Dr Catherine Suttle School of Optometry and Vision Science University of New South Wales Sydney NSW 2052 AUSTRALIA E-mail: firstname.lastname@example.org
Background: Ocular dominance is the superiority or preference of one eye over the other in terms of sighting, sensory function (for example, visual acuity) or persistence in binocular rivalry. There is poor agreement between sighting and sensory dominance and findings are equivocal on the possible neural basis of ocular dominance and its significance. Thus, there are questions on the meaning and importance of ocular dominance. Despite the lack of clarity in this area, ocular dominance is used clinically, for example, as the basis for decisions on monovision in contact lens wear and on treatment of anomalies of binocular vision.
Methods: Sighting dominance and three types of sensory dominance (based on resolution acuity, contrast sensitivity and alignment sensitivity) were compared within individuals, with the main aim of determining whether sensory dominance is consistent across spatial visual functions.
Results: Our findings indicate that each type of sensory dominance is insignificant in most individuals and in agreement with previous work that sensory and sighting dominance do not generally agree.
Conclusion: These results demonstrate not only that different types of ocular dominance are not in agreement but also that in the normal visual system sensory dominance as measured here is insignificant in most individuals with normal vision.
Most people tend to use one side of their body more easily, more frequently and with better dexterity than the other side and this is often characterised by left- or right-handedness, or by the dominant use of a left or right foot. It has been demonstrated that the ocular system is not an exception and that an individual will usually prefer to use one eye rather than the other for certain tasks.1 This tendency is known as ocular dominance and may be defined on the basis of sighting (the eye used for viewing, for example, through a telescope or camera eyepiece), sensory function (such as the eye with better visual acuity) or persistence in a binocular rivalry situation (the more persistent image perceived in a dichoptic presentation). There are numerous definitions of ocular dominance and dominance may be determined clinically and in empirical studies in many different ways. Walls2 listed 25 different types of ocular dominance in five categories and discussed the possible meaning and significance of each.
Ocular dominance is not correlated with other types of lateral body dominance, such as handedness,3,4 which is not surprising, as lateral body dominance may be related to dominance of one cerebral hemisphere, while right and left ocular signals are similarly represented in both hemispheres, due to semi-decussation of the visual pathway. Porac and Coren4 reported a significant correlation between dominance of hand and eye in males but not females and suggested this gender difference might be related to environmental pressures for the male group.
While it has been suggested that each individual has an eye that is dominant for both sighting and sensory tasks, it has been demonstrated that the dominant eye for sighting is not necessarily the dominant eye in a sensory task, so the two types of dominance are not in good agreement.2,5 Studies reporting a lack of correlation between sighting and acuity dominance have used distance acuity measures, while near acuity measures may correlate better with sighting dominance.6 This difference between near and distance testing may be related to the fact that physiological diplopia is more apparent at near than at distance, so ocular dominance may be more related to near than to distant vision.
In attempts to understand the neural basis of ocular dominance, imaging studies have shown interocular differences in signal amplitude elicited in response to monocular viewing of patterned stimuli.7 A difference of this kind has also been found to correlate with sensory dominance,8 defined as the eye with better visual acuity. These findings suggest that the differences in sighting or sensory function that define ocular dominance are underpinned by greater cortical activity in response to the dominant eye. It has also been demonstrated that eye dominance may switch between the two eyes, depending on the image size on each retina9 and this seems to suggest that ocular dominance may reflect a stronger retinal signal due to image differences, rather than a fundamentally stronger neural representation of one eye further along the visual pathway. In a critical review of the phenomenon of ocular dominance, Mapp, Ono and Barbeito10 point out that ocular dominance does not correlate with hand-dominance, and sighting and sensory dominance do not agree, and raise questions on the meaning of ocular dominance and its significance, if any, to visual function. These questions remain unanswered and the issue of poor intra-individual agreement between ocular dominance tests has been recognised for some time.11
As outlined above, the sighting and sensory dominant eye may be different in the same individual, suggesting that in the normal visual system a different eye may be preferred or may be superior on different visual tasks. It is not known whether the same is true within the category of sensory ocular dominance. If so, within an individual, some visual functions (for example, letter acuity) may be better served by one eye, and others (for example, vernier acuity) by the fellow eye. Functional distinctions of this kind between the two eyes seem feasible, as the visual systems underpinning different visual functions show varying degrees of overlap and separation. For functions with quite separate underpinning neural bases, ocular dominance may be independent, with one eye being preferred for one task and the fellow eye for a different task. For example, resolution acuity is underpinned by neurons with high spatial resolving ability, while contrast sensitivity at low to moderate spatial frequency is likely to depend on a different set of neurons.
