Project Plan Eye-Tracking Course
1 Alena Holubcová
PURSUING THE SMOOTH PURSUIT
Research questions:
What is the slowest motion of a target that will induce smooth pursuit in a healthy population (not
for example Schizophrenics or people with Parkinson’s disease)? [At what lowest speed of the target
does the smooth pursuit break into fixations and saccades?]
What is the fastest motion of a target that eyes can follow with smooth pursuit before they perform
catch-up saccades? [At what speed do people start with catch-up saccades to keep track of the
target?]
At what speed of a target does the smooth pursuit achieve best accuracy? [At what speed of the
target can the eyes best anticipate the movement = they are moving at exactly the same location (or
the closest possible=least distance and latency possible) as the target? Is this speed different for
different paths of movement? Is the distribution of speeds following a Gaussian curve for a given
path, and is accuracy the best for the “middle speed” and poorest for the fastest speed? Why would
the accuracy be the best at that specific speed  Off topic question: Is the “middle speed” coinciding
with the speed of some objects we are used to in real life, e.g. the average walking speed of human
beings? Or cars in the cities? “Predators” speed? Is our smooth pursuit’s best accuracy adaptive to the
environment? Is accuracy dependent on age?]
Are there always the same latencies for starting to pursue the target? [Does the latency differ for
different speed of the target? Is this difference statistically significant? And if I would change the
position of the target unexpectedly during the experiment, how long would people take before
making a catch-up saccade?]
Can accuracy of smooth pursuit be enhanced by motivation? [Would there be any difference in
performance? Will motivation for following the target lead to sufficient increase in the speed of the
eyes to match the speed of the target as best as possible when people would be rewarded?]
Overview of literature:
Here I will describe what is known about smooth pursuit and the methods how to measure it.
[Literature overview found until now listed at the end.]
Stimuli and task:
There will be two groups of participants, control group and experiment group.
Experiment setup: a 9point calibration, validation, trials (number to be decided according to different
speed of the target – need to set a range of speed, for each “speed” there could be 3 trials in the
whole experiment; possibly changing the path and/or direction of the movement of the target),
calibration (to recognize any shifts in data)
Project Plan Eye-Tracking Course
2 Alena Holubcová
Tasks for control group:
Target (a black dot) will be moving horizontally, starting at different heights, towards the other side
of the monitor, at different speeds (random order, predefined ranges of speed based on what is
known about smooth pursuit + slower and faster on both edges, to make the margins broader). The
path factor not yet decided.
Tasks for experiment group:
Same target, type and speed of movement and number of trials, the only difference is that when the
eyes will get close enough to the target, the target will change the colour to green. [“How many
targets as an overall score can you catch before they disappear on the other side of the monitor?”
This boundary needs to be backed up by relevant theory. Would people stop being interested in
pursuing the target after they caught it? Would adding the length of the path for which they could
follow the dot in “green” as a part of the score solve the interest question? Would this vary as a
function of the speed?]
Eye-tracker:
SR Eyelink – fast changes of display will be required; online detection of velocities – applying the filter
could change the target colour from black to green in the experiment group as a feedback of “good
pursuing”
Analysis:
There is a mathematical model of smooth pursuit developed by Jansson (2014, Mathematical
Modeling of the Human Smooth Pursuit System). The results of the experiment can be compared with
the model to see how accurately the model can predict eye behaviour in reality.
Thoughts to consider:
[Do we do smooth pursuit in vertical direction? What would the distinction be from blinks?]
[Interaction with the task/measure of performance: if the dot would change to green from black
when the eye speed would match it for e.g.90%; would it become a motivation to “catch it”? Would
people stop trying to get on 100% if the target would turn to green already at 90%?]
[Predictability of path: what types of paths are predictable for people? What types of path can be
mentally tracked – e.g.” hiding” the moving target under a white non-transparent rectangle, and
checking if people can follow the moving target mentally, without breaking the smooth pursuit to
fixations and saccades.]
[Should I let people have practise trials?]
Project Plan Eye-Tracking Course
3 Alena Holubcová
[Should I put attention tasks? E.g. two targets moving at different speeds, following only the black
dot and not the grey dot=distractor]
[Is the latency changing as a function of the motivation, how can I tell the difference in the “basic”
motivation to follow the target as a task of the experiment, and the “enhanced motivation” by
changing the colour of the target and counting the score?]
Implications:
Knowing the minimum and maximum possible speed of a target inducing smooth pursuit could be
implemented into smooth pursuit algorithms – e.g. suggesting improvements to the mathematical
model.
The “best accuracy”/”most optimal speed of a moving target for obtaining the best accuracy result
for given path” of a smooth pursuit could become a new measure how to evaluate the quality of
smooth pursuit [How well could they follow it in terms of latencies/speed and location of the eyes?
And is it really a pursuit, and not anticipation of the movement for most of the time?]; and perhaps a
recommendation for the speed of a target for eye-tracking experiments using smooth pursuit.
[Would any such recommendation be useful?]
Bibliography:
Barnes, B., & Asselman, P. (1991). ‘The mechanism of prediction in human smooth pursuit‘. Journal
of Physiology. 439-461pp.
Groh, J., Born,R., & Newsome W. (1997). ‘How Is a Sensory Map Read Out? Rffects of
Microstimulation in Visual Area MT on Saccades and Smooth Pursuit Eye Movements‘. The Journal of
Neuroscience. 17(11), 4312-4330pp.
Holmqvist, K. et al. (2011). Eye tracking: a comprehensive guide to methods and measures. New York:
Oxford University Press.
Investigative ophthalmology & visual science. (1978). Effect of aging on horizontal smooth pursuit.
Sharpe, J.. & Sylvester, T. 465-480pp.
Jansson, D. (2014). Mathematical Modeling of the Human Smooth Pursuit System. (Doctoral
dissertation). Retrieved from < http://uu.diva-portal.org/smash/get/diva2:724406/FULLTEXT01.pdf>,
120pp.
Krauzlis, R. (2004). ‘Recasting the Smooth Pursuit Eye Movement System‘. Journal of
neurophysiology. 91(2), 591-603pp.
Robinson, D., Gordon, J., & Gordon S. (1986). ‘A model of the smooth pursuit eye movement system’.
Biological Cybernetics. 55,1, 43-57pp.
Project Plan Eye-Tracking Course
4 Alena Holubcová
Terry B., & McDonald, J. (1983). ‘Smooth pursuit eye movements in response to predictable target
motions’. Vision Research. 23(12), 1573-1583pp.
Winterson, J., & Steinman, R. (1978). ‘The effect of luminance on human smooth pursuit of perifoveal
and foveal targets’. Vision research. 1165-1172pp.