Factors Inuencing Reaction Time during

Automated Kinetic Perimetry on the Tubingen

Computer Campimeter
Stephan T. Becker,
1
Reinhard Vonthein,
2
Nicholas J. Volpe,
3
and Ulrich Schiefer
1
PURPOSE. To determine the effect of age, examination, location,
luminance, subject, and vigilance on reaction time (RT) in
automated kinetic perimetry.
METHODS. Thirty-six normal volunteers (2070 years old) un-
derwent kinetic campimetric examinations, during which RTs
(time from appearance of a stimulus to response) were re-
corded. Stimuli with a diameter of 26 minutes of arc (Gold-
mann III) were presented on horizontal vectors with an angu-
lar velocity of 2 deg/s. Thirty-two positions in the 30 radius
visual eld were tested six times, at luminance levels of 41.62
and 110 cd/m
2
(background 10 cd/m
2
). An analysis of RT
variance (ANCOVA) was performed.
RESULTS. Median RT increased with age from 370 ms (2030-
year-old subjects) to 440 ms (6070-year-old subjects). There
was a strong dependency of RT from the individual subjects
examined (means range, 313411 ms), from the course of one
examination period, from the examination and stimulus lumi-
nance, and from the location in the visual eld.
CONCLUSIONS. Reaction time during automated kinetic perime-
try varies considerably. This study shows that the factors of
age, the subject examined, his or her daily condition, the
course within one examination period, and the stimulus loca-
tion and luminance are of relevance. For a much more reliable
prediction of individual RTs and thus a precise assessment of
local kinetic thresholds, application of some additional kinetic
stimuli (RT time vectors) within the intact visual eld areas is
necessary. (Invest Ophthalmol Vis Sci. 2005;46:26332638)
DOI:10.1167/iovs.04-1413
S
ince the beginning of the past century, nonperimetric psy-
chophysical studies have shown that simple reaction times
(RTs) increase with increasing stimulus eccentricity
13
and
decrease with increasing light intensity.
38
More recently, RTs
during static perimetric testing have been studied for different
reasons. Frisen
9
described a method in static perimetry that
shortens the examination by individually adjusting the inter-
stimulus intervals to RT, which leads to an increase in the
patients comfort. Moreover, Bengtsson et al.
10
and Olsson et
al.
11
dened RT windows for a new algorithm of threshold
estimation in static automated perimetry called SITA (Swedish
interactive thresholding algorithm). They rated RTs below a
lower limit of 180 ms as false-positive catch trials, thereby
saving time for extra presentations of conventional catch trial
stimuli. Response times can also be used to remove a large
percentage of false-positive responses selectively, further im-
proving the performance of suprathreshold perimetry (Henson
and Artes
12
). Furthermore, Rouland and Hache
13
and Wall et
al.
14
attempted to develop special RT evaluation testing meth-
ods for the early diagnosis of glaucoma in patients.
RT measurement is even more important in kinetic perim-
etry than in static perimetry, because RT induces a systematic
shift of the scotoma border toward the direction of stimulus
movement. As stimuli are typically traveling from nonseeing
areas (periphery of the visual eld) toward seeing areas (center
of the visual eld) in kinetic perimetry, RT causes a systematic
shrinkage of intact visual eld area and a systematic increase
of visual eld defects, respectively. Moreover, variation of RT at
subsequent visits may obscure the course of visual eld
change.
Because traditional kinetic perimeters are operated manu-
ally, the results strongly depend on the perimetrist and indi-
vidual technique. Not only is the angular velocity of the stim-
ulus unknown, but the direction of motion is also
undocumented. Thus, the determination of the exact visual
eld threshold is uncertain, since its measurement varies with
the RT of both the perimetrist and the patient examined.
Threshold measured with modern computer-controlled perim-
eters, which automate stimulus presentation and recording of
the patients response, are only affected by the RT of the
patient. RT measurements with moving stimuli were rst per-
formed by Eischer-Desrivieres,
15
who reported shorter re-
sponse times when moving stimuli were presented in the
center of the visual eld than when they were presented in the
periphery.
