean(t_no_target8pop);
%Next block
next=figure;
text(0.2, 0.7, 'Press any key to begin the next block');
text(0.05, 0.5, 'For the trials press ''t'' if there''s a target and ''n'' if there isn''t a target');
pause;
close;
right12 = 0;
wrong12 = 0;
%trial_block=30;
%3rd block pop-out with 12 objects; we do 30 trials;
n=12;
t_target12pop=[];
t_no_target12pop=[];
z=randperm(trial_block);
for i=1:trial_block
  if(z(i) <=(trial_block/2))
      target='y';
  else
      target='n';
  end;
g=Treisman(n, 'f', target);
tic;
pause;
temp_t=toc;
%t=[t, temp_t];
temp_ans=get(g, 'CurrentCharacter');
%ans=[ans, temp_ans];
close;
  if((strcmp(target, 'y') && strcmp(temp_ans, 't'))) 
  t_target12pop=[t_target12pop, temp_t];
  right12 = right12 + 1;
  elseif((strcmp(target, 'n') && strcmp(temp_ans, 'n')))
  t_no_target12pop=[t_no_target12pop, temp_t];
  right12 = right12 + 1;
  else
  wrong12 = wrong12 + 1;
  end;
end;
t_targ12pop_mean=mean(t_target12pop);
t_no_targ12pop_mean=mean(t_no_target12pop);
%Next block
next=figure;
text(0.2, 0.7, 'Press any key to begin the next block');
text(0.05, 0.5, 'For the trials press ''t'' if there''s a target and ''n'' if there isn''t a target');
pause;
close;
right16 = 0;
wrong16 = 0;
%trial_block=30;
%4th block pop-out with 16 objects; we do 30 trials;
n=16;
t_target16pop=[];
t_no_target16pop=[];
z=randperm(trial_block);
for i=1:trial_block
  if(z(i) <=(trial_block/2))
      target='y';
  else
      target='n';
  end;
g=Treisman(n, 'f', target);
tic;
pause;
temp_t=toc;
%t=[t, temp_t];
temp_ans=get(g, 'CurrentCharacter');
%ans=[ans, temp_ans];
close;
  if((strcmp(target, 'y') && strcmp(temp_ans, 't'))) 
  t_target16pop=[t_target16pop, temp_t];
  right16 = right16 + 1;
  elseif((strcmp(target, 'n') && strcmp(temp_ans, 'n')))
  t_no_target16pop=[t_no_target16pop, temp_t];
  right16 = right16 + 1;
  else
  wrong16 = wrong16 + 1;
  end;
end;
t_targ16pop_mean=mean(t_target16pop);
t_no_targ16pop_mean=mean(t_no_target16pop);
%Now we begin the conjunction search.
next=figure;
text(0.2, 0.7, 'Press any key to begin the conjunction blocks');
text(0.05, 0.5, 'For the trials press ''t'' if there''s a target and ''n'' if there isn''t a target');
pause;
close;
rightc = 0;
wrongc = 0;
trial_block=30;
%1st block conjunction with 4 objects; we do 30 trials;
n=4;
t_target4con=[];
t_no_target4con=[];
z=randperm(trial_block);
for i=1:trial_block
  if(z(i) <=(trial_block/2))
      target='y';
  else
      target='n';
  end;
g=Treisman(n, 'c', target);
tic;
pause;
temp_t=toc;
%t=[t, temp_t];
temp_ans=get(g, 'CurrentCharacter');
%ans=[ans, temp_ans];
close;
  if((strcmp(target, 'y') && strcmp(temp_ans, 't'))) 
  t_target4con=[t_target4con, temp_t];
  rightc = rightc + 1;
  elseif((strcmp(target, 'n') && strcmp(temp_ans, 'n')))
  t_no_target4con=[t_no_target4con, temp_t];
  rightc = rightc + 1;
  else
  wrongc = wrongc + 1;
  end;
end;
t_targ4con_mean=mean(t_target4con);
t_no_targ4con_mean=mean(t_no_target4con);
next=figure;
text(0.2, 0.7, 'Press any key to begin the next block');
text(0.05, 0.5, 'For the trials press ''t'' if there''s a target and ''n'' if there isn''t a target');
pause;
close;
rightc8 = 0;
wrongc8 = 0;
trial_block=30;
%2nd block conjunction with 8 objects; we do 30 trials;
n=8;
t_target8con=[];
t_no_target8con=[];
z=randperm(trial_block);
for i=1:trial_block
  if(z(i) <=(trial_block/2))
      target='y';
  else
      target='n';
  end;
