Full text of DOI:10.1093/geronb/57.3.P241

Journal of Gerontology: PSYCHOLOGICAL SCIENCES
Copyright 2002 by The Gerontological Society of America
2002, Vol. 57B, No. 3, P241–P245
Aging and Bilateral Symmetry Detection
Andrew M. Herbert,1 Olga Overbury,2 Jason Singh,2 and Jocelyn Faubert3
1Department of Psychology, University of North Texas, Denton.
2Sir Mortimer B. Davis Jewish General Hospital, McGill University, Montreal, Canada.
3School of Optometry, University of Montreal, Canada.
The salience of bilateral symmetry varies as a function of the orientation of the symmetry axis. Vertical symmetry
is most salient, followed by horizontal and then oblique orientations. We tested symmetry detection in different
age groups to determine whether performance of this intermediate-level visual task is affected by normal, non-
pathological aging. We tested forty participants and analyzed the results with respect to age group and symmetry
orientation (vertical, horizontal, and 45 degree oblique). There was a vertical symmetry detection advantage for
all participants, where sensitivity was highest for vertical symmetry, followed by horizontal symmetry, and then
the oblique symmetry. Sensitivity to symmetry did not differ for the two younger age groups (aged 19–39 and 40–
60), but declined significantly for the group aged 61–70, and declined again for the oldest group aged 71–80. This
age-related difference in sensitivity to symmetry was not reflected in a measure of bias, where there were no dif-
ferences as a function of age.
HERE are clear age-related declines in visual acuity,
contrast sensitivity, temporal-frequency sensitivity, vi-
cause of changes in the efficiency of low-level visual func-
tions and because of age-related changes in higher cortical
functions. Thus, changes in the salience of symmetry with
increasing age could provide an index of how long-range vi-
sual connections change during the life span.
T
sual fields, and motion sensitivity during normal (nonpatho-
logical) aging (Crassini, Brown, & Bowman, 1988; Owsley
& Sloane, 1990; Rubin et al., 1997; Spear, 1993). In some
cases the changes in lower level visual abilities correlate
with age-related deficits in higher level tasks, such as the
link observed between poor face recognition and declines in
contrast sensitivity (Owsley, Sekuler, & Boldt, 1981; Ows-
ley & Sloane, 1987). In other cases, visual abilities relying
on the integrity of the visual system are unaffected by aging,
such as in the case of Vernier acuity thresholds (Lakshmi-
narayanan & Enoch, 1995). Thus, deficits in one aspect of
vision in elders do not always translate to predictable de-
clines in another area, and we do not understand why this
occurs (Hoyer & Plude, 1982; Walsh, 1988).
One possible cause of age-related declines in visual func-
tioning is that processes requiring multiple stages become
less efficient with age, whereas those processes relying only
on the output of computations completed in the primary vi-
sual cortex are unaffected. That is, as more synaptic connec-
tions are involved in a sequence of computations required to
complete any process, there is a greater chance of an error
occurring in that sequence (Habak & Faubert, 2000; Hoyer
& Plude, 1982; Walsh, 1988). Thus, processes such as ob-
ject and face recognition show age-related declines, whereas
other visual functions that could be completed in fewer
stages, such as static Vernier acuity, are less affected by
aging (Lakshminarayanan, Aziz, & Enoch, 1992; Lakshmi-
narayanan & Enoch, 1995; Odom, Vasquez, Schwartz, &
Lindbergh, 1989; Whitaker, Elliott, & McVeigh, 1992).
