Spatial Localization in Esotropia: Does Extraretinal Eye Position Information Change?

Abstract

purpose. To investigate the accuracy of spatial localization in children with a specific type of convergent strabismus, fully accommodative esotropia.

methods. Two groups of children, with right and left fully accommodative esotropia, respectively, pointed at targets located centrally and eccentrically on a computer touchscreen without being able to see their hands. The size and the direction of the horizontal pointing responses were recorded under two conditions: when their eyes were aligned (wearing spectacles) and when they were squinting (not wearing spectacles). A group of children without strabismus but with hypermetropia were assessed as controls.

results. For both fully accommodative groups, the pointing responses to the central target shifted in the direction of the nonsquinting eye when deviations were manifest. No difference was found for the eccentric targets. No difference was found for the hypermetropia group with any target.

conclusions. These results are consistent with an alteration in the extraretinal eye position information (efference copy, extraocular muscle proprioception, or both) that is used in spatial localization.

Spatial localization refers to the ability to accurately locate the position of a target in surrounding visual space. It is an essential part of normal visual function in both children and adults and is important in all manner of activities, from simply reaching for an object to more complex tasks such as driving a car. To accurately determine the position of an object relative to the person, the brain must know the direction in which the person is looking, and it relies predominantly on visual (i.e., retinal) input for this information. However, because the eyes are free to move within the orbits, retinal information on its own is not sufficient to specify this visual direction. Additional nonvisual (i.e., extraretinal) information is necessary to interpret the retinal data and thus determine the direction of gaze. This extraretinal eye position information is thought to be derived from two distinct extraretinal sources: monitoring of the motor command sent to the extraocular muscles (efference copy or corollary discharge ^{1 2} ) and inflowing sensory information derived from extraocular muscle proprioceptors. ³ However, the respective contribution of each has been the subject of considerable debate for many years. ^{4 5}  

The task of pointing at targets in surrounding space is a recognized method of assessing spatial localization. ^{6 7 8 9} The combination of retinal and extraretinal information is derived from the sources outlined and then integrated to determine precisely the location of an object of interest. If the person wants to reach out or point to this object, an appropriate motor command is sent to the arm and hand that enables the person to perform the task efficiently, with visual feedback used to ensure greater accuracy. Persons with strabismus are known to make errors when asked to perform such tasks of spatial localization, ^{6 9 10} particularly when they are unable to see the pointing hand (a procedure that eliminates visual feedback). However, the cause of these errors is not fully understood. In addition, interpretation of the results of such studies is hindered by the inclusion of subjects of various ages (both children and adults) and also with different types of strabismus. 

We therefore decided to investigate spatial localization in more detail in children of similar ages, with one specific form of strabismus: fully accommodative esotropia. This is a manifest convergent strabismus, caused by uncorrected hypermetropia and insufficient fusional divergence, in which full correction of the hypermetropic refractive error restores eye alignment. ¹¹ This therefore allowed us to compare the binocular pointing responses of these subjects when the eyes were aligned (when wearing glasses) and with a manifest squint (without glasses). A group of comparable subjects without strabismus but with hypermetropia (nonstrabismus group) was assessed for comparison. 

Methods

All procedures conformed to the Declaration of Helsinki for research involving human subjects. Ethics committee approval was obtained, and parental informed consent was given in all cases. 

Procedure

Subjects were seated and viewed a computer touchscreen (luminance 57 candelas [cd]/m²; IBM, Greenock, UK) from a distance of 26.5 cm. Their heads were stabilized using chin rests and cheek pads. Pictures of three vertical poles were presented on the screen: in the center and 20° to the left and 20° to the right of center (Fig. 1) . A green target (in the shape of a dragon, 4 × 2 cm, luminance 45 cd/m²) appeared on the top of one pole, and the subject was asked to touch the screen, with the outstretched index finger of the dominant hand, at the bottom of the pole on which the dragon had landed. The dragon then jumped randomly to the top of another pole, and the subject was asked to touch the location of this pole. No limit of time was placed on the pointing response. The target was presented on 10 separate occasions at each pole, and the location of each pointing response was stored on-line for later analysis. The mean touched location of the 10 presentations was taken as the horizontal pointing error for that pole. A trial run was allowed in which the subjects could initially visualize the pointing hand to enable them to become familiar with the testing procedure. A cardboard sheet covered with black cloth was then used to mask the lower part of the screen thereby preventing the pointing hand’s being seen. The subjects were allowed to practice with the screen in place. The formal testing session then began. Each subject was tested both with and without hypermetropic refractive correction. To try to avoid any bias due to a learning effect, we tossed a coin to determine which condition was tested first. All children were tested binocularly, with the hypermetropic group also tested monocularly. 

