Content uploaded by Keishoku Sakuraba
Author content
All content in this area was uploaded by Keishoku Sakuraba on Jan 16, 2015
Content may be subject to copyright.
23
Yoshimitsu KOHMURA*1, Kazuhiro AOKI*2, Kazuhiro HONDA*1,
Hiroshi YOSHIGI*2 and Keishoku SAKURABA*2
*1 Graduate school of Health and Sports Science, Juntendo University
1-1 Hiragagakuendai, Inba-Mura, Inba-Gun,Chiba 270-1695
yoshimitsu_koumura@hotmail.co.jp
*2 School of Health and Sports Science, Juntendo University
Received July 11, 2008 ; Accepted November 4, 2008
The purpose of this study was to clarify the relationship between dynamic visual acuity and
saccadic eye movement. Twenty-seven young adults, mean age 21.3±2.4 years, participated in this
research. Electrooculography (EOG) was employed for analysis of the saccadic eye movement.
Saccadic eye movements were recorded during measurements of dynamic visual acuity. Peak velocity,
angle, and latency of saccadic eye movement were measured employing EOG. As a result, there
were no relationships between dynamic visual acuity and peak velocity and angle of saccadic eye
movement. However, there was signicant correlation between dynamic visual acuity and latency of
saccadic eye movement (at target velocity 49.5 rpm: r=-.734, p=.000, 47.6 rpm: r=-.619, p=.001, 45.1
rpm: r=-.538, p=.004, 42.5 rpm: r=-.600, p=.001, 40.0 rpm: r=-.478, p=.012). It was suggested that
early start of saccadic eye movement is one of the important factors in the accurate discrimination of
a moving target at high speed.
Key words :
Electrooculography, Latency, Peak Velocity
Human Performance Measurement Vol. 5, 23-30 (2008)
The relationship between dynamic visual acuity and saccadic
eye movement
1. Introduction
While static visual acuity for identifying a stationary
target is one of the basic visual functions, there are many
other important visual functions. In particular, the visual
function designed to identify moving targets has been
examined by a considerable number of researchers. Brug
(1966) and Ludvigh & Miller (1958) have published a
number of reports on dynamic visual acuity (DVA), the
ability to identify a target which moves horizontally
in front of the eyes. DVA is reported to decline with
increasing age after the 20th year of life (Burg, 1966;
Ishigaki & Miyao, 1994). In general, the ability to
identify a moving object is considered essential for daily
activities, such as driving a car or participation in sports
activities. It has been reported that athletes have excellent
DVA (Stine et al., 1982; Ishigaki & Miyao, 1993), and the
effect of training on DVA has also been examined (Long
& Rourke, 1989).
While the methods and results of DVA measurement
have often been reported, few researchers have examined
the elements and mechanisms of DVA. Hoffman et al.
(1981) have reported that the human attributes involved
in DVA are the resolving power of the retina, peripheral
awareness, oculomotor abilities, and the psychological
ability of the individual to interpret what is seen.
Measured and assessed by a method that requires the
individual to track a target which moves horizontally,
DVA is thought to have a particularly strong association
with eye movement. Some research has been conducted
on the relation between eye movement and DVA (Brown,
1972; Reading, 1972). Among the various types of eye
movement, saccadic eye movement plays a central role
in tracking a target which moves at a high velocity; and
DVA is thought to be related to saccadic eye movement
(Barmack, 1970; Ishigaki, 2000).
Reading (1972) reported on eye movement during DVA
measurement based on the results of analysis utilizing
Paper
24 KOHMURA ・ AOKI ・ HONDA ・ YOSHIGI ・ SAKURABA
electrooculography (EOG). The velocity of the target, the
object of analysis in his study, was low in comparison with
that employed by current DVA measurement methods.
