Transcript Eye Movements of Younger and Older Drivers

Eye Movements of Younger and
Older Drivers
Professor: Liu
Student: Ruby
Objective
• The study experiment younger and older
people’s eye movements in task-directed and
task-undirected visual search.
• Recorded the search times, number of fixations,
and fixation durations from one digit to the
next.
References
• Older people have poorer visual tracking skills
because their peripheral vision was decreased.
(Whiteside,1974; Bono et al., 1996; Kanayama
et al., 1994; Kline, 1994; Olincy, Ross, Young,
& Freedman, 1997).
• Using the UFOV discovered that crashes were
related to older drivers’ age, visual acuity, or
mental status. (Ball, Beard, Roenker, Miller, &
Griggs, 1988; Ball & Owsley, 1991; Ball et al.,
1993; Ball & Rebok, 1994)
References
• The inattention was a more reliable factor of
crashes than visual functions or eye health.
(Ball et al., 1993)
• Older participants’ mean search times were
longer than younger participants, therefore
older people have a less ability to search.
(Schieber, 1989)
Method
• Participants:
– Older group: 3 women and 2 men.
age from 20 to 30 years old.
median driving experience: 37 years.
– Younger group: 3 women and 2 men.
age from 62 to 80 years old.
median driving experience: 8 years.
The eyesight required in 6/9.
Method
• Stimulus photos and equipments
– Four images were chosen which included
pedestrians, oncoming, passing, and
intersecting cars.
– Two images which used in Schieber’s research,
one had numeric in the image and the other did
not have.
– 386 PC and 14-inch monitor.
– ASL 210 eye tracking system and SpectraSensors which collect the data.
The CVRT had numerical in the image
Method
• Procedure
– Training section: participant was presented with a
sample image.
– The numerical overlay image was always shown
second and fifth in the series.
– As each number was found, the participant pressed
a button.
– A scanning time of 10 s was allocated per image.
Results & Discussion
• Numerical Overlay Pictures
– Search time
• The mean search time to complete the task for the
older participants was significantly higher than
younger participants. t = 2.33, p =.01.
Results & Discussion
• The longest search times of the older
participants were significantly longer
than younger participants t(29) = 4.7,
p < .001.
Results & Discussion
• Fixation
– The number of fixations showed significantly
more fixations for the older group, F(1, 8) =
16.85, p = .003.
• The longer total search time for the older participants
was caused by an increase in the number of fixations
rather than an increase in their duration.
– There was a significant drop in saccade amplitude
from the first presentation (mean = 6.7°) to the
second one (mean = 6.1°), F(1, 8) = 11.04, p =.01.
Results & Discussion
• Fixation
– Search duration correlated negatively with the
mean saccade amplitude, r = –.45, p = .01.
• Participants were having a difficult search, which
their saccade amplitudes tended to be shorter.
Results & Discussion
• CVRT. (the number reached (out of 14) in 10 s.
– The number of fixations related with the participant’s
test scores, r = –.73, p = .02.
– Lower relation between saccadic amplitude and
CVRT, r = .62, p = .06.
– A small negative correlation between CVRT test score
and search time.
Results & Discussion
• Comparing Visual Search of the Numerical
Overlay and Traffic Scene Images
– There was a high relation between the images, r =
0.79, p = .01.
• Saccadic amplitude is also a consistent mark of a person’s
visual search behavior.
– Rayner & Pollatsek in 1992 discovered that the
thickness of information on the image which
influence strongly on saccadic extent.
Results & Discussion
• Traffic Scene Images
– The authors divided the four traffic scene images
into areas of interest (AOI) appropriate for driver
attention.
• Mean dwell time outside the AOIs was 20%–34%.
• There was no significant effect of age, F(1, 6) = 2.39,
p = .17.
Results & Discussion
• There were significant interactions between
age group and AOI for:
– Image 3, F(3, 18) = 10.5, p < .001.
– Image 4, F(2, 16) = 4.01, p = .04.
• The older participants in both of images tended to
ignore some AOIs while focusing for an very long
time on a subset of the AOIs.
Results & Discussion
• Concentration ratio
– The degree to which certain areas are fixated on
more than expected.
– Older people really concentrate their fixations
on a more limited number of AOIs.
• Concentration ratio :
Σ|Fi – Ai|.
Results & Discussion
• Older participants showed increased concentration
in a smaller number of AOIs.
• The mean concentration ratio of the old group was
significantly higher than young group, F(1, 6) =
18.5, p = .005.
• There was significant interaction between image
and age, F(3, 18) = 6.1, p = .005.
– Only three out of the four images reflect difference in
the concentration ratio.
Results & Discussion
• Two possible visual search models could explain
the differences between the older and younger
participants.
– Older participants confine the number of AOIs that they
scan and dwell for longer periods of time.
– Older participants’ fixations jump one AOI to the next,
and the confine number of AOIs on which they fixate
reflects saccadic movements to AOIs.
Results & Discussion
• stability ratio: Px|y(x|y)/Px(x).
– If the ratio is close to 1, then the participant’s
fixation pattern is fairly random or uniform
across the whole image.
– The participant’s fixation in a particular AOI is
independent of the previous fixation.
– As the ratio gets higher, the visual search
becomes stable.
Results & Discussion
• The younger participants were stable than the
older participants, F(1, 6) = 13.02, p = .01.
• The main effect of image and the interaction
between image, F(3, 18) = .69, p = .57 and
age were not significant, and F(3, 18) = .39, p
= .76.
Conclusion
• Older persons need longer visual search times than
younger persons to get the same information.
• Older people’s visual search behavior is characterized:
(a) Large variability in the data from the older group.
(b) Occasional lapses that occurs in searching
difficulty.
Conclusion
• The older participants’ eye movements were
characterized by shorter saccades and increased
numbers of fixations caused by:
(a) A poor visual search process with ineffective
use of the peripheral field
(b) A failure to get amounts of information from
areas that were before fixated on.
Conclusion
• The older participants focused on a smaller subset
of the areas of interest within the images.
• The younger participants distributed their attention
more equal different parts of the scene.
• People tend to have the same general type of
saccadic movement for different types of tasks.
• Saccadic eye movements may be a stable indication
of individual differences.