Bilateral symmetry detection is an example of an inter-
mediate level visual task where the structure can be detected
by the filtering of the outputs of cells in the striate cortex
(e.g., Dakin & Watt, 1994; Gurnsey, Herbert, & Kenemy,
1998; Osorio, 1996; Rainville & Kingdom, 2000). One
might expect symmetry detection to decline with age be-
Bilateral symmetry (where one half of a pattern is a mir-
ror reflection of the other half) salience varies as a function
of the orientation of the symmetry axis. Vertical symmetry
is most salient, followed by horizontal and then oblique ori-
entations (e.g., Herbert & Humphrey, 1996; Wagemans,
1996). There have been various accounts for this pattern of
orientation differences, resting either on the idea that there
is a hard-wired preference for vertical symmetry (based on
ideas initially proposed by Mach, 1906, and elaborated on
by Braitenberg, 1990, and Julesz, 1971). Bilateral symmetry
may be detectable based on processes in the primary visual
cortex (see, for example, Dakin & Herbert, 1998; Gurnsey
et al., 1998; Rainville & Kingdom, 2000), in which case it
should be resistant to age-related declines. In contrast, some
research on symmetry detection has indicated that top-down
influences occur, in that attention and priming influence the
salience of symmetry at different orientations (Pashler,
1990; Wenderoth, 1994; Wenderoth & Welsh, 1998). It has
been suggested that higher-level factors, such as the ob-
server’s frame of reference, affect the selection of a possible
symmetry axis (Palmer & Hemenway, 1978; Rock & Lea-
man, 1963). If such a top-down influence exists, then age-
related declines in symmetry detection, associated with
changes in higher-level cortical function that result from
aging, may occur. To date, there has been no research on
symmetry detection in aging.
METHODS
Participants
We recruited forty participants from staff and patients at
the Ophthalmology department of the Sir Mortimer B.
P241

P242
HERBERT ET AL.
Davis Jewish General Hospital in Montreal, Canada. The re-
search followed the tenets of the Declaration of Helsinki.
Participants gave informed consent after having the experi-
mental procedure explained to them, and the research was
approved by the review boards of the Hospital and Univer-
sity of Montreal. Minimum acceptable corrected acuity
(monocular) was 20/30 (1.5 LogMAR). We took the visual
acuities from the participants’ charts, with their permission.
These values represent their best corrected far acuities, and
all participants were tested with their best eye and using any
needed corrective lens. We tested the acuity of each partici-
pant using Snellen and LogMAR charts in the Ophthal-
mology department of the Sir Mortimer B. Davis Jewish
General Hospital. No participants reported any visual or
neurological problems. We excluded potential volunteers if
they had a history of diabetic retinopathy, glaucoma, or
macular degeneration. All participants completed the testing
in one visit to the laboratory based in the ophthalmology
clinic at the Hospital.
After testing, we divided participants into different groups
based on their age: young adults (19–39 years, n ϭ 10,
mean LogMAR acuity ϭ 1.13, s ϭ 0.21); middle-aged (40–
60 years, n ϭ 10, mean LogMAR acuity ϭ 1.1, s ϭ 0.21);
young-old (61–70 years, n ϭ 8, mean LogMAR acuity ϭ 1.34,
s ϭ 0.19); and old-old (71–80 years, n ϭ 9, mean LogMAR
acuity ϭ 1.33, s ϭ 0.22). We did not analyze the results
from 3 other participants because performance was at
chance for all symmetry orientations (1 in the middle-aged
group, and 2 in the old-old group).
Figure 1. (A) Reverse contrast versions of the kinds of patterns
presented to subjects. Note that the dots were circular blobs when
presented on the cathode ray tube screen. (B) Mean dЈ and c for each
age group plotted as a function of the orientation of the symmetry
axis. Error bars indicate the standard error of the mean for dЈ.
Apparatus and Stimuli
We conducted testing using a Macintosh 7200 and Apple
high-resolution monitor (Apple Computer, Cupertino, CA).
Data collection and stimulus presentation were controlled
by Pixx, an application developed by the Vision Laboratory
at Concordia University, Montreal, Canada. The viewing
distance (114 cm) was maintained using a combination
chin/head rest. A fixation cue was presented on the screen
between each trial.
screen as small circular white blobs rather than as square
points as in the figure.
Procedure
Testing was monocular using the participant’s preferred
eye. We presented a total of 60 asymmetric and 60 symmet-
ric patterns in each of three blocks of trials. We tested each
symmetry orientation on a different block of trials. On each
trial participants had to respond yes or no that the pattern
was symmetric. The experimenter (Jason Singh) entered
each response on the keyboard. We adopted this procedure
because pilot testing indicated that some of the older partic-
ipants found it difficult to look at the screen, make their
judgment, find the right key to press, then return to the task.
Trials were self-paced, so we initiated each trial when the
participant was ready to respond and was fixating upon the
fixation dot placed in the center of the screen between trials.