A similar tendency is observed in other studies on DVA
and eye movement conducted around the same period
(Barmack, 1970; Brown, 1972). What is more, these studies
were rather inadequate in terms of the number of subjects
examined in the analysis of DVA and eye movement. In
the studies on saccadic eye movement, peak velocity and
latency, which is the starting time of eye movement, are
often used as an index (Fukushima et al., 2000; Moschner
& Baloh, 1994). According to Moschner & Baloh (1994),
the saccadic eye movement of young subjects has shorter
latency and higher peak velocity in comparison with that
of elderly subjects. Brown (1972) has reported that latency
becomes longer as the velocity of the target decreases when
the target velocity in DVA measurement is comparatively
slow. Considering these study results, there may be merit
in investigating the relation between DVA and saccadic
eye movement for the sake of further clarification.
EOG is a representative method for the analysis of eye
movement. Taking advantage of the fact that the cornea
has positive electric potential to the retina and the fact that
voltage change associated with eye movement has a near
proportional relation to eye rotation angle, EOG detects
any electric change through the electrodes applied to the
skin surrounding the eyes (Arai et al., 2001; Yonemura,
2007; Yamauchi et al., 2003). Because of problems in
obtaining accurate positional information in the analysis
of eye movement, EOG is thought to be inappropriate
for the detection of prolonged eye movement (Kuno
et al., 2003; Sakashita et al., 2006). On the other hand,
EOG does not require any apparatuses that block vision,
such as the video camera that needs to be placed on the
examinee when employing the corneal reflex method,
another method used in eye movement measurement.
Therefore, EOG is less burdensome for the subject than
other methods, though the electrodes applied to the
skin do have the potential to cause slight discomfort.
Employment of EOG also makes it possible to examine
a relatively large number of study subjects and to detect
a wide range of eye movement (Miyashita et al., 2008;
Sakashita et al., 2006).
In actuality, there are relatively few studies that have
been published on the mechanism of DVA. Much of the
relation between DVA and saccadic eye movement has
yet to be clarified. It is meaningful to conduct a basic
investigation of DVA in order to elucidate the mechanism
and relevant elements of DVA, and to contribute to the
enhancement of DVA measurement accuracy. This study
aimed to quantify the characteristics of saccadic eye
movement and to investigate the relation between DVA
and saccadic eye movement.
2. Procedures
2.1. Subjects
The subjects of this study were university undergraduate
and graduate students whose corrected or uncorrected
visual acuity was 1.0 or higher (static visual acuity) and
who exercised on the regular basis or who had experienced
6 to 12 years of participation in sports activities.
Subject selection was not based on specialization in any
particular sport or sports. The number of the subjects
was 27 (including 8 females); namely, 9 subjects with an
uncorrected visual acuity of 1.0 or higher, 13 contact lens
wearers, and 5 wearers of framed corrective lenses. The
mean age of subjects was 21.3±2.4. Prior to entry into
the study, written informed consent was obtained from
the individual subjects after a detailed explanation of the
content of the experiment and the object of the study. This
study was conducted with the approval of the Research
Ethics Committee of the Juntendo University Graduate
School of Health and Sports Science.
2.2. Measurement of Dynamic Visual Acuity
(DVA)
DVA measurement was conducted with the use of a
dynamic visual acuity analyzer (Kowa Co. Ltd. HI-10). In
the HI-10, a Landolt ring, which is projected and reflected
by a mirror, moves from left to right on a semi-circular
screen with a visual angle of 90º. The screen was situated
80cm in front of the subject. The subject was required to
track the Landolt ring and identify the gap in the ring. The
subject was also directed to place his or her chin on the
chin support and to not move his or her head. The Landolt
ring has a visual angle of 10 minutes. Use of a larger-
sized Landolt ring causes the measurement values to be
topically distributed around the peak value (Kohmura &
Yoshigi, 2004). Considering this, a smaller Landolt ring
25
Dynamic Visual Acuity and Saccadic Eye Movement
was adopted in this study.