g=Treisman(n, 'c', target);
tic;
pause;
temp_t=toc;
%t=[t, temp_t];
temp_ans=get(g, 'CurrentCharacter');
%ans=[ans, temp_ans];
close;
  if((strcmp(target, 'y') && strcmp(temp_ans, 't'))) 
  t_target8con=[t_target8con, temp_t];
  rightc8 = rightc8 + 1;
  elseif((strcmp(target, 'n') && strcmp(temp_ans, 'n')))
  t_no_target8con=[t_no_target8con, temp_t];
  rightc8 = rightc8 + 1;
  else
  wrongc8 = wrongc8 + 1;
  end;
end;
t_targ8con_mean=mean(t_target8con);
t_no_targ8con_mean=mean(t_no_target8con);
next=figure;
text(0.2, 0.7, 'Press any key to begin the next block');
text(0.05, 0.5, 'For the trials press ''t'' if there''s a target and ''n'' if there isn''t a target');
pause;
close;
rightc12 = 0;
wrongc12 = 0;
trial_block=30;
%3rd block conjunction with 12 objects; we do 30 trials;
n=12;
t_target12con=[];
t_no_target12con=[];
z=randperm(trial_block);
for i=1:trial_block
  if(z(i) <=(trial_block/2))
      target='y';
  else
      target='n';
  end;
g=Treisman(n, 'c', target);
tic;
pause;
temp_t=toc;
%t=[t, temp_t];
temp_ans=get(g, 'CurrentCharacter');
%ans=[ans, temp_ans];
close;
  if((strcmp(target, 'y') && strcmp(temp_ans, 't'))) 
  t_target12con=[t_target12con, temp_t];
  rightc12 = rightc12 + 1;
  elseif((strcmp(target, 'n') && strcmp(temp_ans, 'n')))
  t_no_target12con=[t_no_target12con, temp_t];
  rightc12 = rightc12 + 1;
  else
  wrongc12 = wrongc12 + 1;
  end;
end;
t_targ12con_mean=mean(t_target12con);
t_no_targ12con_mean=mean(t_no_target12con);
next=figure;
text(0.2, 0.7, 'Press any key to begin the last block');
text(0.05, 0.5, 'For the trials press ''t'' if there''s a target and ''n'' if there isn''t a target');
pause;
close;
rightc16 = 0;
wrongc16 = 0;
trial_block=30;
%4th block conjunction with 16 objects; we do 30 trials;
n=16;
t_target16con=[];
t_no_target16con=[];
z=randperm(trial_block);
for i=1:trial_block
  if(z(i) <=(trial_block/2))
      target='y';
  else
      target='n';
  end;
g=Treisman(n, 'c', target);
tic;
pause;
temp_t=toc;
%t=[t, temp_t];
temp_ans=get(g, 'CurrentCharacter');
%ans=[ans, temp_ans];
close;
  if((strcmp(target, 'y') && strcmp(temp_ans, 't'))) 
  t_target16con=[t_target16con, temp_t];
  rightc16 = rightc16 + 1;
  elseif((strcmp(target, 'n') && strcmp(temp_ans, 'n')))
  t_no_target16con=[t_no_target16con, temp_t];
  rightc16 = rightc16 + 1;
  else
  wrongc16 = wrongc16 + 1;
  end;
end;
t_targ16con_mean=mean(t_target16con);
t_no_targ16con_mean=mean(t_no_target16con);
n=[4, 8, 12, 16]';
t_target_pop=[t_targ4pop_mean, t_targ8pop_mean, t_targ12pop_mean, t_targ16pop_mean]';
t_target_con=[t_targ4con_mean, t_targ8con_mean, t_targ12con_mean, t_targ16con_mean]';
t_no_target_pop=[t_no_targ4pop_mean, t_no_targ8pop_mean, t_no_targ12pop_mean, t_no_targ16pop_mean]';
t_no_target_con=[t_no_targ4con_mean, t_no_targ8con_mean, t_no_targ12con_mean, t_no_targ16con_mean]';
[M1, p1]=corrcoef(n, t_target_pop);
M1=M1(1,2);
p1=p1(1,2);
[M2, p2]=corrcoef(n, t_target_con);
M2=M2(1,2);
p2=p2(1,2);
[M3, p3]=corrcoef(n, t_no_target_pop);
M3=M3(1,2);
p3=p3(1,2);
[M4, p4]=corrcoef(n, t_no_target_con);
M4=M4(1,2);
p4=p4(1,2);
sprintf('The correlation coefficient is %0.5g and the p-value is %0.5g for the popout trial when the target was present', M1, p1)
sprintf('The correlation coefficient is %0.5g and the p-value is %0.5g for the conjunction trial when the target was present', M2, p2)