Despite unanswered questions around the concept, measures of ocular dominance form the basis for a number of clinical decisions made by ophthalmic practitioners, such as the choice of eye for distance viewing in monovision contact lens wear. McMonnies12 considered seven criteria on which decisions on the choice of eye for distance and near vision may be based in monovision lens prescribing (for example, prescribing the distance lens for the eye with better acuity). In an extensive review of monovision, Evans13 points to the disagreement between tests of ocular dominance and discusses the difficulties of selecting the eye to correct for distance vision in monovision. Thus, from a clinical point of view and to better understand the issues surrounding ocular dominance, it is important to clarify whether ocular dominance is consistent within individuals, in terms of agreement in dominance for different sensory functions.
The aim of the present study was to investigate whether sensory dominance is consistent within individuals for the spatial functions of resolution acuity, contrast sensitivity and alignment sensitivity and whether these measures are consistent with sighting dominance in the same individual. Different visual functions are underpinned by different mechanisms in the visual system. We hypothesised that the same eye would not necessarily be dominant for all three of the tested visual functions but that dominance, as defined by a significant interocular difference, would be apparent for each of the three functions.
Twenty-one subjects (11 female) aged from 19 to 31 years (mean 23 years) were recruited to the study. Inclusion criteria were anisometropia less than 1.00 DS, astigmatism less than 1.00 DC, letter acuity (Bailey-Lovie chart) 6/6 or better each eye, with less than one line interocular difference and no ocular pathology detected on direct ophthalmoscopy. The study was approved by the Human Research Ethics Advisory of the University of New South Wales and each subject signed a declaration of informed consent prior to participation.
The sighting dominant eye was determined by the hole-in-card test. The subject was seated and held a white card measuring 12 × 20 cm with a central circular hole of three centimetres diameter. The subject's task was to view a 3/60 letter E presented directly ahead at a distance of four metres, through the hole, with the card held with both hands, at arm's length. While viewing the target, the subject's eyes were alternately occluded briefly and the subject was asked to report when the target was visible. The sighting dominant eye was determined as the eye viewing the target, according to the subject's response.
Sensory dominance was determined on the basis of monocular resolution acuity, contrast sensitivity and alignment sensitivity measures. All stimuli were generated using a VSG 2/5 system (Cambridge Research Systems, Rochester, UK) housed in a Dell Pentium computer and displayed on a 21-inch (1024 × 769 pixels; 0.34 mm per pixel) gamma-corrected Clinton Monoray CRT monitor (Richardson Electronics, Illinois). Background luminance of the monitor was 170 cd/m2. All viewing was monocular, with the fellow eye covered with a dark opaque occluder. The order of testing was pseudo-randomised.
The display was viewed from a distance of four metres. On each trial, an E target of 70 per cent Michelson contrast was presented at the centre of the display in one of four orientations (gaps to the left, right, up or down) and the subject's task was to indicate the orientation in a four-option forced-choice double staircase (two-down, one-up) procedure, with fixed step size of 0.02 logMAR. Starting stimulus levels were -1.1 and 0.1 logMAR for the first and second staircases, respectively. The letter was generated on a 5 × 5 grid, with bars and gaps of identical size. Prior to letter presentation, a fixation target consisting of two horizontal and two vertical bars subtending one degree in length and 0.1 degree width were presented one degree away from the centre of the display, while the subject was instructed to look toward the centre. The fixation target was replaced by a uniform display with the E target at centre for 142 ms and the subject was allowed unlimited time to respond. The mean of eight reversals was taken to be resolution acuity and for each acuity measure, 95% confidence intervals were calculated.
The display was viewed from a distance of 1.3 metre. Contrast steps were specified on a decibel (dB) scale, where the relationship between contrast and decibels is:
A Gabor target with carrier horizontal sine wave of spatial frequency one cycle per degree (cpd) and standard deviation of 1.3 degrees was presented for 500 ms at the centre of the display. The Gabor was surrounded by a low-luminance ring centred on the centre of the display and of diameter 11 degrees and thickness 0.33 degrees. In a two temporal alternative forced-choice double staircase (three-down, one-up) procedure, step size 2 dB (starting levels zero and -12 dB for the two staircases), the subject's task was to indicate whether the Gabor appeared in the first or second interval. The inverse of the mean of eight reversals was taken to indicate contrast sensitivity and 95% confidence intervals were calculated.
The display was viewed from a distance of one metre. Two identical Gabor targets with carrier vertical sine wave of 2.6 cpd and Michelson contrast of 98 per cent were presented in vertical alignment. A third identical Gabor was presented vertically mid-way between these two, in horizontal misalignment. Centre-to-centre separation of the Gabor targets was 1.88 degrees at this viewing distance. The subject's task was to indicate whether the central Gabor was to the right or the left of the outer Gabors, in a two-alternative forced-choice, one-right one-left double staircase, with fixed step size of three minutes of arc. Starting points of the two staircases were 10 and -8.5 minutes of arc. Mean and standard deviation of eight reversals were calculated from the psychometric function. Mean was not used because this indicates the subject's right/left bias in alignment setting. Standard deviation was taken to indicate sensitivity to misalignment, with higher standard deviation indicating poorer sensitivity.