There was no feedback as to whether or not the response
was correct.
Each block of experimental trials was preceded by a prac-
tice block of 10 trials (5 asymmetric and 5 symmetric, or-
dered randomly) to familiarize the participants with the task
and with detecting symmetry at each particular orientation.
We tested each symmetry orientation in a separate block of
trials, and participants knew which orientation we were as-
sessing. We also tested vertical symmetry first and randomly
selected the subsequent orientations. We did this because
naive participants associate vertical symmetry with the con-
The stimuli consisted of asymmetric and symmetric
dot patterns presented for 40 frames (0.533 seconds at
the refresh rate of 75 Hz) centered on the high-resolution
monitor. The patterns were composed of 100 white dots
(80 cd/m2) placed within an invisible circular region 2
degrees in diameter. The monitor screen was dark aside
from the patterns, and the contrast was 99% (background
screen luminance of 0.5 cd/m2). The computer generated
all patterns at random. The dots could appear anywhere
within the circular region. Patterns were symmetric about
an axis oriented vertically, horizontally, or at 45 degrees
clockwise from vertical (only this one oblique orientation
was tested, both to limit the number of trials to be com-
pleted in the test session and because there are no indica-
tions of differential performance for the left and right ob-
lique orientations in the literature, Herbert & Humphrey,
1996). On symmetric trials, a random pattern of 50 dots
was generated and reflected about the axis specified for
that block of trials. Schematic versions of a vertically
symmetric and an asymmetric pattern are provided in
Figure 1A. Note that the dots appeared on the monitor

SYMMETRY DETECTION IN AGING
P243
cept of bilateral symmetry, so this made explaining the task
easier. Each sequence of trials was randomly generated
using the presentation software, with the only proviso being
that no more than six symmetric or asymmetric patterns
could appear in a row.
of symmetry orientation, F(2,66) ϭ 6.3, p Ͻ .01, reflected
in values of c closer to zero for vertical symmetry compared
with the other two orientations (which were not signifi-
cantly different from one another). This analysis revealed
that participants were more likely to identify asymmetric
patterns as symmetric when looking for nonvertical symme-
try orientations.
RESULTS
The symmetry detection performance for each age group
is presented in Figure 1B, where dЈ and c are plotted for
each symmetry orientation tested. These two statistics give a
measure of sensitivity and bias, respectively. Both are com-
puted from z scores for “Symmetric” responses made by
participants. When an individual said “Symmetric” and was
correct, the response was defined as a hit. When the re-
sponse “Symmetric” was given to an asymmetric pattern, it
was termed a false alarm. Minimizing false alarms and
maximizing hits results in a high dЈ value. Higher dЈ values
indicate greater sensitivity. A dЈ equal to 1 is equivalent to
an overall proportion of correct responses of 70%. The for-
mula for c is analogous to that for dЈ, but rather than taking
the difference between the proportions of hits and false
alarms, it is computed by taking their sum multiplied by
Ϫ0.5 (Macmillan & Creelman, 1991). If the hit rate is high
and false alarm rate low, c results in a score near 0. Scores
less than 0 indicate a preponderance of “Symmetric” re-
sponses regardless of trial type (lots of hits and false
alarms), and positive values of c occur if there are more
“Asymmetric” responses regardless of stimulus type (few
hits and few false alarms). In some cases perfect perfor-
mance was obtained, where the z scores used in calculating
dЈ would be infinite. An arbitrary z score of Ϯ2.5 was as-
signed when no misses or false alarms were made so that a
dЈ could be calculated. Thus, the maximum possible dЈ was 5.
We analyzed the results using a split-plot analysis of vari-
ance with Age Group as a between participants factor (4 lev-
els) and Symmetry Orientation as a within-subjects factor (3
levels). For dЈ, there was no significant Age ϫ Symmetry
Orientation interaction, F(6,66) ϭ 1.6, p Ͼ .1. However,
there was a significant main effect for age group, F(3,33) ϭ
11.2, p Ͻ .01, and a significant main effect for symmetry
orientation, F(2,66) ϭ 42.3, p Ͻ .01. These results confirm
the pattern of results illustrated in Figure 1B. Sensitivity did
not differ between the young and middle-aged groups, but
all other age-group comparisons were significant (p Ͻ .05,
Newman-Keuls test). The old-old group had the lowest sen-
sitivity, the young and middle-aged groups the highest
sensitivity, with the young-old group falling between the
others. For each group the mean dЈ differed significantly for
each symmetry orientation (p Ͻ .05). Performance was best
for vertical symmetry, followed by horizontal symmetry.
