I have Fixation details in 2D, X-Y Coordinates in pixels. Distance between participants and screen.. Saacade duration for each saccade..
So how I can calculate saccade Amplitude in degree for each saccade.. Which software, formula, algorithms need to saccade Amplitude
If you have the distance to the screen and the dimensions of the screen, to calculate the saccade amplitude is simple using the trigonometry sin, cos, tg functions. I am surprised that tobii does not export it... but see this link for help: https://www.sr-research.com/visual-angle-calculator/
João Miguel Castelhano thanks for your response!
I have fixation point X-Y coordinates in [MCS norm] and [DACS px] form. what is required to calculate euclidean distance between 2-fixations?
Deepike hi,
I just have a remark for future recordings. If you would need an entire set of eye-tracking metrics (including the amplitude) calculated and visualized automatically, then our software, EventIDE (www.okazolab.com), can do it in real-time. The software supports all Tobii trackers and exports results in any desired format.
I have a custom Matlab script I made for exactly this purpose with the Tobii output. Let me know if you’re interested. I calculate amplitude using Euclidean distance. I calculated in pixels but you can convert that to degrees.
Shannon Ross-Sheehy Would it be possible to share the Matlab script you have made for calculating saccade amplitude with Tobii output? It would be very helpful as I am trying to calculate the same metric.
The calculation itself is not terribly difficult, the hard part is getting the tobii data imported in a way that allows you to identify trial starts and stops (or other interest periods). Here is a chunk of my matlab code that assumes you've exported out of tobii using the IVT, imported into MatLAB and have created two arrays with trial starts and stops (mine are called StartTrialRow and EndTrialRow below, I used a Cedrus stimtracker to create stimmarks, but if you used Tobii studio, it includes even marks too). This produces a single mean euclidian distance for each trial (i) but you could tune that how ever you like. I can send you my full code, but it's really specific to my setup (however it sometimes helps to dig through someone else's code). Please let me know if you find any bugs! :-)
Copy and paste below into Matlab:
% Calculate eculidian dist for consecutive points (pdist and pdist2 find all
% pairwise distances which is NOT what we want here -srs
% Calculate eculidian dist for consecutive points (pdist and pdist2 find all
% pairwise distances which is NOT what we want here -srs, NOTE that Tobii by
% default uses pixels so you must convert to angle using a formula that fits
% your setup. Here mine was the standard TX300 high def monitor
% (surface of 45.5° (w) by 26.76° (h) at a distance of 65 cm but I used
% monitor specs for vis angle calc below so more trustworthy).
x = [FixationPointXMCSpx(StartTrialRow(i):EndTrialRow(i))]; x = x(~isnan(x));
y = [FixationPointYMCSpx(StartTrialRow(i):EndTrialRow(i))]; y = y(~isnan(y));
edist = sqrt( sum( abs( diff( [x,y] ) ).^2, 2 ) );
edist([edist==0])=NaN;
% Convert the saccade distance score (currently in pixels) to degrees
% NOTE, each pixel is .024 degrees for our Tobii TX300 monitor
edist = edist * .024;
SaccDist(i) = mean(edist,'omitnan');
Hope this helps!
- Badges
- Science topic
- Similar topics
- Correction