For each subject, a dominance ratio ([OD threshold - OS threshold]/[OD threshold + OS threshold]) was calculated for each of the three visual functions. In addition, interocular differences in resolution acuity and contrast sensitivity were tested for significance based on non-overlapping 95% confidence intervals. Any overlap was taken to indicate a lack of significant interocular difference. Overlap of confidence intervals was tested as follows: (Meanmax - Meanmin) + (CIR + CIL), where Meanmax and Meanmin are the higher and lower of the right and left eye threshold values, respectively and CIR and CIL are right and left confidence intervals. A positive value indicated no overlap of confidence intervals. An F-ratio test was used to test for significance in interocular difference in the standard deviation measures of alignment sensitivity.
Figure 1 shows a dominance ratio for resolution acuity, contrast sensitivity and alignment sensitivity, for each of the subjects. Ratios greater than zero indicate thresholds lower for OS than for OD and ratios less than zero indicate thresholds lower for OD than for OS. Thus, for each of the three functions tested here, negative values indicate right eye sensory dominance and positive values indicate left eye sensory dominance. Thus, within individuals, the eye showing sensory dominance (lower threshold) is not necessarily the same across the three functions. Figure 2 shows scatter plots of acuity, contrast sensitivity and alignment sensitivity of the dominant and non-dominant eyes. The plots show that, at least for acuity and contrast sensitivity, the relationship between the two is linear with a slope close to 1, as shown by the similarity in slope between the solid and broken lines. This indicates that differences between the non-dominant and dominant eye in terms of acuity and contrast sensitivity do not depend on the absolute level and thus, are consistent across the range from low to high threshold. Spearman's rho indicates that the correlation between the two eyes is significant (p ≤ 0.001) for each visual function. Not surprisingly and consistent with the definition of sensory dominance used here, the dominant eye has lower thresholds than the non-dominant eye. Figure 3 shows resolution acuity and contrast sensitivity for each eye of each subject. Error bars indicate confidence intervals. For resolution acuity, overlapping confidence intervals indicate that in 17 of the 21 subjects there is no significant interocular difference and thus no significant sensory dominance. For contrast sensitivity, overlapping confidence intervals indicate no significant interocular difference in 20 of the 21 subjects. Interocular difference in alignment sensitivity was tested by F-ratio and this indicated no significant interocular difference in any of the 20 subjects (in one subject, alignment sensitivity data were not available) and for this reason the data are not shown. These findings indicate that, although inter-ocular threshold differences may suggest that one eye is dominant for a particular visual function, the difference may not be sufficient to reflect a true superiority of function of one eye over that of the fellow eye.
In those subjects showing significant dominance in resolution acuity or contrast sensitivity, sighting and sensory dominance are compared in Table 1. Sighting and sensory dominance agree in three of five subjects. Sensory dominance does not agree across the three functions in any of the 21 subjects, as alignment sensitivity dominance is lacking in all subjects, four subjects show dominance in acuity but not contrast sensitivity and one subject shows dominance in contrast sensitivity but not acuity.
Table 1. For each subject showing a significant interocular difference in resolution acuity (RA) or contrast sensitivity (CS), the better (sensory dominant) eye and the sighting dominant eye (SDom) are indicated. No data are shown for alignment sensitivity because none of our subjects showed a significant interocular difference in this function.
We measured sensitivity of three spatial visual functions (resolution acuity, contrast sensitivity and alignment sensitivity) monocularly, in individuals with normal vision, to determine whether the same eye has sensory dominance in each of these functions within individuals. We also measured sighting dominance, to determine whether sighting and sensory dominance are in agreement. Interocular differences in threshold indicated that the dominant eye is not consistent across the three functions. Further analysis indicated that interocular differences in these three functions were not significant in most individuals. Consequently, the question of whether sensory dominance is consistent within individuals across spatial visual functions cannot be satisfactorily addressed here. Instead, we focus our discussion on the main finding of a lack of significant sensory dominance in the normal visual system.
