Effects of Strabismic Amblyopia on Visuomotor Behavior:
Part II. Visually Guided Reaching

Abstract

Purpose.: To examine the effects of impaired spatiotemporal vision on reaching movements in participants with strabismic amblyopia and to compare their performance to those with strabismus only without amblyopia and to visually normal participants.

Methods.: Sixteen adults with strabismic amblyopia, 14 adults with strabismus only, and 16 visually normal adults were recruited. Participants executed reach-to-touch movements toward targets presented randomly 5° or 10° to the left or right of central fixation in three viewing conditions: both eyes, monocular amblyopic eye (nondominant eye for participants without amblyopia), and monocular fellow eye (dominant eye for participants without amblyopia). Visual feedback of the target was removed on 50% of the trials at the initiation of reaching.

Results.: Both groups with abnormal binocular vision (strabismic amblyopia and strabismus only) had reach latency, accuracy, and precision comparable to visually normal participants when viewing with both eyes and fellow (dominant) eye. Latencies were significantly delayed by more than 30 ms in all participants with reduced binocularity during amblyopic eye or nondominant eye viewing compared with controls (P < 0.0001). Participants with strabismic amblyopia and negative stereopsis also had reduced reach precision (i.e., increased variability) during amblyopic eye viewing. In contrast, participants with strabismus only and negative stereopsis had comparable precision across all viewing conditions. Participants with strabismus only and those with strabismic amblyopia used a similar motor strategy; regardless of viewing condition, reach peak acceleration was significantly reduced (P < 0.05) and the duration of acceleration phase was extended in comparison with visually normal participants. There were no significant differences for the deceleration phase.

Conclusions.: Participants with strabismic amblyopia and those with strabismus only attain relatively normal reach accuracy and precision. However, they use a different reach strategy that involves changing the motor plan. A similar compensatory strategy was reported previously in participants with anisometropic amblyopia. Our results provide further support that normal binocular vision during development provides important input for the development of visually guided reaching movements.

Introduction

Vision provides important sensory input for guiding goal-directed upper limb movements, including reaching, grasping, and manipulation of objects. ^1,2 For example, to reach for and grasp a cup of coffee, vision is used to localize the cup and to program the direction and distance of the reaching movement. ^3,4 Visual information about the size and orientation of the cup is also used to program the appropriate grasp aperture. ⁵  

Amblyopia is a unilateral (or less commonly, bilateral) reduction of best-corrected visual acuity that cannot be attributed only and directly to a structural abnormality of the eye. ⁶ It is caused by abnormal visual experience early in life and cannot be remedied immediately by spectacle glasses alone. ⁶ It is defined clinically as at least a two-line difference in best-corrected acuity between the eyes. ⁶ Amblyopia is the number one cause of monocular blindness, affecting 3% to 5% of the population worldwide. ^7–13 The two most common disruptions of visual input that lead to amblyopia are unequal refractive error (anisometropic amblyopia) and eye deviation (strabismic amblyopia). ¹⁴ Regardless of etiology, the hallmark of amblyopia is the impairment in spatiotemporal visual processing, which includes deficits in monocular vision (i.e., acuity and contrast sensitivity). In addition, amblyopia is often associated with binocular visual function deficits (stereopsis). ^15–18 In addition to reduced monocular acuity in the amblyopic eye, patients have difficulties with extracting global form and contour integration, ^19–21 discriminating global motion, ^22,23 and identifying complex objects presented in different viewpoints, ²⁴ as well as impairments in attentional ²⁵ and decision-making processes. ²⁶ Although these deficits are most pronounced during amblyopic eye viewing, they also are evident to some degree when patients view binocularly or with the fellow eye. ^24,26–29  