The DVA of each subject was measured according
to a method in which the speed of the target gradually
decreases. The velocity of the target was set initially at
49.5rpm, the maximum velocity of this measurement
device, and reduced gradually thereafter. The subjects
were required to f lip a switch at the moment the ring
gap was identified and to identify the position (up, down,
right, or left) of the gap before the Landolt ring appeared
at the center of the screen. The rotating velocity at the
moment the subjects correctly f licked the switch was
recorded as the DVA value. Measurement was repeated
until the subjects correctly identified the position of the
ring gap 10 times. The subjects were directed to exercise
caution in correctly identifying the position of the gap.
The mean value of DVA for each subject was calculated
from the DVA values for the 10 correct measurements,
excluding the maximum and minimum values. The mean
value obtained by this method was used as the DVA value
for the individual subjects.
2.3. EOG Recording
Following earlier studies (Brown et al., 2006;
Reading,1972; Munoz et al., 1998), all EOG signals were
electrically amplified by an amplifier (EOG100C) and
were digitally recorded with the use of a data collection
and analysis system (Biopac Systems Co. Ltd.: MP150;
Monte System Corporation). Using AcqKnowledge
software, peak velocities, latencies, and angles were
calculated as the indexes for analyses. An electrode was
placed laterally to each eye and a ground electrode was
attached to the frontal region of the head. Trigger stimulus
was input to another channel at the moment the eye-
tracking target appeared on the screen. Voltage and angle
were calibrated with respect to each subject and to each
measurement.
The peak velocity of the first saccadic eye movement
after the input of the trigger stimulus was calculated. In
terms of latency, the length of time from the input of trigger
stimulus to the point at which the velocity of saccadic eye
movement (i ncluding pe ak veloc it y) exc eeded 10 0 deg ree/
sec was calculated. The change in angle from the point at
which the velocity of saccadic eye movement exceeded
100 degree/sec to the point at which it fell below 100
degree/sec was also calculated. Though the reference
velocity varies depending on the researcher, the velocity
of 100 degree/sec was adopted in this study based on
previously published study literature (Yonemura, 2007).
Depending on the subject, when the velocity is 37.5rpm
or lower, more than one peak of saccadic eye movement
can be observed. Considering this, only those angles
obtained when the target velocity was 40rpm or higher
were analyzed in this study.
All the saccadic eye movement values with the
above-mentioned indexes were averaged to determine
the central value for each subject at each velocity, with
which analyses were conducted. In principle, any case
which failed to fulfill the above-mentioned criteria, which
lacked peak velocity, latency, or angle data, or which had
major saccadic eye movement prior to the input of trigger
stimulus was excluded from analysis.
Any change in the velocity of the target during the
recording of EOG can affect the elements of saccadic eye
movement. The method in which the subject is required
to stop the measurement device at the moment he or she
identifies the position of the gap in the target Landolt
ring whose velocity is gradually decreasing presents
difficulties in the unification of analytical items due to the
fact that measurement conditions and recording times can
vary depending on subject or measurement. Considering
these factors, EOG recording was conducted in this study
as follows:
EOG was recorded utilizing a DVA measurement
device with fixed target transfer velocity. Each subject
was directed to track the target until it passed ahead of
him or her 10 times. The examiner counted the number of
passes and stopped the measurement device after the 10th
pass of the target. No matter whether the subject identif ied
the gap in the Landolt ring or not, EOG continued to be
recorded until the 10th pass of the target. The subject
was required to identify the position of the ring gap at
the moment it was identified and to continue tracking the
target thereafter until it passed ahead of him or her 10
times. The results were recorded by the examiner. Velocity
was set at 15 levels with approximate intervals between
levels set at 2.5rpm. Due to the mechanical constraints
of the measurement device, velocity intervals of exactly
2.5rpm were difficult to achieve. The 15 levels of velocity
that were adopted were 49.5, 47.6, 45.1, 42.5, 40.0, 37.5,
35.0, 32.6, 30.0, 27.5, 25.0, 22.5, 19.9, 17.5, and 15.0 rpm.