In an effort to determine more accurately the location of
kinetic thresholds, we studied the inuence of subject age,
inter- and intraindividual variations, stimulus location in the
visual eld, stimulus luminance, and vigilance on the RT of
healthy subjects during automated kinetic perimetry.
METHODS
Subjects
Thirty-six healthy subjects (ratio of women to men, 20:16) from three
age groups (2030 years, 7:5; 4050 years, 8:4; 6070 years, 5:7)
participated in the study. The protocol adhered to the tenets of the
Declaration of Helsinki. After appropriate informed consent and sub-
ject instruction, a health questionnaire was completed including ques-
tions about previous vision problems and use of any medications with
potential effect on CNS function. Patients were given an allowance to
defer travel expenses.
Ophthalmic inclusion criteria for the subjects were as follows: best
corrected distance and near visual acuity equal or better than 20/20
OU, spherical ametropia less than or equal to 6 D, and cylindrical
From the
1
Department of Pathophysiology of Vision and Neuro-
Ophthalmology, University Eye Hospital, Tubingen, Germany; the
2
De-
partment of Medical Biometry, University of Tubingen, Tubingen, Ger-
many; and the
3
Scheie Eye Institute, Department of Ophthalmology,
University of Pennsylvania School of Medicine, Philadelphia, Pennsyl-
vania.
Submitted for publication December 3, 2004; revised March 1,
2005; accepted March 28, 2005.
Disclosure: S.T. Becker, None; R. Vonthein, None; N.J. Volpe,
None; U. Schiefer (P)
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be marked advertise-
ment in accordance with 18 U.S.C. 1734 solely to indicate this fact.
Corresponding author: Ulrich Schiefer, Department of Patho-
physiology of Vision and Neuro-Ophthalmology, University Eye
Hospital, Schleichstrasse 12-16, D-72076 Tubingen, Germany;
[email protected].
Investigative Ophthalmology & Visual Science, July 2005, Vol. 46, No. 7
Copyright Association for Research in Vision and Ophthalmology 2633
ametropia less than or equal to 2 D. Furthermore, we required
isocoric pupils normal stereopsis (all gures identied, using the Lang
II test), normal ocular alignment, normal ocular motility with no
history of diplopia, and no relative afferent pupillary defect. The IOP of
both eyes had to be less than or equal 20 mm Hg (determined after
perimetry by either air pulse or applanation tonometry), normal refrac-
tive media as assessed with slit lamp and normal optic disc and macular
appearance by direct ophthalmoscopy. Each subject had one eye
examined, chosen in random fashion.
Tubingen Computer Campimeter
The Tubingen Computer Campimeter (TCC) presents kinetic stimuli
on a calibrated high-resolution video display unit
1620
(1024 768
pixels, 75 dpi; Barco, Kippenheim, Germany). At a viewing distance of
30 cm, the 20-in. monitor covers approximately 35 horizontally and
24 vertically of the visual eld radius. The background had a homog-
enous luminance of 10 cd/m
2
. The xation mark was diamond-shaped
and consisted of four red circles (10 minutes of arc in diameter,
identical luminance as the background) that were located at 1 eccen-
tricity. Subjects responded to stimulus perception by pressing a but-
ton. Fixation was directly monitored by the examiner through an
infrared camera attached to the chin and head support, which also
recorded pupil position and horizontal diameter.
Perimetric Settings
The size of the stimuli tested was 26 minutes of arc (corresponding to
Goldmann III). Targets moved linearly along vectors with an angular
velocity of 2 deg/s within the central 30 visual eld (Fig. 1). The test
vectors were located at 16 positions at 0, 15, 20, and 30 eccentric-
ity on the straight and oblique meridians. The pattern was symmetric
with respect to the vertical axis, but the blind-spot area was not tested.
A pair of vectors with opposite directions originated at each location.
From the initial point, one target moved horizontally, either to the right
or to the left. The order of presentation of the vectors was randomized.