The effects of amblyopia on spatiotemporal visual processing have been studied extensively, both behaviorally and using imaging techniques ^15,30,31 ; however, relatively fewer studies have investigated the effects of this visual impairment on visuomotor behaviors. Because visual input is used to guide most of our motor behaviors, it is important to understand the impact of impaired vision on spatial localization, movement planning, sensorimotor integration, and reach execution. Previous studies have used a variety of localization tasks, including pointing, to systematically investigate the effect of amblyopia and strabismus on spatial representations. ^32–36 These studies showed that people with strabismic amblyopia have greater deficits in accuracy and increased endpoint variability in the central visual field in comparison with the peripheral field during amblyopic eye viewing. Spatial distortions and loss of precision during alignment tasks also were reported in several studies. ^37,38 Because target localization is necessary for all visuomotor behaviors, the results from these studies suggest that people with amblyopia might have deficits in subsequent processing stages leading to movement planning and execution. Grant et al. ³⁹ conducted the first study to examine the kinematic trajectories of reaching and grasping movements in a group of people with different amblyopia subtypes. They reported that patients had more difficulty shaping the appropriate grasp aperture and exhibited more errors during grasp application, whereas movements were comparable with visually normal participants during the transport phase. Although their study provided important information on the effects of amblyopia on prehension movements, it included people with various amblyopia subtypes. Because the sample size was small (n = 20), their study could not properly assess the effects of amblyopia etiology and severity of the acuity or stereoacuity deficits on reaching and grasping movements. 

There are several important differences among patients with different amblyopia subtypes. For example, patients with anisometropic amblyopia and patients with strabismic amblyopia show a different pattern of deficits across the visual field. Patients with strabismic amblyopia show the greatest deficits in localization and alignment tasks in the central visual field; however, their performance on these tasks in the peripheral visual field is similar to that in visually normal subjects. ³⁴ In contrast, patients with anisometropic amblyopia show deficits on localization and alignment tasks across the entire visual field (central and peripheral). ⁴⁰ Another important difference between strabismic and anisometropic amblyopia was highlighted in a seminal study by McKee and colleagues. ⁴¹ They showed that patients with strabismic amblyopia are more likely to have greater deficits in binocular visual function (i.e., reduced or negative stereopsis) than patients with anisometropic amblyopia with a comparable level of acuity loss in the amblyopic eye. 

The goal of our research is to characterize the effects of amblyopia etiology, as well as severity of the acuity and stereoacuity deficits on eye movements and visually guided reaching movements. Our group has systematically investigated the effects of amblyopia on saccadic eye movements in people with anisometropic ⁴² and strabismic amblyopia. ⁴³ We also have previously reported the effects of anisometropic amblyopia on reaching movements. ⁴⁴ In this study, we examined reaching movements in people with strabismic amblyopia and compared their performance to visually normal participants, as well as to people with strabismus only without amblyopia. Because binocular vision provides important input for planning and for online control of reaching movements, ^45–54 we hypothesized that individuals with strabismic amblyopia will show greater deficits in the accuracy and precision of their reaching movements as compared with individuals with anisometropic amblyopia. We also predicted that for participants with amblyopia, these deficits will be related to the loss of binocular visual function: individuals with negative stereopsis will have a greater impairment in reach accuracy and precision in comparison with individuals with residual stereopsis. 