26 KOHMURA ・ AOKI ・ HONDA ・ YOSHIGI ・ SAKURABA
2.4. Analytical and Statistical Processing
The level of statistical significance in this study was
5%.
2.4.1. Relation between DVA and Correctness of
Identication
In order to record EOG under stable conditions, the
velocity of the moving target and the number of passes
were fixed in a unified manner. While DVA measurement
is customarily conducted with a fixed target velocity (Burg,
1966, Millslagle, 2004), only the subjects' identification
of the ring gap was pursued in this study for the sake of
EOG recording. Both the ability to identify the target at
a declining velocity and the ability to identify the target
moving at a certain velocity can be regarded as the ability
to identify a target which moves horizontally; therefore, it
is necessary to examine the relation of these two abilities.
Since only the subjects' identification of the ring gap
was measured in this study, the Pearson's correlation
coefficients of the number of correct identifications
during the 15-level EOG recording and the evaluated
values of DVA were calculated and the partial correlation
coefficients thereof, excluding the effects of static visual
acuity, were determined.
2.4.2. Relation between DVA and Elements of Saccadic
Eye Movement
The relationships between DVA values and the
elements of saccadic eye movement were analyzed. A
decline of eye movement velocity and an extension of
latency were expected when the velocity of the target
was low. If the velocity of the target at which most of the
subjects correctly identified the ring gap had been used
for analysis, the relation between DVA and the respective
elements might have been affected. In the EOG record for
the velocity of 37.5rpm or lower, the number of the subjects
who misidentified the ring gap was less than one third
of all the subjects, an extremely small value. Therefore,
an analysis of saccadic eye movement recorded when the
velocity of the target was 40rpm or higher was conducted.
Pearson's product-moment cor relation coefficient was
used for correlation analysis.
The subjects who correctly and incorrectly identified
the ring gap at a target velocity of 40rpm or higher were
compared and analyzed in terms of latency. t-test was
used for the comparison.
3. Results
From the results of the analysis of the relation between
DVA and the correctness of identification during EOG
recording, the correlation coefficient was determined to
be 0.672 (p = 0.000). The partial correlation coefficient
between DVA and correct identification excluding the
effect of static visual acuity was determined to be 0.661
(p = 0.000).
Table 1 shows the mean values and SDs of the elements
of eye movement at the velocity of 40.0rpm or higher.
Table 2 shows the correlation coefficients between the
evaluated values of DVA and the elements of saccadic eye
movement. There were significant correlation coefficients
between latency and the evaluated values of DVA (target
velocity 49.5 rpm: r =-.734, p =.000, 47.6 rpm: r =-.619, p
=.001, 45.1 rpm: r =-.538, p =.004, 42.5 rpm: r =-.600, p
=.001, 40.0 rpm: r = -.478, p =.012).
Table 3 shows the results of the comparison of latencies
between the subjects who correctly and incorrectly
identified the ring gap. The numbers of subjects of the
respective groups (C: subjects who correctly identified;
M: subjects who misidentified) at the respective velocities
were as follows: 49.5 rpm: C-10, M-17, 47.6 rpm: C-11,
Mean SD Mean SD Mean SD Mean SD Mean SD
Peak Velocity (dgree/sec)668.8106.3665.191.4673.3 80.2 665.188.9 666.190.5
Latency (sec) .111 .016 .108 .018 .108 .015 .111 .015 .112 .015
Angle (dgree)57.110.357.310.457.2 10.5 56.510.457.310.2
49.547.645.142.5 40.0
Target Velocity (rpm)
Table1. Saccadic eye movement measurements
27
Dynamic Visual Acuity and Saccadic Eye Movement
M-16, 45.1rpm: C-11, M-16, 42.5 rpm: C-16, M-11, 40.0
rpm: C-16, M-11. Figure 1 shows a waveform chart of the
subjects with correct and incorrect identification during
the EOG recording.
