With the help of Eyetracker (Eyelink in this case), we can get the realtime gaze position and therefore we can monitor the gaze position to make sure it falls into a certain area before the participants start the trial.

% make sure participants are fixating on the central
while isEyeLink  % if Eyelink was connected
status = Eyelink('newfloatsampleavailable');  % check to see if everything is fine in Eyelink
if status /=0; % if all is fine
evt = Eyelink('newestfloatsample'); % get the newest float sample from Eyelink
realtime.x = evt.gx(1);  % get the x axis of gaze
realtime.y = evt.gy(1);  % get the y axis of gaze
end;

if abs(realtime.x-512)\<=50&abs(realtime.y-384)\<=50  % if the gaze falls into  a 50\*50 pixel square around the cross
break; % jump out the loop and start the trial
end
end

1 load raw data

2 channel location

3 re-reference

4 filtering (sometimes the data is very huge and therefore the filtering will be very slow, so we can do the filtering later)

5 Create EEG event list (ERPLAB from now on)

6 Assign bins (how to operate on bins? see here)

7 extract bin-based epoch

8 artifact rejection

-> choose channel around the eye. -> update and then reject

9 compute average erps

10 filter if you don’t do it in step 4

11 bin operations if necessary 

12 Plot ERPs

13 Save figures if you want

Remember, these are basic steps to illustrate how to process the EEG data using EEGLAB and ERPLAB toolbox, you’re suggested not to strictly follow these steps.

Code:

%% merge behavioral data into EEG

% Sometimes, it's hard to lively send behavioral data to EEG system, i.e.

% oral response. Here is a code example to merge behavioral result into EEG

% data.

% Author: Jian Chen

% saturn.jian.chen@gmail.com

% 13/Dec/2016

%+++++++++++++%

%% load EEG data first, finish 'channel location', 're-reference'

% get event data via 'event = EEG.event'

event = EEG.event;

event\_back = event;

%% load behavioral data

fid = fopen('YourResponse\_acc.txt');  % here is a vector with length of 540

acc = textscan(fid,'%d');

fclose(fid);

%% set the latency between target code and inserted code

time = 100;

%% creat response event

for i = 2 : 2 : length(event)

Nevent(i/2).bepoch    = event(i).bepoch;

Nevent(i/2).bini      = event(i).bini;

Nevent(i/2).binlabel  = event(i).binlabel;

Nevent(i/2).codelabel = event(i).codelabel;

Nevent(i/2).duration  = event(i).duration;

Nevent(i/2).enable    = event(i).enable;

Nevent(i/2).flag      = event(i).flag;

Nevent(i/2).item      = event(i).item + 1;

Nevent(i/2).latency   = event(i).latency + time;

Nevent(i/2).type      = acc{1}(i/2);

end

%% add new events to original event list

for i = 1:length(Nevent)

event(2*i+i+1:end+1) = event(2*i+i:end); % move rows down

event(2\*i+i) = Nevent(i); % insert response mark

end

%% save event in case of overwrite

save event\_merge.mat event

%% if the EEG data was currently loaded, simply replace the EEG.event with event, then operate as normal

EEG.event = event;

%replace ERPLAB event if you are using ERPLAB plugin

ERP.EVENTLIST.eventinfo = event;

We often use errorbar function to plot errorbar but it’s impossible for Matlab to put error bars on a grouped bar graph. Matlab simply gives a shit if you command it straightforwardly. For example:

The error bars are located in the center of each grouped bars, rather than the individual bar in each grouped bars.

Here is the note I made, the main code is from the link below:

Code example:

Data set:

mm: 4 by 6 matrix; ci: 4 by 6 matrix



mm =

[0.2081 0.3440 0.3470 0.7018 0.8422 0.9874
0.2628 0.3043 0.4032 0.7287 0.8462 1.1055
0.2928 0.3139 0.4176 0.7241 0.8319 1.0483
0.2287 0.2538 0.3665 0.4895 0.8999 1.1309]

ci =

[0.0176 0.0473 0.0443 0.0724 0.0855 0.0931
0.0312 0.0201 0.0407 0.0621 0.0604 0.0874
0.0291 0.0330 0.0403 0.0756 0.0732 0.0847
0.0228 0.0150 0.0264 0.0598 0.0922 0.0861]

figure;
h = bar(mm');
set(h,'BarWidth',0.8);    % The bars will now touch each other
set(gca,'fontsize', 18);
set(get(gca,'YLabel'),'String','RT in Seconds')
lh = legend('Valid Left', 'Invalid left', 'Valid right', 'Invalid right');
set(lh,'Location','BestOutside','Orientation','horizontal')
hold on;

% Aligning errorbar to individual bar within groups
% Based on barweb.m by Bolu Ajiboye from MATLAB File Exchange
groupwidth = min(0.8, 4/(4+1.5));
for i = 1:4
x = (1:6) - groupwidth/2 + (2*i-1) * groupwidth / (2*4);
errorbar(x,mm(i,:),ci(i,:),'k', 'linestyle', 'none');
end

Here is the solution:

Select Preferences -> Browse Packages… to open your Packages folder, and open the Objective-C subfolder. Open Objective-C.tmLanguage in Sublime as an XML file (open it in sublime), and comment out (or remove entirely) Line 7:

<string>m</string>

Save the file and close it. Next, open Objective-C++.tmLanguage and do the same thing to Line 8:

<string>M</string>

Save and close that file as well.