Materials and Methods

Participants

Forty-six adults were tested: 16 participants with strabismic amblyopia (nine males; mean age 32 ± 9 years), 14 participants with strabismus only without amblyopia (seven males; mean age 30 ± 5 years), and 16 visually normal participants (nine males; mean age 31 ± 9 years). The saccadic eye movements of 43 of these 46 participants have been reported previously. ⁴³ All participants underwent a complete orthoptic assessment, including visual acuity testing using the Snellen chart, measurement of eye alignment using the prism cover test, and stereoacuity testing using the Titmus test. Refractive error was obtained from the participant's most recent prescription (see Supplementary Table S1 for participants' clinical characteristics). Strabismic amblyopia was defined as an interocular acuity difference of 2 or more chart lines with a history of childhood strabismus and manifest eye deviation. The difference in refractive error between the two eyes was 1 diopter (D) or less of spherical or cylindrical power to rule out anisometropia as a potentially amblyogenic component (i.e., to rule out mixed mechanism amblyopia). All participants with strabismic amblyopia had visual acuity between 20/30 and 20/200 in the amblyopic eye, and 20/20 or better in the fellow eye. Eleven participants had mild amblyopia, with acuity in the amblyopic eye ranging from 20/30 to 20/70. Five of these 11 participants with mild amblyopia had residual stereopsis (range, 120–800 arc seconds), whereas the other six participants had no stereopsis. Five participants had severe amblyopia (20/100 or worse in the amblyopic eye) and no stereopsis. Participants with strabismus only without amblyopia had visual acuity of 20/25 or better in both eyes and manifest eye deviation. Nine participants had no stereopsis, two had stereoacuity of 3000 arc seconds, and the remaining three had stereoacuity between 50 and 70 arc seconds. Seven participants in the strabismus-only group had received patching therapy previously, suggesting that they had a history of amblyopia. Visually normal participants had corrected-to-normal visual acuity (20/20 or better) in both eyes and stereoacuity of 40 seconds or less of arc. 

Exclusion criteria were any ocular cause for reduced visual acuity, prior intraocular surgery, or any neurologic disease. All participants were right-handed. The study was approved by the Research Ethics Board at The Hospital for Sick Children and all protocols adhered to the guidelines of the Declaration of Helsinki. Informed consent was obtained from each participant. 

Apparatus and Experimental Procedure

Reaching movements of the upper limb were recorded using an infrared illumination-based motion capture system (Optotrak Certus; Northern Digital, Waterloo, ON, Canada). Details of the apparatus and calibration procedure have been previously described. ⁴⁴ Briefly, two infrared markers were affixed to the tip of the index finger and the wrist joint of a participant's right hand. Testing was conducted in a dimly lit room. Participants were seated at a table in front of a cathode ray tube monitor (viewing distance was 42 cm) with their heads stabilized on a chin rest. The visual stimulus was a white circle (visual angle 0.5°) presented on a black background. Stimulus presentation was controlled by a custom-written Matlab script (MathWorks, Natick, MA, USA) using the ViSaGe visual stimulus generator (Cambridge Research Systems, Rochester, UK). A force-sensitive resistor (FSR; Tekscan, Boston, MA, USA), 15 mm in diameter, was placed on the table at the participant's midline 28 cm from the computer screen. The FSR was used to trigger the initiation of each trial and to control when the visual target was switched off during a trial. At the start of each trial, participants fixated a white cross on the screen that was centered vertically at their eye level and horizontally along their midsagittal plane. After a variable delay ranging between 1.5 and 3.0 seconds, the fixation cross was extinguished and the target appeared at one of four locations along the azimuth (±5° or ±10° from fixation). Participants were instructed to look and reach to the target as fast and as accurately as possible. On 50% of the trials, the target was switched off at the onset of hand movement, that is, as soon as the finger was lifted off the FSR (target OFF condition). On these trials, participants were instructed to reach to the remembered location of the target. On the remaining 50% of the trials, the target remained visible throughout the trial (target ON condition). The target OFF and ON conditions were randomly interleaved. 

Table

 

Results of Multiple Regression Analyses

Viewing Condition	Peak Acceleration		Duration of Deceleration Phase			P Value
	β	P Value	β	P Value
Control participants, n = 16
Both eyes	0.03	0.903	−0.56	0.065	0.23	0.071
Fellow, dominant eye	0.14	0.54	−0.65	0.013	0.47	0.006
Amblyopic, nondominant eye	0.17	0.225	−0.80	<0.0001	0.72	<0.0001
Participants with amblyopia, n = 16
Both eyes	0.63	0.026	0.08	0.739	0.26	0.058
Fellow, dominant eye	0.50	0.044	−0.28	0.231	0.37	0.021
Amblyopic, nondominant eye	0.35	0.232	0.208	0.472	0.02	0.458
Participants with strabismus only, n = 14
Both eyes	−0.09	0.812	−0.74	0.067	0.36	0.035
Fellow, dominant eye	0.03	0.924	−0.62	0.103	0.31	0.052
Amblyopic, nondominant eye	−0.34	0.345	−0.71	0.087	0.14	0.178