4. Discussion
In this study, investigators attempted to quantify the
characteristics of saccadic eye movement and examine
the relation between DVA and saccadic eye movement.
There are both advantages and disadvantages in analyzing
saccadic eye movement by the EOG method. Considering
that this method was suitable for the clarification of the
characteristics of saccadic eye movement when tracking
the target with the use of a DVA measurement device,
however, the EOG method was adopted in this study.
Since it was difficult to obtain accurate positional data for
eye movement by the EOG method (Miyashita et al.,2008;
Sakashita et al., 2006) and to define the transferred angles
in saccadic eye movement with multiple peak velocities,
analysis of angles was only partially conducted.
As a result of analysis, it was clarified that the
correlation coefficient between the number of the correct
answers and the evaluated values of DVA was 0.672
and the partial correlation coefficient was 0.661 when
excluding the inf luence of static visual acuity. Although
both expressed the ability to discern a target which moved
horizontally, a sufficiently high correlation coefficient
was not obtained. This may have resulted from the
difference in measurement methods and from the choice
of the number of correct answers as an analysis object.
However, both evaluation items have been used for DVA
assessment and both have been thought to express the
ability to identify a target which moves horizontally.
The relation between these methods in terms of DVA is
not thought to be affected by static visual acuity. Since
the number of correct answers were all the data of DVA
obtained during EOG recording in this study, the relation
r .076 .072 .056 -.090 .059
p .707 .721 .781 .656 .770
r -.734
**
-.619
**
-.538
**
-.600
**
-.478
*
p .000 .001 .004 .001 .012
r .155 .012 -.081 -.062 .074
p .439 .954 .689 .759 .713
Peak Velocity
Latency
Angle
*:p<.05,**:p<.01
40.049.547.645.1 42.5
Target Velocity (rpm)
Target Velocity (rpm)
Latency (sec) Mean SD Mean SD Mean SD Mean SD Mean SD
correct answer .101 .012.100.018.105 .013 .105 .014 .110 .012
miss .117 .014.114.015.110 .016 .121 .010 .115 .018
t-value
p
49.5 47.6 45.1 42.5 40.0
3.26
.00
.83
.41
2.83
.01
2.21
.04
.83
.41
Table2. Correlation coefcients between DVA and elements of saccadic eye movement
Table 3. Comparison of latencies between correct and incorrect answer
28 KOHMURA ・ AOKI ・ HONDA ・ YOSHIGI ・ SAKURABA
between the saccadic eye movement and the evaluated
values of DVA was emphasized in the analysis.
In terms of peak velocities and angles, no significant
correlation coefficients were found between DVA
and elements of saccadic eye movement. Meanwhile,
significant correlation coefficient was seen in terms of
latency (Table 2). From this, it was suggested that the
subjects who exhibited proficient DVA tended toward
being quick to begin saccadic eye movement. Latency
of saccadic eye movement has been used as one of the
evaluation indexes. It has been reported that latency
becomes shorter with the development of saccadic eye
movement (Fukushima et al., 2000) and that young
subjects have shorter latency than elderly subjects do
(Moschner & Baloh, 1994; Munoz et al., 1998). According
to Fukushima et al. (2000), latency involves a complex
combination of procedures in brain for visual information
processing from the trigger towards visual acuity to the
start of saccadic eye movement and for exercise signal
transformation. They reported the development of latency.
DVA has been reported to develop at around the time that
latency develops (Ishigaki & Miyao, 1994). While the
DVA of athletes often draws public attention, professional
cricket players have been reported to have shorter latency
than low-level cricket players have (Land & McLeod,
2000). Considering these reports, the results of this study
which showed a significant relation between DVA and
latency cannot be judged as inconsistent.