Each vector was tested six times with a stimulus luminance of 110 and
41.62 cd/m
2
each. Invisible stimuli (10 cd/m
2
, same luminance as
background) were also presented three times per location. Testing
without visible stimuli served as a false-positive catch trial to see how
often a subject pressed the response key without actually having seen
a stimulus. The visual eld examination could be interrupted anytime
according to the subjects or the examiners needs.
The stimulus appeared suddenly without any acoustic cueing and
moved on the monitor until the subject pressed the response button or
until the stimulus had moved a distance of 6. At this point, it disap-
peared immediately. The time (in milliseconds) between the appear-
ance of a stimulus and the subjects response was dened as RT. To
avoid a rhythm effect, we alternated stimulus presentation after pauses
of various lengths (1200 or 1800 ms).
Data Management
For analysis, the vector pattern of the left eye was mirrored by the
vertical midline to match the right eye pattern. This enabled the
comparison of vectors with the same characteristics. Values above
800 ms and below 170 ms were discarded as not arising from imme-
diate response or response to the preceding stimulus, respectively. The
limits were similar to those used in experiments by other research
groups.
2123
Statistical Analysis
A Box-Cox-transformation was estimated in an analysis of covariance
(ANCOVA) of the raw readings. As a result, the reciprocal value,
distance per 2 seconds, was used as the response in the prediction
model, to reach a closer approximation to a normal distribution. In
graphs and tables, the values were transformed back for clarication.
Data were analyzed with ANCOVAs. For expected values, the 95%
condence interval (CI) was computed.
Experiment Age
In the rst experiment, the effect of age on RT was estimated. The RT
data of 36 subjects examined once were analyzed by the factors
luminance of the signal, age, code, and the covariables eccentricity,
number of presentations (nop) and nop
2
. Three age groups (2030,
4050, and 6070 years) were compared. The factor code allowed an
estimation of an individual effect for each subject. Eccentricity was the
eccentricity (in degrees) of the location where the target appeared in
the visual eld. Stimulus presentations were assigned consecutive
numbers. The covariables number of presentations (nop) and its
square nop
2
were considered to describe a parabolic fatigue trend over
the course of one examination, as each curvature was described by a
quadratic (parabola) and a linear (ascending-descending) term. The
factor effects were given for the 240th of 480 stimulus presentations.
Every subject had his or her own fatigue curve. Because of this, a mix
of fatigue and learning effect for each individual was possible.
Experiment Code
RTs of the rst group (age, 2030 years) were measured on two
different days (interval: 12153 days). In the ANCOVA model, the
factors luminance (of the target presented), eccentricity in the visual
eld, side of the visual eld (nasal, midline, temporal), half (upper part,
midline, lower part), code, examination, examination * code (* de-
notes interaction of two factors), nop, nop
2
, nop * code, and nop
2
*
code were analyzed. Only stimuli that started from the horizontal or
vertical midline were classied as midline. Examination had two pos-
sible values: rst or second visit. In our model, each subject exhibited
a systematic deviation from the general difference between visits (i.e.,
the examination * code interaction). The parabola over nop, too, was
estimated for each subject and as a general trend.
RESULTS
Experiment Age
For analysis of the inuence of age on RT, 36 normal adults in
three groups (2030, 4050, and 6070 years) underwent one
FIGURE 1. Scheme showing vector locations within the examined
visual eld. The test vectors were located at 16 positions at 0, 15,
20, and 30 eccentricity on the straight and oblique meridians. The
pattern was symmetric with respect to the vertical axis, but the blind
spot area was not tested. A pair of vectors with opposite directions
originated at each location. From the initial point, one target moved
horizontally either to the right or to the left. The order of presentation
of the vectors was randomized. Each vector was tested six times, with
a stimulus luminance of 110 and 41.62 cd/m
2
each.
2634 Becker et al. IOVS, July 2005, Vol. 46, No. 7
examination on the TCC. The mean RT of the 13,637 valid
measurements was 409 ms, with a residual SD of 52 ms.
Because of the limitation of the RTs to the preset interval, 61
(0.44%) were too short (i.e., 170 ms), and 129 (0.93%) were
too long (i.e., 800 ms) and were therefore excluded from the
evaluation.