Oblique symmetry was most difficult to detect. As reported
in the Methods section, average LogMAR acuities increased
with age group, but the range of acuities did not differ with
age. We reanalyzed the dЈ data with acuity as a covariate.
This analysis showed no significant influence of acuity on dЈ
(F Ͻ 1), but the significant effect of age remained, F(3,33) ϭ
19.8, p Ͻ .01.
DISCUSSION
The observed symmetry orientation difference replicates
the results of many other studies, and the sequence of orien-
tations was the same for all age groups (sensitivity for Verti-
cal greater than Horizontal, greater than Oblique). If prac-
tice affected symmetry detection, the sequencing of trial
blocks we used would have minimized the chances of ob-
taining the vertical advantage. The fact that this advantage
was obtained despite testing Vertical first is another indica-
tion of the strength of that effect. Whatever the source, this
orientation anisotropy does not vary across different age
groups. The results demonstrate a decline in the overall abil-
ity of participants older than age 60 to see symmetry. How-
ever, it must be stressed that although the change in dЈ for
the older groups was significant, it was relatively small. The
young-old and old-old participants could still detect sym-
metry at dЈ consistent with 70% accuracy.
Although we used participants’ far acuities in determin-
ing which eye to test and then tested at a viewing distance
of 114 cm, we believe acuity is not a significant factor for
the following reasons: (a) we asked all participants if they
could see the individual dots, and they reported they
could; (b) The fixation cue presented between trials was
the same size as an individual element composing the pat-
terns, and this dot could be seen by all participants; and,
(c) There is evidence that there is little change in the abil-
ity to detect symmetry in spatially filtered patterns and
blurred patterns (Barlow, 1980; Dakin & Herbert, 1998).
Thus, we did not find it surprising that the minor decrease
in average acuity for the older participants did not covary
with symmetry detection accuracy in this study. Age-
related declines in contrast thresholds should have had no
effect on the results because the individual dots and the
patterns as a whole were suprathreshold. We maximized
the contrast by presenting white dots on a black back-
ground, and all participants reported being able to resolve
the individual dots and the patterns. Although we screened
participants using their far acuities and tested at an inter-
mediate distance, the absence of significant influence of
acuity on dЈ argues against differences in acuity account-
ing for the age-group differences. The sequence of orien-
tation differences did not change across the age groups,
arguing against changes in the frame of reference with
age. The pattern of these results is consistent with a gen-
eral decline in performance associated with nonpathologi-
cal aging. Although low-level explanations cannot be
ruled out (the components of the dot patterns should be
very salient to all participants, but this was not measured
explicitly), we speculate that the general decline in perfor-
mance is consistent with some unspecified changes in in-
termediate and higher level visual functions with aging.
The deficit in symmetry detection for both groups of older
We repeated the analyses for c. There was neither a main
effect of age (F Ͻ 1) nor an interaction between age and
symmetry orientation (F Ͻ 1). There was a significant effect

P244
HERBERT ET AL.
Address correspondence to Andrew M. Herbert, PhD, University of
North Texas, Department of Psychology, P.O. Box 311280, Denton, TX
76203-1280. E-mail: herberta@unt.edu
participants suggests they have more difficulty matching
elements across the symmetry axis. Such comparisons re-
quire an intact neural network spanning a few degrees of
visual angle, which correlates to a large area in the visual
cortex with interconnections between areas separated by
relatively large cortical distances.
References
Appelle, S. (1972). Perception and discrimination as a function of stimulus
orientation: The “oblique effect” in man and animals. Psychological
Bulletin, 78, 266–278.