The difference between the mean pointing errors with glasses (not squinting) and without glasses (squinting) was then calculated for individual subjects for each pole. Statistical analysis was performed by computer (Prism; GraphPad Software, San Diego, CA). 

Subjects

Ninety children participated in the study and were divided into three groups of 30 as follows: (1) right fully accommodative esotropia (mean age, 5 years, 9 months; range, 4 years, 4 months to 7 years); (2) left fully accommodative esotropia (mean age, 5 years, 11 months; range, 4 years, 5 months to 7 years, 2 months); (3) hypermetropia (mean age, 5 years, 9 months; range, 4 years, 6 months to 7 years). The number of subjects required in each group was determined using a power calculation. ¹² None of the children had a history of strabismus surgery, and none had any medical history of note. Any fully accommodative subjects who noted diplopia when the deviation was manifest were excluded. At the time of testing, none had any evidence of amblyopia, as defined by uniocular visual acuity of less than 0.250 (log minimum angle of resolution [MAR]) or an interocular acuity difference of more than 0.1 log units. ¹³ Visual acuities were recorded unaided and with refractive correction using logMAR-crowded tests at a distance of 3 m ¹⁴ and MacLure Reading Type for Children at a distance of 25 cm. When appropriate, the angles of the esodeviation for 6 m and 33 cm, unaided and with refractive correction, were measured using the prism cover test. The strength of the spectacle correction worn was also noted. A summary of these clinical details is given in Table 1 . 

Results

All the children were able to perform the test without any difficulty, both with and without spectacle correction. For children with right fully accommodative esotropia, the mean decrease in distance visual acuity when not wearing spectacle correction was 0.14 ± 0.1 log units (SD) for the right eye and 0.1 ± 0.08 log units for the left eye. For children with left fully accommodative esotropia, the mean decrease in distance visual acuity when not wearing spectacle correction was 0.1 ± 0.1 log units for the right eye and 0.13 ± 0.07 log units for the left eye. For children with hypermetropia the mean decrease in visual acuity when not wearing spectacle correction was 0.1 ± 0.09 log units for both eyes. Near visual acuity was N5 in all the fully accommodative groups and in children with hypermetropia, with and without spectacle correction. 

A significant shift in the mean pointing response to the central target was found in both the left and right fully accommodative groups when comparing responses with glasses (not squinting) to those without glasses (squinting; Fig. 2A ). The right fully accommodative esotropes showed a mean pointing shift to the left of 1.7° ± 2.1° (SD) when they were squinting (P < 0.001, t = 5.1, Student’s t-test). This effect was observed in 24 (80%) subjects. The left fully accommodative esotropes showed a mean pointing shift to the right of 1.1° ± 1.2° when they were squinting (P < 0.001, t = 4.96). This effect was observed in 25 (83%) subjects. There was no significant shift in the localization position for the two eccentric targets in either strabismus group when they were tested without glasses. For example, for the left 20° target, the right fully accommodative esotropes showed a mean pointing shift to the left of 0.3° ± 2.6° (P = 0.5, t = 0.6), and the left fully accommodative esotropes showed a mean shift of 0.4° to the right (SD ± 2.9; P = 0.5, t= 0.7; Fig. 2B ). For the right 20° target, the right fully accommodative esotropes showed a mean pointing shift to the left of 0.1° ± 3.0°; P = 0.8, t = 0.2), and the left fully accommodative esotropes showed a mean shift of 0.2° ± 2.7° to the right (P = 0.6, t = 0.5; Fig. 2C ). No correlation was found among refractive error, the age of the subjects, the angle of the deviation, and the pointing shift. 

No significant localization shift was observed in the hypermetropia group for any of the three targets when the subjects were tested binocularly without their glasses (Fig. 2A 2B 2C) . For example, for the left eccentric, center, and right eccentric targets they showed a mean shift of 0.2° ± 3.3° to the right (P = 0.7, t = 0.4), 0.2° ± 2.6° to the left (P = 0.7, t = 0.4), and 0.4° ± 3.0° to the right (P = 0.5, t = 0.7), respectively. In addition, no significant localization shift was observed in the hypermetropia group for any of the three targets when the subjects were tested monocularly without their glasses (Fig. 2D) . For example, for the left eccentric, center, and right eccentric targets they showed a mean shift of 0.1° ± 3.5° to the left (P = 0.9, t = 0.1), 0.2° ± 2.9° to the right (P = 0.8, t = 0.3), and 0.4° ± 3.6° to the left (P = 0.3, t = 1), respectively. 