Age effect had a probability of 0.0004. The least square
means (LSM) were 370 ms (95% CI: 351392 ms) in the 20- to
30-year age group, 425 ms (95% CI: 400453 ms) at 40 to 50
years, and 440 ms (95% CI: 413470 ms) at 60 to 70 years (Fig.
2). The factors luminance, age, code, eccentricity; the variables
nop and nop
2
; and the interactions examination * code, nop *
code, and nop
2
* code had P 0.0001 in this ANCOVA. At the
higher luminance level (110 cd/m
2
vs. 41.62 cd/m
2
) the LSM
RTs were shorter by 18 ms.
However, the model accounted for just R
2
46% of total
variance. The largest part (24%) was explained by interindi-
vidual variation in the average RTthe so-called code effect.
The eccentricity effect was responsible for 5% of the variance.
Approximately 1% of the variance depended on the luminance
levels (110 cd/m
2
vs. 41.62 cd/m
2
) of the target presented and
on age, 0.45% on nop during the experiment, and 0.23% on
nop
2
.
The frequency distribution of 427 RTs of all 36 subjects at
the center of the visual eld at a stimulus luminance of
110 cd/m
2
is provided in Figure 3 (two vectors starting from
the center, one in the temporal and the other in the nasal
direction) as an example.
Experiment Code
Twelve young normal subjects were examined two times on
different days to evaluate the inuence of interindividual dif-
ferences, individual interexamination differences, and the
course of the examination (nop). The RT mean of 9157 valid
measurements equalled 356 ms and the residual SD was 47 ms.
By limiting the obtained RTs to the predened interval (170
800 ms), we declared 30 (0.33%) to be too short and 30
(0.33%) to be too long in this experiment. The model ex-
plained 30% of total variance. The ANCOVA results are shown
in Table 1.
At the lower stimulus luminance level (41.62 cd/m
2
) mean
RT equalled 318 ms (95% CI: 314323) versus 307 ms (95% CI:
301314) at the higher luminance level (110 cd/m
2
). RT in-
creased with eccentricity on average by 2 ms per degree. At
the center, LSM equalled 326 ms (95% CI: 322330 ms), and at
30 eccentricity 386 ms (95% CI: 376397 ms). RT was 8 ms
longer in the nasal hemield than in the temporal (nasal mean,
360 ms, 95% CI: 358362 ms; temporal mean, 352 ms; 95% CI:
350355 ms). RT for stimuli that originated from the vertical
midline was 363 ms (95% CI: 361366 ms). In the lower
hemield, a mean RT of 353 ms (95% CI: 350355 ms) was
observed; in the upper hemield, 367 ms (95% CI: 365369
ms); and at the horizontal midline, 356 ms (95% CI: 354359
ms). The mean of the individual RTs (code) ranged between
313 and 411 ms. In the second examination, RT was on average
10 ms higher than in the rst visit (SE, 1 ms). Repeatability
variance was 0.57% of total variance.
The average RTs in all 12 subjects in both examinations and
the related 95% CIs are shown in Figure 4. Highlighted by this
summary are the interindividual differences in RT, but also
intraindividual ones, that were between 19 and 46 ms.
On average, over all subjects, the predicted RTs increased
from 326 ms for the rst stimulus to 344 ms for the last one.
Average curvature of the fatigue trend indicated that an initial
increase in RT was reversed to a decline at the end of the
examination period. Each subject examined had a separate
course modeled by a linear (nop) plus a quadratic (nop
2
) term.
The regression lines for each subject are presented in Figure 5.
DISCUSSION
Our experiments on normal subjects on the Tubingen Com-
puter Campimeter represent the rst carefully performed trials
undertaken in kinetic perimetry to determine the effects of
age, the subject examined, his or her daily condition, and the
course within one examination period, as well as stimulus
location and luminance on RT. We found them all to have
signicant effects on RT. Our ndings are consistent with
previous similar reports of static perimetry RTs but highlight
some of the key differences in these methodologies and how
various factors alter RT.