The experiments were performed under three viewing conditions: both eyes, monocular amblyopic eye, and monocular fellow eye viewing. A black eye patch was used during the monocular viewing conditions. For participants with strabismus only and control subjects, viewing was both eyes, monocular with the dominant eye, and monocular with the nondominant eye. Eye dominance for visually normal participants was determined using Dolman's “hole-in-card” test. ⁵⁵ Data were collected in blocks for each viewing condition, and the order of viewing conditions was randomized across subjects. Participants completed 10 trials in each combination of the experimental conditions for a total of 240 trials. Practice trials were completed before the start of the experiment to familiarize subjects with the experimental procedure. 

Data Analysis

Hand position data were filtered using a second-order dual-pass (bidirectional) Butterworth filter with a cutoff frequency of 7.5 Hz. Hand velocity was computed using a 2-point differentiation method. Position data were differentiated twice to obtain acceleration. A custom-written Matlab script (MathWorks) was used to identify the initiation of the hand movement, defined here as when the velocity of the finger in the y-coordinate (i.e., elevation) exceeded 30 mm per second. The end of the reaching movement was identified when the finger reached the computer screen and the velocity of the finger in the z-axis (depth) fell to and stayed below 30 mm per second. All trials were inspected visually to ensure that the reaching movement was identified correctly by the program. 

Reach Outcome Measures: Latency, Accuracy, and Precision.

The coordinate system was defined relative to the computer monitor used to present the visual stimulus as follows: x, horizontal plane (azimuth); y, vertical plane (elevation); and z, median plane (depth). Reaching performance was quantified by calculating the movement latency, accuracy, and precision. Latency was defined as the interval between the onset of the visual stimulus and the initiation of reaching. Accuracy was assessed by calculating end point constant error (response bias) along the azimuth and elevation directions separately, as well as the overall error. Constant error was the signed error defined as the mean distance between the fingertip and the target location along the azimuth and elevation at the end of the movement. Overall accuracy was assessed by calculating radial error, which was the resultant error along the azimuth and elevation (unsigned error). Precision was assessed by calculating the SD across the reaching trials in a given experimental condition along the azimuth and elevation, as well as the radial error. 

Reach Performance Measures: Kinematics.

The kinematics of the reaching movement were assessed by calculating the following parameters: total movement time (the interval between reaching initiation and the end of movement), peak acceleration, peak velocity, the duration of the acceleration phase (the interval from movement onset to peak velocity, i.e., the zero-crossing on the acceleration trajectory), and the duration of the deceleration phase (the interval from peak velocity to the end of the hand movement). Peak acceleration, peak velocity, and the duration of the acceleration phase reflect the programming of the movement (i.e., feedforward control), whereas peak deceleration and the duration of deceleration phase reflect on-line (i.e., feedback) control. ⁵⁶ All kinematic parameters were calculated based on the z-axis of the finger, which represents the primary direction of movement in depth. 

Reaching accuracy (constant and radial error) and precision (variable error), as well as reach kinematic measures, were submitted to a repeated measures mixed ANOVA with Group as a between-subjects factor (participants with strabismic amblyopia, participants with strabismus only, visually normal participants) and three within-subject factors: Viewing Condition (both eyes, monocular fellow eye, and monocular amblyopic eye viewing; for participants with strabismus only and visually normal participants viewing was both eyes, and monocular with the dominant and nondominant eye), Target Visual Feedback (target ON, and target OFF), and Target Location (±5°, ±10°). 