sprintf('The correlation coefficient is %0.5g and the p-value is %0.5g for the popout trial when the target was absent', M3, p3)
sprintf('The correlation coefficient is %0.5g and the p-value is %0.5g for the conjunction trial when the target was absent', M4, p4)
comment='p-values < 0.05 are considered significant'
figure;
plot(n, t_target_pop);
hold on;
plot(n, t_target_con, 'r');
title('Correctly Identified the Presence of Target');
legend('Pop Out', 'Conjunction');
xlabel('Set Size');
ylabel('Mean Reaction Time (sec)');
hold off;
figure;
plot(n, t_no_target_pop);
hold on;
plot(n, t_no_target_con, 'r');
title('Correctly Identified the Absence of Target');
legend('Pop Out', 'Conjunction');
xlabel('Set Size');
ylabel('Mean Reaction Time (sec)');
hold off;
%These are the results for the sprintf function for the trial I ran on
%myself. The graphs are given as well.
%The correlation coefficient is -0.57912 and the p-value is 0.42088 for the popout trial when the target was present
%The correlation coefficient is 0.99994 and the p-value is 6.4571e-005 for the conjunction trial when the target was present
%The correlation coefficient is -0.032356 and the p-value is 0.96764 for the popout trial when the target was absent
%The correlation coefficient is 0.92974 and the p-value is 0.070256 for the conjunction trial when the target was absent
%p-values < 0.05 are considered significant
%For a qualitative assessment of the slopes of the graphs I would say that
%the popout task does not seem to require more mental processing time as 
%the set size increases (ie the slope is relatively flat.). However, 
%for the conjunction task there is statistically significant linear
%increase in reaction time to complete the task.
%I realize that one p-value for the conjunction task is not quite below
%0.05, so the statement made above is not exactly true from this small sample size.
%Final note in the trials ran I always scored more than 20 correct/block,
%so the mean reaction time is based on at least 20 data points as required.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Steven Lord 2016-3-20

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/274448-matlab-for-neuroscientists-cap-3-project#answer_214271

As far as I'm aware that's not the name of a function in any MathWorks toolbox, nor is it a file on the File Exchange. A search of the whole MathWorks site found only this Answers message. The code you posted only appears to have uses of Treisman, not the definition of that function, so it's going to be hard/impossible to help you determine why it's not doing what you expect.

4 个评论
显示 2更早的评论隐藏 2更早的评论

Walter Roberson 2016-3-20

We have no idea what the purpose of any of this code is. We would have to read the chapter to begin to find out. The purpose of the code is not documented in the code. We would have to download the code and run it to even find out what it does in practice, since it doesn't say.

I would speculate it might have something to do with Feature Integration Theory developed by Anne Treisman and Garry Gelade. But you, having read the chapter, would have a much better guess about what the code is for then we would.

Gonzalo Lerner 2016-3-20

Ohh, i get it. Sorry for the trouble, since i'm new i didn't understand that it was a function that i actually had in the same folder, so that's where it was picking it up from.

Now that i read the function i did understand what it is for.

Thank you!

请先登录，再进行评论。