According to Finlay,
24
an object is perceived as moving
immediately when it moves with an angular velocity of 1 to
FIGURE 3. Frequency distribution of 427 RTs of all 36 subjects at the
center of the visual eld at a stimulus luminance of 110 cd/m
2
(two
vectors, each tested six times).
FIGURE 2. Mean RT and related 95% CI by age group (age experi-
ment).
IOVS, July 2005, Vol. 46, No. 7 Factors Affecting Perimetric Reaction Time 2635
2 deg/s. Thus, the target velocity in our study (2 deg/s) was
more conservative than that proposed by Johnson et al.
25
,
being 4 deg/s.
Age
In general, our RT data with moving stimuli are in accordance
with several but not all studies regarding similar testing of
healthy subjects. A signicant effect of age on RT was demon-
strated in this study. The mean RT was 55 ms higher in the
group aged 40 to 50 years than in the group aged 20 to 30 years
(370 ms) and 70 ms higher in the group aged 60 to 70 years
than in the group aged 20 to 30 years, with no overlap in the
related condence intervals. The increase in RT between the
group aged 40 to 50 years and the group aged 60 to 70 years
was smaller than between the groups aged 20 to 30 years and
40 to 50 years. Comparable data regarding the inuence of age
on RT in kinetic perimetry have not been reported. Porciatti et
al.
21
used two different age groups (mean ages of 29 and 70
years) to differentiate between the inuences of motoric (re-
sponding to a stimulus) and sensory (seeing the stimulus)
components of delayed RT with age. These investigators mea-
sured RTs for motion onset, as the study required normal
subjects to respond as quickly as possible by button-press
(simple RT) to a visual stimulus. At the lower speed level
(1 deg/s) it was found that the older group had an average
increase of RT of 75 ms compared with the younger group.
According to our data, the difference between the two groups
with comparable ages was similar (70 ms) in kinetic perimetry.
An increase in RT of 9 ms per decade in measurements of static
suprathreshold perimetry in patients has also been described
however, with a weak correlation (Pearson r 0.18; Artes et
al.
23
). This study had a group mean RT (451 ms) that was
slightly longer than ours (409 ms). Only one study (Rouland
and Hache
13
) found no signicant difference in the RT mea-
surements between two age groups studied during static pe-
rimetry with normal subjects (mean RTs approximately 470 ms).
However, in their research, the age groups were closer in age
(2640 and 4169 years).
Stimulus Luminance and Location
The inuence of stimulus luminance on RT in perimetry has
been described in the literature. Cattell
26
found that there is a
shorter RT with a brighter stimulus luminance, whereas
Pieron
7
went further to develop his own law, which describes
a dependency of RT on luminance. A detailed history of the
developments in the inuence of luminance on RT can be
found in Manseld,
6
where RT is described by two compo-
nents: a xed one and one that follows a power function of
TABLE 1. ANCOVA Table: Experiment Code
Source DF Sum of Squares Variance (%) Mean Square F Ratio P
Luminance 1 29.37 1.26 29.37 164.36 0.0001
Eccentricity 1 126.73 5.45 126.73 709.24 0.0001
Side 2 9.82 0.42 4.91 27.48 0.0001
Half 2 22.27 0.96 11.14 62.32 0.0001
Code 11 200.57 8.62 18.23 102.05 0.0001
Examination 1 13.34 0.57 13.34 74.63 0.0001
Examination*Code 11 61.22 2.63 5.57 31.15 0.0001
nop 1 19.55 0.84 19.55 109.41 0.0001
nop*Code 11 15.79 0.68 1.44 8.04 0.0001
nop*nop 1 2.31 0.10 2.31 12.91 0.0003
nop*nop*Code 11 7.91 0.34 0.72 4.03 0.0001
Model 53 700.47 30.10 13.22 73.97
Error 9, 103 1, 626.54 69.90 0.18
Total 9, 156 2, 327.02
DF, degrees of freedom.
FIGURE 4. Mean RT and related 95% CI by examination in 12 subjects
who underwent two examinations on different days (code experi-
ment).
FIGURE 5. Individual regression lines over the course of one exami-
nation in 12 subjects (code experiment).