The subjects who gave correct answers had shorter
latency than those who gave incorrect answers (Table
3). This result was not unexpected due to the fact that
there was an association between the evaluated values
of DVA and the correctness of the answers during the
recording of EOG. It should be considered, however,
that this result could include the subjects who answered
correctly or incorrectly only at a certain velocity. Figure
1 is a waveform chart of a typical case. It may be possible
that cor rect identif ication of a target which moves rapidly
becomes difficult when a delay in the start of saccadic eye
movement occurs. Correct identification of the target is
not exclusively attributed to an early start of saccadic eye
movement. Considering that the target disappears from
the screen in approximately 0.3 second at a velocity of
49.5rpm, however, an early start of saccadic eye movement
may be one of the conditions for correct identification.
Judging from the results of this study, it may be
difficult to conclude that DVA becomes better as the
peak velocity becomes higher and as the angle becomes
Figure1. Image of typical EOG at a target velocity of 49.5 rpm
Left 45°
Right 45°
Appearance of the target Typical waveforms of two subjects
who could not identify targets
Typical waveforms of two subjects who
could indentify targets
29
Dynamic Visual Acuity and Saccadic Eye Movement
larger (Table 2). There is a study reporting that the peak
velocity of saccadic eye movement of young subjects is
higher than that of elderly subjects (Moschner & Baloh,
1994). The results of this study could be affected by the
fact that all the subjects were approximately the same age
and that the experiment required them not only to move
their eyes quickly but also to track and identify the target
as well. Considering that peak velocity develops earlier
than both DVA and latency develop (Ishigaki & Miyao,
1994; Fukushima et al., 2000), however, peak velocity of
saccadic eye movement may be poorly linked to the level
of DVA.
When a subject tracks the moving target, he or she does
not perform only saccadic eye movement at a constant
velocity but also various types of complex eye movement.
In this study, the mechanism of eye movement and accurate
positional information of the gaze were not clarified.
Therefore, the analyses in this study are supposed to be
insufficient in terms of the accuracy of eye movement
towards the target. In an earlier study, positional and
velocity errors of eye movement towards a target were
analyzed, though not employing the EOG method (Brown,
1972). While no significant relation was observed between
the transfer angle in one saccadic eye movement and DVA
in this study, it will be necessary to consider positional
error and angle in future investigations. Considering
that a significant relation was observed in this study
between latency and DVA, some relation may be observed
between the ability to track the target correctly and the
level of DVA. In recent years, various types of extremely
expensive and highly-accurate eye mark recorders have
become available. Quantification of eye movement with
the use of one of these newly developed devices will
allow researchers to accomplish more detailed analyses.
Eye movement employed in the acquisition of a target
moving at a declining velocity was not analyzed in this
study. This should be investigated with the use of more
advanced research methods and equipment in the future
studies, while paying attention to partial differences in
measurement methods.
5. Conclusion
Based on the results of the analyses of saccadic
eye movement by the EOG method in this study, it is
suggested that DVA has a stronger relation with latency
than with the peak velocity of saccadic eye movement
and angle. While having complex physiological elements,
latency is reported to develop at around the time when
DVA develops. It is speculated that the subjects who are
excellent in DVA tend to have a shorter latency. It will be
necessary to develop research methods and equipment for
the quantification of the accuracy of tracking a target in
DVA measurement and to investigate this field further in
the future.
References
Arai, S., Itsuki, N., and Kubo, M. (2001) Characteristics
of electrooculogram generated with eyeball rotation
on eccentric eye position. IEICE technical report, ME
and bio cybernetics 101(181): 67-70. (in Japanese
with English abstract)
Barmack, B. (1970) Dynamic visual acuity as an index
of eye movement control. Vision Research 10: 1377-
1391.
Brown, B. (1972) The effect of target contrast variation
on dynamic visual acuity and eye movement. Vision
Research 12: 1213-1224.
Brown, M., Marmor, M., Vaegan, Zrenner, E,m Brigell,
M., and Bach, M. (2006) ISCEV standard for
clinical electro-oculography (EOG). Documenta
ophthalmologica 113: 205-212.