It is noteworthy that there was no difference in symmetry
detection for participants younger than age 60. In combina-
tion with evidence that low-level visual abilities are pre-
served in aging, the observed decline in symmetry detection
performance after age 60 suggests that observed difficulties
in higher level tasks may result from problems at all stages
of visual processing that accumulate with the number of
steps required to complete the visual computations (Habak
& Faubert, 2000; Hoyer & Plude, 1982; Walsh, 1988). A
diffuse neural loss occurring during nonpathological aging
would be expected to have a greater effect on higher level
tasks than on lower level tasks based on the increasing num-
bers of neurons and synapses involved in more complicated
processing (Spear, 1993; Weale, 1982). Habak and Faubert
(2000) showed that the perception of simple motion stimuli
(moving gratings) was similar in young and elderly groups,
but complex motion perception was worse in the older
group (where moving dot textures are presented). Complex
motion perception requires more processing steps than are
required for perceiving simple motion, and Habak and
Faubert argue that this is why the difference as a function of
age is primarily seen for certain types of moving stimuli. We
speculate that the decline in performance observed in the
present study reflects the fact that symmetry detection relies
on processes occurring after visual information arrives in
the primary visual cortex (Dakin & Herbert, 1998; Gurnsey
et al., 1998). This conclusion is speculative and requires fur-
ther study, because it has been demonstrated that older
adults use binocular cues just as well as younger subjects
when detecting masked Gabor patterns (Speranza, Mora-
glia, & Schneider, 1995). That finding is consistent with a
conclusion that more central processes related to binocular
vision must be intact in the older population. There were
some minor differences between the older and younger par-
ticipants in Speranza and colleagues’ study, which is consis-
tent with our finding that older participants could still detect
symmetry, just not as well as younger participants.
Barlow, H. B. (1980). The absolute efficiency of perceptual decisions.
Philosophical Transactions of the Royal Society of London, B290, 71–
82.
Braitenberg, V. (1990). Reading the structure of brains. Network, 1, 1–11.
Crassini, B., Brown, B., & Bowman, K. (1988). Age-related changes in
contrast sensitivity in central and peripheral retina. Perception, 17,
315–332.
Dakin, S. C., & Herbert, A. M. (1998). The spatial region of integration for
visual symmetry detection. Proceedings of the Royal Society of Lon-
don, B26, 659–664.
Dakin, S. C., & Watt, R. J. (1994). Detection of bilateral symmetry using
spatial filters. Spatial Vision, 8, 393–413.
Essock, E. A. (1980). The oblique effect of stimulus identification consid-
ered with respect to two classes of oblique effects. Perception, 9, 37–
46.
Gurnsey, R., Herbert, A. M., & Kenemy, J. (1998). Bilateral symmetry em-
bedded in noise is detected accurately only at fixation. Vision Research,
38, 3795–3803.
Habak, C., & Faubert, J. (2000). Larger effect of aging on the perception of
higher-order stimuli. Vision Research, 40, 943–950.
Herbert, A. M., & Humphrey, G. K. (1996). Bilateral symmetry detection:
Examining a “callosal” hypothesis. Perception, 2, 463–480.
Hoyer, W. J., & Plude, D. J. (1982). Aging and the allocation of attentional
resources in visual information-processing. In R. Sekuler, D. Kline, &
K. Dismukes, (Eds.), Aging and human visual function (pp. 245–263).
New York: A. R. Liss.
Julesz, B. (1971). Foundations of cyclopean perception. Chicago, IL: Uni-
versity of Chicago Press.
Lakshminarayanan, V., Aziz, S., & Enoch, J. M. (1992). Variation of the
hyperacuity gap function with age. Optometry and Vision Science, 69,
423–426.
Lakshminarayanan, V., & Enoch, J. M. (1995). Vernier acuity and aging.
International Ophthalmology, 19, 109–115.
Mach, E. (1906). The analysis of sensations and the relation of the physical
to the psychical. New York: Dover.
Macmillan, N. A., & Creelman, C. D. (1991). Detection theory: A user’s
guide. New York: Cambridge University Press.
Odom, J. V., Vasquez, R. J., Schwartz, T. L., & Lindbergh, J. V. (1989).
Adult Vernier thresholds do not increase with age; Vernier bias does.
Investigative Ophthalmology & Visual Science, 30, 1004–1008.