Peak acceleration provides information about feedforward control (i.e., movement planning), whereas the duration of deceleration phase provides information about feedback control during reach execution. A multiple regression analysis was performed to assess the relative contributions of feedforward (peak acceleration) and feedback (duration of deceleration phases) to reach precision. Standardized regression coefficients (β) and their associated P values were calculated for the three groups in different viewing conditions. The value of the β weight reflects the proportion of the variance explained in the dependent variable (reach precision) by the changes in the variance of independent variables (peak acceleration and the duration of the deceleration phase). Separate multiple regressions were performed on the group mean data from both patient groups and visually normal participants for the three viewing conditions. 

The Effects of Severity of Amblyopia and Stereopsis.

To investigate further the effects of severity of amblyopia and stereopsis, a separate repeated-measures ANOVA was performed on each outcome measure. For this analysis, participants with amblyopia were stratified into three subgroups: (1) mild amblyopia (i.e., acuity ≤ 20/70) and positive stereopsis (mild & stereo⁺, n = 5); (2) mild amblyopia and negative stereopsis (mild & stereo⁻, n = 6); and (3) severe amblyopia and negative stereopsis (i.e., acuity ≥ 20/100; n = 5). The ANOVA had Subgroup as a between-subjects factor (i.e., mild & stereo⁺, mild & stereo⁻, severe) and two within-subjects factors: Viewing Condition (both eyes, monocular fellow eye, and monocular amblyopic eye), and Target Location (±5°, ±10°). 

All statistical analyses were performed using the SAS 9.2 software package (SAS Institute, Inc., Cary, NC, USA). The significance level was set at P less than 0.05. Any significant main effects and interactions were analyzed further using Tukey's HSD test. 

Results

Reach Outcome Measures

Latency.

There was a significant main effect of Viewing Condition for reach latency (F _2,86 = 8.25, P = 0.0005). Overall, reach latencies were longer when viewing with the amblyopic eye or the nondominant eye of patients with strabismus without amblyopia. For visually normal participants, reach latency was similar for both monocular viewing conditions (left eye 338 ± 70 ms; right eye 333 ± 84 ms), which was longer in comparison to viewing with both eyes condition (323 ± 85 ms). Participants with amblyopia had longer latencies only during amblyopic eye viewing (404 ± 114 ms) in comparison to both eyes (382 ± 99 ms) and fellow eye (369 ± 79 ms) viewing. Similarly, for participants with strabismus only, reach latency was longer only during nondominant eye viewing (389 ± 103 ms) in comparison to both eyes (363 ± 77 ms) and dominant eye viewing (357 ± 71 ms). 

Accuracy.

There was no significant difference between the groups for reaching accuracy (i.e., constant error) along both the azimuth (F _2,43 = 0.28, NS) and elevation (F _2,43 = 1.05, NS) direction. Accuracy also was similar between the groups in the different viewing conditions along azimuth (F _4,86 = 0.93, NS) and elevation (F _4,86 = 1.98, NS). In contrast, radial error (unsigned overall error) was significantly greater for participants with amblyopia viewing with the amblyopic eye in comparison to the other viewing conditions, as well as to the other groups in all viewing conditions (F _4,86 = 4.58, P = 0.002; Fig. 1a). There were no significant differences in accuracy between groups when reaching to targets at different locations across the viewing conditions. 

Figure 1

 

(a) Overall reaching accuracy across viewing conditions. Participants with strabismic amblyopia had significantly worse accuracy when viewing with the amblyopic eye in comparison with the other viewing conditions and the other groups (P < 0.01). (b) Overall reaching precision across viewing conditions. Participants with strabismic amblyopia had significantly worse precision when viewing with the amblyopic eye in comparison with the other viewing conditions and the other groups (P < 0.01).

Precision.