2636 Becker et al. IOVS, July 2005, Vol. 46, No. 7
light intensity. Because we studied only two different intensi-
ties, we cannot analyze our data by a power function. How-
ever, consistent with previous reports, we found a decrease in
RT with increasing intensity.
In this experiment, the increase of RT per degree eccen-
tricity was 2 ms on average. Wall et al.
22
found an increase of
RT of approximately 1 ms/deg when using suprathreshold
stimuli in static perimetry with increasing visual eld eccen-
tricity, from the center to 50 (374 ms at the center). Other
investigators have reported an increase of approximately 20 ms
between the RT value at the center and at an eccentricity of
30 (average increase of 0.7 ms/deg) for static stimuli pre-
sented with a stimulus-luminance of 14 cd/m
2
on a cathode ray
tube (Ando et al.
27
). Osaka
28
reported an increase of 40 ms
from the center to an eccentricity of 30 (average increase, 1.3
ms/deg) with a signal luminance of 8.5 cd/m
2
with static
stimuli. In addition to that, Osaka described a nasotemporal
difference in RT. In the temporal area of the visual eld RTs
were shorter than in the nasal, as seen in the two tested
subjects. These differences increased from 4.5 ms (at 10
eccentricity) to 11 ms (at 50 eccentricity). In agreement with
these results, we found RT to be, on average, 8 ms longer in the
nasal area than in the temporal part of the visual eld. Osaka
theorized that these differences in RT depend on interhemi-
spheric transmission time. The extent of the inuence of ec-
centricity seems to depend on several different experimental
conditions. It ranges from 1 to 2 ms per degree of eccentricity.
Our data suggest a greater effect of eccentricity on RT in
kinetic situations than in static ones.
Individual Variation
In a previous report
29
we demonstrated that RTs showed
considerable inter- and intraindividual variation with individual
LSMs ranging from 305 to 454 ms. In the current study, we
expand on these observations and present RT data from 12
subjects whose LSMs varied over a range of 98 ms (313411
ms). Remember that we tested only healthy subjects between
20 and 30 years of age. Interindividual RT differences have
been reported by other authors: Artes et al.
23
reported consid-
erable RT differences between subjects examined with su-
prathreshold static perimetry and reported individual means
that differed between 316 and 908 ms. However, these sub-
jects (aged between 12 and 81 years) were patients attending
an optometric practice for routine eye care, potentially having
risk factors for glaucoma or neurologic disease. Thus, patients
may be expected to have a much broader range of interindi-
vidual differences in RT. Frisen
9
reported about the interindi-
vidual differences of mean RT (320660 ms) in 100 subjects
during automated static perimetry on a ring screener, while
Henson and Artes
12
reported large intersubject differences in
the average latency and its dispersion in measurements of
suprathreshold perimetry. Another report about intersubject
variability in RT was published by Breitmayer and Breier
30
(mean RT ranged over 120 ms), describing experiments con-
cerning the inuence of background color on RT to incre-
mentdecrement spot stimuli of varying diameters. The sub-
jects tested were four normal subjects, whose ages ranged
from 20 to 46 years.
The examination effect caused some variance, showing that
in each examination period most subjects had different mean
RTs. The average RT-increase between the rst and second
session was small (10 ms). The interaction examination * code
was also signicant. Some individuals were faster in the rst
examination and others in the second one, by various degrees.
Kobrick and Sleeper
31
examined RT for light signal detection
in 24 male soldiers aged 18 to 35 years on different days. They
repeated the task three times, without nding any signicant
effect between measurements on different days under the
same conditions. On the contrary, Ando et al.
27
reported sig-
nicant training success over a period of several days. How-
ever, the stimuli in this study were presented at only three
different locations in the visual eld to subjects having prac-
ticed 45 times. As the mean RT of our second examination was
greater than in the rst, there seems to be no learning effect
regarding RT. This lack of a learning effect may depend on the
low number of task repetitions.