Osorio, D. (1996). Symmetry detection by categorization of spatial phase,
a model. Proceedings of the Royal Society of London, B263, 105–110.
Owsley, C., Sekuler, R., & Boldt, C. (1981). Aging and low-contrast vision:
Face perception. Investigative Ophthalmology & Visual Science, 21,
362–365.
Owsley, C., & Sloane, M. E. (1987). Contrast sensitivity, acuity, and the
perception of “real-world” targets. British Journal of Ophthalmology,
71, 791–796.
Owsley, C., & Sloane, M. E. (1990). Vision and aging. In F. Boller & R. J.
Grafman (Eds.), Handbook of neuropsychology (Vol. 4, pp. 229–249).
Amsterdam: Elsevier.
Palmer, S. E., & Hemenway, K. (1978). Orientation and symmetry: Effects
of multiple, rotational, and near symmetries. Journal of Experimental
Psychology: Human Perception and Performance, 4, 691–702.
Pashler, H. (1990). Coordinate frame for symmetry detection and object
recognition. Journal of Experimental Psychology: Human Perception
and Performance, 16, 150–163.
Rainville, S. J. M., & Kingdom, F. A. A. (2000). The functional role of ori-
ented spatial filters in the perception of mirror symmetry—Psycho-
physics and modelling. Vision Research, 40, 2621–2644.
Rock, I., & Leaman, R. (1963). An experimental analysis of visual symme-
try. Acta Psychologica, 21, 171–183.
In conclusion, the salience of symmetry as a function of
orientation does not change, because this is most likely a di-
rect result of early visual processing, where different orien-
tations are treated differentially by cortical neurons (Ap-
pelle, 1972; Essock, 1980). The differences in symmetry
detection across age groups observed in the present study
are consistent with normal, but less effective, operation of
whatever neural circuitry is involved in symmetry detection.
Acknowledgments
This research was supported by Medical Research Council of Canada
Grant MT-14777 to Jocelyn Faubert. We thank the observers and staff at Sir
Mortimer B. Davis Jewish General Hospital for their time and patience. Pre-
liminary results were presented at the 1999 annual meeting of the Associa-
tion for Research in Vision and Ophthalmology, Ft. Lauderdale, FL. Many
thanks to Dr. Bert Hayslip and three anonymous reviewers for their com-
ments on the manuscript. The ANCOVA was run in response to comments
from the reviewers.
Rubin, G. S., West, S. K., Munoz, B., Bandeen-Roche, K., Zeger, S.,
Schein, O., et al. (1997). A comprehensive assessment of visual impair-

SYMMETRY DETECTION IN AGING
P245
ment in a population of older Americans. Investigative Ophthalmology
& Visual Science, 38, 557–568.
Sekuler, D. Kline, & K. Dismukes (Eds.), Aging and human visual
function (pp. 161–171) New York: A. R. Liss.
Wenderoth, P. (1994). The salience of vertical symmetry. Perception, 23,
221–236.
Spear, P. D. (1993). Neural bases of visual deficits during aging. Vision Re-
search, 33, 2589–2609.
Wenderoth, P., & Welsh, S. (1998). The effects of cuing on the detection of
bilateral symmetry. Quarterly Journal of Experimental Psychology,
A51, 883–903.
Whitaker, D., Elliott, D., & McVeigh, D. (1992). Variations on hyperacuity
performance with age. Ophthalmic & Physiological Optics, 12, 29–32.
Speranza, F., Moraglia, G., & Schneider, B. A. (1995). Age-related changes
in binocular vision: Detection of noise-masked targets in young and old
observers. Journal of Gerontology: Psychological Sciences, 50B,
P114–P123
Wagemans, J. (1996). Detection of visual symmetries. In C. W. Tyler (Ed.),
Human symmetry perception (pp. 9–32). Utrecht, The Netherlands:
VSP Publishers.
Walsh, D. A. (1988). Aging and visual information processing: Potential
implications for everyday seeing. Journal of the American Optometric
Association, 59, 301–306.
Received June 27, 2000
Accepted July 9, 2001
Weale, R. A. (1982). Senile ocular changes, cell death, and vision. In R.
Decision Editor: Margie E. Lachman, PhD