The overall variable error (i.e., the SD of radial error) was significantly greater in participants with amblyopia viewing with the amblyopic eye (F _4,86 = 4.49, P = 0.002; Fig. 1b). Furthermore, the reduced precision was significant along the azimuth (F _4,86 = 3.24, P = 0.016) and elevation (F _4,86 = 4.75, P = 0.002) directions. There was a main effect of Target Visual Feedback (F _1,43 = 23.46, P < 0.0001). Reaching movements were less precise when the target was switched off at the initiation of the reach in comparison to when the target was present throughout the trial in participants with amblyopia, participants with strabismus only, and visually normal participants, regardless of viewing condition. 

Reach Performance Measures: Kinematics

Total Movement Time.

The main effect of Group did not reach significance (F _2,43 = 2.74, NS). The interaction of Group and Viewing Condition also was not significant (F _4,86 = 0.39, NS). For visually normal participants, movement times were 521 ± 97 ms when viewing with both eyes, 534 ± 105 ms for dominant eye viewing, and 540 ± 106 ms for nondominant eye viewing. In comparison, participants with reduced binocularity had longer movement times (participants with amblyopia: both eyes 636 ± 161; fellow eye 639 ± 145 ms; amblyopic eye 633 ± 158 ms; participants with strabismus only: both eyes 628 ± 164; dominant eye 655 ± 221 ms; nondominant eye 644 ± 199 ms). 

Total movement time was significantly longer for all groups when visual feedback of the target was present (i.e., target ON condition) as compared with when visual feedback was absent (i.e., target OFF condition) (F _1,43 = 34.87, P < 0.0001). For visually normal participants, movement time increased from 522 ± 99 ms to 541 ± 108 ms when the visual target remained visible during the reaching movements. In participants with amblyopia, total movement time increased from 619 ± 155 ms to 653 ± 152 ms when visual feedback was present. Similarly for participants with strabismus only, movement time increased from 631 ± 193 ms in the target OFF condition to 653 ± 199 ms in the target ON condition. No other significant main effect or interaction was observed for total movement time. 

Acceleration Phase.

Group mean acceleration trajectories for visually normal participants, participants with strabismus only, and participants with amblyopia for the three viewing conditions are shown in Figure 2. Regardless of viewing condition, visually normal participants had significantly higher peak acceleration (11.6 ± 5.2 m/s²) in comparison to participants with amblyopia (7.95 ± 3.3 m/s²) and participants with strabismus only (8.3 ± 3.9 m/s²) (F _2,43 = 3.63, P = 0.033; Figs. 2, 3). 

Figure 2

 

Group mean acceleration trajectory (solid line) and the corresponding SD (dashed lines) for visually normal participants (a–c), participants with strabismus only (d–f), and participants with strabismic amblyopia (g–i) when they reached to the 10° target. Viewing was with both eyes (top: [a, d, g]), monocular with the fellow/dominant eye (middle: [b, e, h]), or monocular with the amblyopic/nondominant eye (bottom: [c, f, i]). In all three viewing conditions, participants with strabismus only and strabismic amblyopia had lower mean peak acceleration and a prolonged duration of the acceleration phase (indicated by the delayed zero-crossing), when compared with visually normal participants.

Figure 3

 

Mean reach peak acceleration across viewing conditions. Visually normal participants had significantly higher peak acceleration in comparison with participants with strabismic amblyopia and participants with strabismus only in all viewing conditions (P < 0.05).

Figure 2 also shows that participants with amblyopia and participants with strabismus only attained peak velocity later in comparison to visually normal participants (i.e., delayed zero-crossing on the acceleration trace). Statistical analysis confirmed that the mean duration of the acceleration phase was significantly shorter in visually normal participants in comparison to both patient groups, regardless of viewing condition (F _2,43 = 4.25, P = 0.021). As shown in Figure 4, the mean duration of the acceleration phase was 177 ± 57 ms for visually normal participants, 223 ± 79 ms for participants with amblyopia, and 233 ± 63 ms for participants with strabismus only. 