Intraexamination Course
In our measurements we found a trend over the course of one
examination (nop). This effect is similar to short-term uctua-
tion, the scatter of the differential luminance threshold ob-
served during a single visual eld test in static perimetry. For
nop, 1 was the rst stimulus presented, 2 was the second, up
until 480 for the last stimulus. To obtain an optimal t of the
model, we used a linear and a quadratic term in the prediction
model. The coefcient of the common linear effect suggested
that there is a general fatigue effect. The predicted RT in-
creased from 326 ms for the rst stimulus to 344 ms for the last
one. Common curvature was such that the average fatigue
curve would increase at rst. However, it would then decline
at the end of the examination period.
To analyze these courses further, each individual was to
have his or her own course with a linear and quadratic part
(nop * code, nop
2
* code). Because of this, a particular mix of
fatigue and learning effect was possible for each individual. It
is notable that in 11 of 12 participants, a fatigue effect was
detected. In one, a slight shortening of RT appeared, which
could be interpreted as a learning effect. In this single partic-
ipant, a fatigue effect also appeared in the rst half of the
examination that was compensated for in the second part.
Marra and Flammer
32
detected neither a learning nor a fatigue
effect during a perimetric session using threshold determina-
tions in static perimetry common to healthy persons and pa-
tients with glaucoma or cataract. To date, there have been no
comparable studies published concerning the interpretation of
RT measurements over the course of one examination.
Nevertheless, it should be mentioned again that we tested
only healthy subjects and that these results are applicable only
to this group. For common diseases such as glaucoma or other
optic neuropathies, as well as neuroophthalmic diseases, fur-
ther investigation is necessary. It is also especially important to
know whether individual RT varies when approaching the
scotoma border.
CONCLUSION
We demonstrated that RT during automated kinetic perimetry
varies considerably, even in normal subjects. Our experiments
show that well-known factors such as location in the visual
eld and stimulus luminance are relevant, but that age, and
individual inter- and intrasubject variability are also important
factors.
Precise RT predictions can be made when presenting only
a few (approximately 10) so-called RT vectors, which can be
easily integrated in modern computer-controlled perimetric
test procedures. By use of this method, one can assess and
reduce the systematic shift of scotoma borders measured with
moving stimuli and more accurately determine the actual
threshold.
References
1. Payne WH. Reaction time as a function of retinal location. Vision
Res. 1966;6:729732.
IOVS, July 2005, Vol. 46, No. 7 Factors Affecting Perimetric Reaction Time 2637
2. Poffenberger ATJ. Reaction time to retinal stimulation with special
reference to the time lost in conduction through nerve centers.
Arch Psychol. 1912;23:173.
3. Haines RF. Peripheral visual response time and retinal luminance-
area relations. Am J Optom Physiol Opt. 1975;52:8596.
4. Rains JD. Signal luminance and position effects in human reaction
time. Vision Res. 1963;3:239251.
5. Bartlett NR, Macleod S. Effect of ash and eld luminance upon
human reaction time. J Opt Soc Am. 1954;44:306311.
6. Manseld RJ. Latency functions in human vision. Vision Res. 1973;
13:22192234.
7. Pieron H. Recherches sur les lois de variation des temps de latence
sensorielle en function des intensites excitatrices. Ann Psychol.
1914;20:1796.
8. Ulrich R, Rinkenauer G, Miller J. Effects of stimulus duration and
intensity on simple reaction time and response force. J Exp Psy-
chol Hum Percept Perform. 1998;24:915928.
9. Frisen L. Computerized perimetry: possibilities for individual ad-
aptation and feedback. Doc Ophthalmol. 1988;69:39.
10. Bengtsson B, Olsson J, Heijl A, Rootzen H. A new generation of
algorithms for computerized threshold perimetry, SITA. Acta Oph-
thalmol Scand. 1997;75:368375.
11. Olsson J, Bengtsson B, Heijl A, Rootzen H. An improved method to
estimate frequency of false positive answers in computerized pe-
rimetry. Acta Ophthalmol Scand. 1997;75:181183.
12. Henson DB, Artes PH. New developments in supra-threshold pe-
rimetry. Ophthalmic Physiol Opt. 2002;22:463468.