Finally, just to make sure your Open all with current extension as… command worked, go to Packages/User and check for the existence of a file called MATLAB.sublime-settings. Open it as a JSON file, and double-check that its contents are as follows:

{
     "extensions";
     [
         "m"
     ]
}

If for some reason the file doesn’t exist, create it with the above contents and save it.

Preference – User – above contents – save it as ‘MATLAB.sublime-settings

You should now be able to open any .m file and have the syntax automatically set to MATLAB.

(This solution mainly come from http://stackoverflow.com/questions/24016186/setting-default-syntax-for-sublime-doesnt-work)

The way to implement this method is easy, just to insert a line of code in your matlab script. To even simplify the triggering job, I wrote a function here to share. So you basically only need to input the triggers you want to send into the trigger.m function. For example, trigger(1); trigger(2).

Download all three files at here:

pp_ppMexDownload

trigger.m

function trigger(sti_trigger)

% [input] = pp(pins,[output],[slowChecks],[port],[address])
% pp(uint8([2 3 4 5 6 7 8 9]), [0 0 0 0 0 0 0 0], 0, uint8(0), uint64(53248));

if sti_trigger == 0
    bin_value = [0 0 0 0 0 0 0 0];
    pp(uint8([2 3 4 5 6 7 8 9]), bin_value, 0, uint8(0), uint64(53248));
end

if sti_trigger == 1
    bin_value = [1 0 0 0 0 0 0 0];
    pp(uint8([2 3 4 5 6 7 8 9]), bin_value, 0, uint8(0), uint64(53248));
end

if (1 < sti_trigger)&&(sti_trigger < 4)
    bin_value = strcat('000000',dec2bin(sti_trigger));
    bin_value = fliplr(bin_value);
    bin_value = bin_value - '0' ;
    pp(uint8([2 3 4 5 6 7 8 9]), bin_value, 0, uint8(0), uint64(53248));
end

if (3 < sti_trigger)&&(sti_trigger < 8)
    bin_value = strcat('00000',dec2bin(sti_trigger));
    bin_value = fliplr(bin_value);
    bin_value = bin_value - '0' ;
    pp(uint8([2 3 4 5 6 7 8 9]), bin_value, 0, uint8(0), uint64(53248));
end

if (7 < sti_trigger)&&(sti_trigger < 16)
    bin_value = strcat('0000',dec2bin(sti_trigger));
    bin_value = fliplr(bin_value);
    bin_value = bin_value - '0' ;
    pp(uint8([2 3 4 5 6 7 8 9]), bin_value, 0, uint8(0), uint64(53248));
end

if (15 < sti_trigger)&&(sti_trigger < 32)
    bin_value = strcat('000',dec2bin(sti_trigger));
    bin_value = fliplr(bin_value);
    bin_value = bin_value - '0' ;
    pp(uint8([2 3 4 5 6 7 8 9]), bin_value, 0, uint8(0), uint64(53248));
end

if (31 < sti_trigger)&&(sti_trigger < 64)
    bin_value = strcat('00',dec2bin(sti_trigger));
    bin_value = fliplr(bin_value);
    bin_value = bin_value - '0' ;
    pp(uint8([2 3 4 5 6 7 8 9]), bin_value, 0, uint8(0), uint64(53248));
end

if (63 < sti_trigger)&&(sti_trigger < 128)
    bin_value = strcat('0',dec2bin(sti_trigger));
    bin_value = fliplr(bin_value);
    bin_value = bin_value - '0' ;
    pp(uint8([2 3 4 5 6 7 8 9]), bin_value, 0, uint8(0), uint64(53248));
end

if sti_trigger > 127
    bin_value = dec2bin(sti_trigger);
    bin_value = fliplr(bin_value);
    bin_value = bin_value - '0' ;
    pp(uint8([2 3 4 5 6 7 8 9]), bin_value, 0, uint8(0), uint64(53248));
end

[update 2/12/2016]

An minor error was found in trigger.m and has been fixed.

Here is the source code:

Pupil_faceDownload

How to use:

Just feed face images into the source_face folder (1 arrow) and then open and run the ‘pupil_face.m’ file (3 arrow) in the current directory. Remember to connect to the internet as I use some online face analysis tools. After computation, just go to the output_face folder to pick up processed face images (3 arrow).

Run the pupil_face.m file, then it will automatically recognise the pupil position and resize the face to a grayscale pupil-distance constant face. Remember to connect to the Internet cause I use some online source.

Results showing as below:

Download and try it!