Previous studies in which sensory (acuity or contrast sensitivity) dominance is measured have not assessed the significance of interocular difference in threshold. For example, Ooi and He14 measured contrast sensitivity in each eye and used the ratio of right to left log sensitivity as a measure of interocular difference, while Porac, Whitford and Coren6 counted the number of correct responses to letter or other targets and assigned a score for each eye. The lack of significant interocular difference in acuity and contrast sensitivity found in the present study suggests that, in the normal visual system, sensory dominance of this kind may not be meaningful and may explain the widely reported lack of significant correlation between sighting and sensory dominance. Previous work7 using fMRI has demonstrated significantly higher cortical activation in response to stimulation of the dominant eye, defined by sighting, than the non-dominant eye. In a separate study,8 fMRI shows significantly higher cortical activation in response to the dominant eye than the non-dominant eye when defined by visual acuity but not when defined by sighting. In the latter study, visual acuity was measured using a Snellen chart and in subjects with no interocular difference in Snellen acuity, using a grating stimulus under computer control. The findings of the present study indicate that intra-individual interocular differences in visual acuity are not significant and thus, suggest that correlations between acuity and cortical activation may not be significant when compared within individuals.
Previous findings6 indicate that acuity and sighting dominance may correlate for near, but not for distance acuity. In addition, the sighting dominant eye performs better than the non-dominant eye on a ‘pop-out’ visual search task viewed at a distance of 57 cm.15 Though not noted by the authors, this finding may reflect better correlation between sighting and other forms of eye dominance at near than at distance; further investigation using pop-out tasks at different viewing distances would be required to test this possibility. In the present study, resolution acuity was measured at six metres, while contrast sensitivity and alignment sensitivity were both measured at nearer viewing distances (1.3 m and 1.0 m, respectively). Thus, sensory dominance was measured using distant and nearer targets but was found to be insignificant in most subjects for all tasks. These findings indicate that sensory dominance defined by interocular differences in these spatial functions is insignificant at relatively near viewing distances and that weak ocular dominance at longer viewing distances cannot fully explain poor correlations between sighting and sensory dominance.
In the present study, sensory dominance was measured according to intra-individual interocular differences in each of three spatial functions. We asked whether differences were significant by testing for overlap of 95% confidence intervals (for acuity and contrast sensitivity) or by F-ratio (for alignment sensitivity). Less stringent tests would allow more comparisons to pass as significant. We measured acuity and both contrast and alignment sensitivities using a particular set of psychophysical methods and again the outcomes would be different if different methods were used for the assessment of thresholds. It would be interesting to investigate whether sensory dominance, if significant using a different set of tests, is repeatable and consistent within individuals and thus whether it is meaningful.
Eye care practitioners sometimes use ocular dominance when making clinical decisions. For example, in monovision contact lens wear, the dominant eye is usually corrected for distance and the non-dominant eye for near, based on the idea that the dominant eye will be less easily suppressed by the relatively blurred image in the fellow eye.13 In practice, dominance is measured using a range of techniques and may be defined according to motor factors contributing to sighting or by sensory factors contributing to a better-seeing eye on sensory tests such as acuity. Our findings add to previous work indicating that acuity or other measures of spatial vision may not be good indicators of ocular dominance and may not be useful as a basis for clinical decision-making in eye care, in the normal visual system. It should be noted that in conditions such as unilateral amblyopia, one eye is significantly and consistently superior and ocular dominance, defined according to inter-ocular difference in thresholds, is likely to be significant for a range of spatial visual functions.
The lack of significant ocular dominance reported here adds to previous work investigating the significance of ocular dominance. For example, Seijas and colleagues5 measured sighting and sensory ocular dominance using nine methods, in young and older adults with normal vision and found poor correlations between the tests, indicating poor intra-individual agreement between different measures of ocular dominance. They concluded that ocular dominance did not exist in most of their subjects, as the dominant eye was method-dependent and not consistent across tests. Some previous studies found that interocular interactions, such as suppression of a viewing eye during dark occlusion of the fellow eye, are stronger when the dominant eye rather than the non-dominant eye is occluded.16,17 This effect may depend on the method of ocular dominance determination and on test conditions.18 Ocular dominance has also been found to depend on horizontal gaze angle, with the right eye more dominant in rightward gaze and the left in leftward gaze and this effect may be due to interocular retinal image size differences.19
Previous studies showing a significant effect of eye dominance on cortical activation or interocular interactions have tested for significance using group data,8,17 while significance of eye dominance in the present study was measured within individuals. Perhaps eye dominance effects are more apparent when assessed within groups rather than within individuals. Paired t-tests indicate no significant interocular difference in acuity, contrast sensitivity or alignment sensitivity in the present study (p > 0.2), so this factor does not explain the difference between the present and previous findings.
The results presented here are in agreement with previous work indicating that measures of ocular dominance show poor agreement between methods within individuals and that within one method the outcome may depend on factors other than the relative strength of the right or left eye signal. The present and previous findings raise questions on the significance of ocular dominance and its meaning in normal visual function. Our findings differ from those of a number of previous studies, indicating that at least some measures of visual system function (for example, strength of interocular interactions and of cortical signal) depend on ocular dominance. Taken together, the present and previous work suggest that ocular dominance may be significant in some conditions but that sensory dominance measured by threshold difference is not significant in observers with normal visual function.