Discussion

This study demonstrates that for centrally located targets, spatial localization in children with fully accommodative esotropia shifts in the direction of the nonsquinting eye when the deviation is manifest. To our knowledge changes in spatial localization in this well-defined group of children with strabismus has not been reported previously. Assessing pointing responses when the subject is unable to see the pointing hand is a recognized method of assessing the accuracy of spatial localization in both children and adults. ^{7 8 9} The variability of responses observed in this study is not surprising, considering that we were testing children with a mean age of less than 6 years and that findings in such studies by their very nature (i.e., pointing when unable to see the hand) tend to be variable. 

The shift in pointing response was only observed for the central target. A small effect may have been present at the eccentric positions but could have been masked by the greater variability of the pointing responses, as noted by the larger standard deviations. In addition, it is conceivable that peripherally located targets could have stimulated retinal loci outwith the suppression scotomas in the deviating eye. This would have provided further visual information that would help determine the altered direction of gaze of the squinting eye, thereby preventing a localization shift. Interestingly, Fronius and Sireteanu ⁹ tested the pointing responses in a heterogeneous group of children with strabismus and concluded that spatial localization may be altered to varying degrees within different areas of the visual field. Although they assessed only four patients under experimental conditions similar to those in our study (i.e., subjects unable to see the pointing hand), their results highlight the variable nature of spatial localization, particularly among patients with strabismus. 

For the central target, the direction of the pointing shift was noted to be in the direction in which the squinting eye was looking. That the position of one (presumably suppressed) eye can influence the perception of visual direction when viewing with the dominant contralateral eye is not surprising. When Ono and Weber ¹⁵ studied the pointing responses of normal adult subjects, they found that during monocular viewing, a shift in spatial localization occurred. The direction of this shift was related to the direction of the phoria of the occluded eye, indicating that the position of both eyes is taken into account when performing such tasks. A similar finding was reported by Mann et al., ⁶ who studied a group of constantly suppressing esotropic and exotropic strabismus patients. They noted that positional information from the dominant eye influenced the pointing responses when subjects viewed targets monocularly with the suppressed eye. They also found that the size of the localization shift correlated with the angle of strabismus. However, we failed to identify such a relationship, a result that is in keeping with the findings of Fronius and Sireteanu, ⁹ who emphasize that this noncorrelation is not unexpected, given the complex causes of pointing errors in children with strabismus. Although there are similarities between our study and those of Mann et. al. ⁶ and Fronius and Sireteanu, ⁹ it should be remembered that they examined monocular spatial localization in contrast to our binocular testing procedure, which is perhaps more relevant to everyday tasks. In addition, their subjects were significantly older (age range, 6–32 years) and had several types of strabismus (including esotropia and exotropia). In contrast, we assessed younger children (age range, 4 years, 6 months to 7 years) who were all fully accommodative esotropes. 

Why should we have observed a shift in the pointing response when the children were squinting? As was discussed in the introduction we rely on a combination of both retinal and extraretinal information for accurate spatial localization. Therefore, a change in one of these may account for our findings (discussed in more detail later in the article). We are assuming, of course, that the localization shifts we observed were not related to any alteration in the motor control of the pointing arm. We have no reason to believe otherwise. 

Change in Retinal Information

The fully accommodative subjects’ distance visual acuities declined when they removed their glasses. It is possible that this decrease in acuity resulted in a greater degree of inaccuracy when performing the pointing test and may therefore explain our findings. This explanation, however, is unlikely, because the hypermetropia control group had a similar reduction in distance acuity but showed no significant change in pointing response when they were tested without their refractive corrections. Although the decrease in distance acuity was greater for the deviating eye, this is unlikely to have been a contributory factor, because none of the subjects mentioned diplopia. It was assumed, therefore, that the retinal image from this eye was suppressed and did not provide visual information about the location of the central target. In addition, the testing was performed at a distance of 26.5 cm (i.e., near), and in both the fully accommodative groups, near vision remained at N5 in all subjects. 

Another aspect of visual function that changed when the children were squinting, was binocularity. When their deviations were manifest they no longer had stereopsis. Could this have affected their spatial perception? Again, this is unlikely, because when the subjects with hypermetropia were tested monocularly (thereby eliminating binocular vision), it did not affect their perceived location of the target. On this basis we can assume that stereopsis was not required to perform the test accurately. 

Change in Extraretinal Information

If visual (retinal) information cannot explain the pointing shift, then a change in the nature of the extraretinal eye position signal that is used to determine visual direction may be the answer. As discussed earlier, the two possible sources of this extraretinal information are efference copy and extraocular muscle proprioception. 