Burg, A. (1966) Visual acuity as measured by dynamic
visual acuity and static test. Journal of Applied
Psychology 50: 460-466.
Fukushima, J., Hatta, T., and Fukushima, K. (2000)
Development of voluntary control of saccadic
eye movements I. Age-related changes in normal
children. Brain & Development 22: 173-180.
Hoffman, G. L., Rouse, M., and Ryan, B. J. (1981)
Dynamic visual acuity: a review. Journal of the
American Optometric Association 52: 883-887.
Ishigaki, H., and Miyao, M. (1993) Differences in dynamic
visual acuity between athletes & nonathletes.
Perceptual and Motor Skills, 77: 835-839.
Ishigaki, H., and Miyao, M. (1994) Implications for
dynamic visual acuity with changes in age and sex.
Perceptual and Motor Skills 78: 363-369.
Ishigaki, H. (2000) On the characteristic of human
dynamic visual acuity. The bulletin of the Research
Institute for Modern Physical Education Kyoto
30 KOHMURA ・ AOKI ・ HONDA ・ YOSHIGI ・ SAKURABA
Sangyo University 9: 61-67. (in Japanese with
English abstract)
Kohmura, Y., and Yoshigi, H. (2004). Present conditions
and problems on measurement and evaluation of
sports vision. Japan Journal of Test and Evaluation of
Physical Education and Sports 4: 37-43. (in Japanese
with English abstract)
Kuno, Y., Yagi, T., Fujii I., Koga, K., and Uchikawa Y.
(1998) Development of eye-gaze input interface
using EOG. Transactions of Information Processing
Society of Japan 39(5): 1455-1462. (in Japanese with
English abstract)
Land, M. F., and McLeod, P. (2000) From eye movements
to actions: how batsmen hit the ball. Nature
Neuroscience 3: 1340-1345.
Long, G. M., and Rourke, D. A. (1989). Training Effects
on Resolution of Moving Targets Dynamic Visual
Acuity. Human Factor 31(4): 443-451.
Ludvigh, E., and Miller, J. W. (1958). Study of the visual
acuity during the ocular pursuit of moving test
objects. Journal of the Optical Society of America
48: 799-802.
Millslagle, D. (2004) Coincidence anticipation and
dynamic visual acuity in young adolescents.
Perceptual and Motor Skills 99: 1147-1156.
Miyashita, H., Hayashi, M., and Okada, K. (2008)
Eye-gaze interface on HMD for virtual reality.
Proceedings of the Virtual Reality Society of Japan,
Annual Conference 13(CS-2): 25-30. (in Japanese
with English abstract)
Moschner, C. and Baloh, R. W. (1994) Age-related
changes in visual tracking. Journal of Gerontology
49(5): 235-238.
Munoz, D. P., Broughton, J. R., Goldring, J. E., and
Armstrong, I. T. (1998) Age-related performance of
human subjects on saccadic eye movement tasks.
Experimental Brain Research 121: 391-400.
Reading, V. M. (1972) Analysis of eye movement
responses and dynamic visual acuity. Pügers Archiv
333: 27-34.
Sakashita, Y., Fujiyoshi, H., and Hirata, Y. (2006)
Measurement of 3D eye movements using image
processing. Journal of the Japanese Society for
Experimental Mechanics 6(3): 236-243. (in Japanese
with English abstract)
Stine, C. D., Arterburn, M. R., and Stern, N. S. (1982)
Vision and Sports -A review of the literature-.
Journal of American Optometric Association 53(8):
627-633.
Yamauchi, Y., Itsuki, N., Kubo, M., and Shinomiya, K.
(2003) New measuring method of electrooculogram
for analysis of battery model of the eyeball. IEICE
technical report. ME and bio cybernetics 103(184):
31-34. (in Japanese with English abstract)
Yonemura, T. (2007) Velocity perception based on pursuit
eye movements. Kyushu University Psychological
Research 8: 143-154. (in Japanese with English
abstract)