13. Rouland JF, Hache JC. Variations du temps normal de reponse du
patient pendant la mesure du champ visuel. Ophthalmologica.
1991;202:4852.
14. Wall M, Maw RJ, Stanek KE, Chauhan BC. The psychometric
function and reaction times of automated perimetry in normal and
abnormal areas of the visual eld in patients with glaucoma. Invest
Ophthalmol Vis Sci. 1996;37:878885.
15. Eischer-Desrivieres J. La perception des objets en mouvement a ` la
peripherie du visual: determinations quantitatives. Bull Soc Oph-
thalmol Paris. 1937;46.
16. Schiefer U, Stercken-Sorrenti G, Dietrich TJ, Friedrich M, Benda N.
Fundus-orientierte Perimetrie: Evaluation eines neuen Gesichts-
feld-Untersuchungsverfahrens bezuglich der Detektion von An-
gioskotomen. Klin Monatsbl Augenheilkd. 1996;209:6271.
17. Schiefer U, Selig B, Dietrich TJ. Automated static campimetry with
locally enhanced spatial resolution. In: Wall M, Wild JM, eds.
Perimetry Update 1998/99. The Hague, The Netherlands: Kugler
Publications; 1999:261272.
18. Schiefer U, Stercken-Sorrenti G, Dietrich TJ, Friedrich M, Benda N.
Fundus oriented perimetry: a new concept for increasing the
efciency of visual eld examination. In: Wall M, Heijl A, eds.
Perimetry Update 1996/97. New York: Kugler; 1997:107109.
19. Dietrich TJ, Selig B, Friedrich M, Benda N, Schiefer U. Calibration
routines for video display units for perimetric examinations (Ab-
stract). Ger J Ophthalmol. 1996;5(suppl):125.
20. Dietrich TJ, Friedrich M, Selig B, Benda N, Schiefer U. Application
of video display units for campimetric purposes: luminance char-
acteristics and calibration procedures. In: Wall M, Heijl A, eds.
Perimetry Update 1996/97. New York: Kugler; 1997:471.
21. Porciatti V, Fiorentini A, Morrone MC, Burr DC. The effects of
ageing on reaction times to motion onset. Vision Res. 1999;39:
21572164.
22. Wall M, Kutzko KE, Chauhan BC. The relationship of visual thresh-
old and reaction time to visual eld eccentricity with conventional
automated perimetry. Vision Res. 2002;42:781787.
23. Artes PH, McLeod D, Henson DB. Response time as a discriminator
between true- and false-positive responses in suprathreshold pe-
rimetry. Invest Ophthalmol Vis Sci. 2002;43:129132.
24. Finlay D. Motion perception in the peripheral visual eld. Percep-
tion. 1982;11:457462.
25. Johnson CA, Keltner JL. Optimal rates of movement for kinetic
perimetry. Arch Ophthalmol. 1987;105:7375.
26. Cattell JM. The inuence of the intensity of the stimulus on the
length of the reaction time. Brain. 1886;9:512515.
27. Ando S, Kida N, Oda S. Practice effects on reaction time for
peripheral and central visual elds. Percept Mot Skills. 2002;95:
747751.
28. Osaka N. Naso-temporal differences in human reaction time in the
peripheral visual eld. Neuropsychologia. 1978;16:299303.
29. Schiefer U, Strasburger H, Becker ST, et al. Reaction time in
automated kinetic perimetry: effects of stimulus luminance, eccen-
tricity, and movement direction. Vision Res. 2001;41:21572164.
30. Breitmeyer BG, Breier JI. Effects of background color on reaction
time to stimuli varying in size and contrast: inferences about
human M channels. Vision Res. 1994;34:10391045.
31. Kobrick JL, Sleeper LA. Effect of wearing chemical protective
clothing in the heat on signal detection over the visual eld. Aviat
Space Environ Med. 1986;57:144148.
32. Marra G, Flammer J. The learning and fatigue effect in automated
perimetry. Graefes Arch Clin Exp Ophthalmol. 1991;229:501
504.
2638 Becker et al. IOVS, July 2005, Vol. 46, No. 7