Could the efferent copy of the oculomotor command have changed when the children were squinting? When the subjects’ deviations were manifest, the fixating, dominant eye, viewed the same targets in the same position as when the subjects’ eyes were aligned. According to Walls ¹⁶ the visual system monitors only the efference command sent to the dominant eye. If this was the case in our subjects with fully accommodative esotropia, then efference copy should have been unchanged when the nondominant eye was squinting. This means that the shift in localization that we have observed cannot be explained by an alteration in efference copy. Bridgeman ⁴ also supports the notion that there is only one copy of the efferent command, which, according to Hering’s law, represents the equal motor innervation sent to both eyes. Although this would be sufficient to specify binocular visual direction when the eyes are aligned, it is not clear what happens to efference copy when one eye is deviated relative to the other. We therefore cannot be sure about the role of efference copy in manifest strabismus and whether it influences the extraretinal eye position signal that contributes to spatial localization under such circumstances. 

The second component of extraretinal eye position information is proprioceptive input from the extraocular muscles. In our children with strabismus, when their deviation is manifest, the relative stretch on the lateral rectus and medial rectus muscles of the squinting eye must be different from that of the fixing, nonsquinting eye. It is reasonable to assume, therefore, that the proprioceptive feedback must be altered. Because the sense of visual direction is partly determined by the afferent input from both eyes, this alteration in feedback could result in an erroneous eye position signal, producing a shift in the perceived location of a target. There is an increasing body of evidence to support such an explanation. For example, Gauthier et al. ⁸ created experimental strabismus in normal subjects by passively rotating one eye using a suction contact lens, a technique believed to modify extraocular muscle proprioception. When subjected to passive rotation, the subjects consistently mislocated targets in the direction of the deviation, a finding consistent with our results. Two other experimental studies have demonstrated that manipulating extraocular muscle proprioception produces mean pointing shifts of 2.5° ¹⁷ and 2.98° ¹⁸ . Although these shifts are slightly larger than in our subjects, they are of a similar magnitude. Alterations in pointing responses after different forms of strabismus surgery have also been reported, ^{7 19} findings believed to result from modified proprioception secondary to surgical damage. The exact site(s) at which proprioception influences spatial perception is not known, although possibilities include the lateral geniculate nucleus and the visual cortex, both of which respond to stimulation of extraocular muscle afferent input. ^{20 21}  

Although we believe the localization shifts we observed were the result of a change in the extraretinal eye position signal, we cannot be sure whether this was due to an alteration in efference copy, proprioception, or both. However, the balance of evidence certainly suggests that modified proprioception was a contributing factor. 

Our findings are not only of theoretical importance, they also have practical implications by providing some insight into the ability of children with strabismus to function with respect to their visual environment. Although we found relatively small shifts in localization (up to 1.8°), this would equate to children’s inaccurately judging the position of targets by between 1 to 2 cm at arm’s length. This has implications even for performing simple tasks in everyday life, such as catching a ball or picking up a cup. Spatial localization is an aspect of visual function that is often overlooked, and although the majority of children probably do not experience any difficulties, those with strabismus perhaps warrant further assessment. 

 Supported by grants from the Chief Scientist Office, Scottish Office, and The W. H. Ross Foundation, Scotland, United Kingdom.

 Submitted for publication April 6, 2000; revised June 28, 2000; accepted July 19, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

A diagram showing the computer touchscreen and the dragon that flies and lands on successive targets, or poles.

Table 1.

 

Summary of Clinical Details for Each Group of Subjects

	Refractive Error*			Angle of Esodeviation, †
		Right Eye	Left Eye	Near	Distance
Right fully accommodative Esotropia	5.3 ± 1.7	5.1 ± 1.8	24.9 ± 9.8	19.9 ± 7.9
Left fully accommodative Esotropia	4.7 ± 1.5	5.3 ± 1.3	29.2 ± 10	25 ± 8.8
Hypermetropia	3.9 ± 1.3	4 ± 1.4

 Values are means ± SD.

*  Mean spherical equivalent in diopters.

†  In prism diopters.

Figure 2.

 

Mean horizontal pointing shifts in degrees for individual children with strabismus and hypermetropia. (A, B, and C) Shifts for binocular viewing for the center target, the left 20° target, and the right 20° target, respectively. (D) Children with hypermetropia viewing all three targets monocularly. Positive values represent shifts to the right; negative values, shifts to the left. Right FA, right fully accommodative esotropes; Left FA, left fully accommodative esotropes; Hyper, hypermetrope control subjects.