P435 Fall 2015
How to develop an eye tracking research question
The focus of eye tracking experiments comes down to two questions:
1) What information is available to do the task?
2) What information do people use?
We will be comparing two types of tasks that use the same stimuli, or two sets of stimuli, or possibly observers from
different groups on the same task and stimuli. Basically you want to have two groups or conditions.
You should be able to identify 3-7 regions of interest on all of your stimuli. Each stimulus must have at least a subset
of the regions on it. For example, if the stimuli were faces, you might identify the eyes, nose, mouth and external
features as four different regions of interest.
We will be tracking the number of eye fixations that fall in the different regions when subjects look at different stimuli
or do different tasks. For example, will subjects look more at the mouth when doing an emotion recognition task than when
doing a famous/non-famous judgment?
You can compare across different types of stimuli (e.g. same- vs other-race faces) or comparing a male vs female
participant on one set of stimuli.
First, come up with a general research topic and a design for the stimuli and comparisons.
Note that your images will need to be static, or possibly animated gifs as long as there isn’t too much motion in the
image. You’ll need to have regions of interest for each trial that indicate where each region is for each image.
Next, pick a set of features that you plan to use and mark them using a masking procedure described later.
Finally, run your design past Tom. Your design should tell us something important about human behavior, and be
generalizable to a variety of settings. You are not limited to faces, but you should avoid more trivial questions that are
specific to one set of stimuli.
Make sure to think about the consequences of the Faces in Autism paper, and design your stimuli appropriately. This might
mean having extreme closeups of faces to encourage the use of foveal mechanisms, or designs that rely on relatively small
You will then collect data from two participants in class and analyze the data using the ExpertEyes software and custom
matlab routines provided by Tom.
Currently, i am taking Psychology intensive writing class. This time, our broad topic was about Eyetracking. And 5 groups
were designed to make each experiment and see that how our hypothesis applied to the experiment. Our group chose to use
Spot the differences for the experiment. We chose 6 to 7 images for our experiment and created Powerpoints with including
Calibration Powerpoint in the beginning. Then we asked our participants to wear some kind of equipment that our professor
made to see their eye movements. Our hypothesis is that given a scene to search differences on two similar pictures, a
proportionate amount of time should be spent searching four types of areas relative to each areas density. Later on, we
used a program to make masks of their eye movements and see where their eye moved the most and how and where they moved
So i would like to ask you to write introduction by using our hypothesis and 3 references to support our hypothesis. There
is an article that we read in the beginning of the class which might be helpful to understand what this essay is about.
(It is called Hayhoe.pdf). Please do not use this article as a reference.
* The calibration that i mentioned can be seen below powerpoint (it is called Sample Calibration Powerpiont Widescreen.pptx).
This was to use before we start our experiment.
* Also, i uploaded project grading so it can help you to figure out what to write about (it is called
* Last, “Howtodevelopaneyetrackingresearchquestion” document will help you to understand how our experiment started