Figure 4

 

Mean duration of the acceleration phase across viewing conditions. Visually normal participants had a shorter duration of acceleration phase in comparison with participants with strabismic amblyopia and participants with strabismus only in all viewing conditions (P < 0.05).

The correlation between the duration of acceleration phase and peak acceleration was significantly higher for visually normal participants compared with both patient groups across all viewing conditions (F _2,43 = 4.73, P = 0.014). For all groups, higher peak acceleration was associated with a shorter acceleration phase. The mean correlation coefficient for visually normal participants was −0.72 ± 0.10 when viewing with both eyes, −0.66 ± 0.11 during dominant eye viewing, and −0.67 ± 0.12 during nondominant eye viewing. For participants with amblyopia, the correlation was −0.53 ± 0.20 when viewing with both eyes, −0.57 ± 0.17 during fellow eye viewing, and −0.60 ± 0.13 for the amblyopic eye. For participants with strabismus only, the correlation was −0.61 ± 0.16 when viewing with both eyes, −0.57 ± 0.22 during dominant eye viewing, and −0.55 ± 0.19 for the nondominant eye. 

Peak Velocity.

There was no significant difference (F _2,43 = 2.24, NS) in mean peak velocity between visually normal participants (1.00 ± 0.31 m/s), participants with amblyopia (0.81 ± 0.19 m/s) and participants with strabismus only (0.83 ± 0.30 m/s). 

Deceleration Phase.

There was no significant difference (F _2,43 = 0.37, NS) in mean peak deceleration between visually normal participants (5.5 ± 2.5 m/s²), participants with amblyopia (4.5 ± 2.4 m/s²), and participants with strabismus only (5.2 ± 4.0 m/s²). There was also no significant difference (F _2,43 = 0.96, NS) in the duration of the deceleration phase between visually normal participants (355 ± 90 ms) and participants with amblyopia (413 ± 139 ms) or participants with strabismus only (410 ± 166 ms). 

The duration of the deceleration phase, however, was affected by visual feedback in all three groups (F _1,43 = 42.47, P < 0.0001). The duration of the deceleration phase was approximately 20 ms longer when target remained visible during the reaching movement, as compared with when the target was switched off at the initiation of the reach. 

Relation Among Peak Acceleration, Duration of Deceleration Phase, and Reach Error

The standardized regression coefficients (β), the associated P value for peak acceleration, and the duration of the deceleration phase are shown in the Table. In visually normal participants, only the duration of the deceleration phase was associated with overall reach precision, but not the peak acceleration. Specifically, longer duration of the deceleration phase was associated with smaller endpoint reach error. In contrast, the duration of the deceleration phase was not significantly related to reach precision in participants with amblyopia. Instead, better reach precision was associated with lower peak acceleration when viewing with both eyes or with the fellow eye. During amblyopic eye viewing, reach error was not significantly associated with peak acceleration or the duration of deceleration phase. For participants with strabismus only, the duration of the deceleration phase or peak acceleration were not significantly associated with reach error, which is evident from the inspection of the standardized coefficients. 

The Effects of Severity of Amblyopia and Stereopsis.

There was a significant difference in reach precision along the azimuth among participants with amblyopia in different viewing conditions (F _4,26 = 3.43, P = 0.022). Regardless of the severity of amblyopia, participants with negative stereopsis had reduced precision during amblyopic eye viewing (severe amblyopia 7.75 ± 3.6 mm; mild amblyopia 6.26 ± 2.1 mm) in comparison to the other viewing conditions (participants with severe amblyopia: both eyes viewing 4.84 ± 2.6 mm, fellow eye viewing 4.34 ± 1.9 mm; participants with mild amblyopia: both eyes viewing 5.57 ± 1.6 mm, fellow eye viewing 5.27 ± 1.9 mm) and participants with mild amblyopia and residual stereopsis (amblyopic eye viewing: 5.16 ± 2.1 mm; both eyes viewing: 3.97 ± 1.3 mm, fellow eye viewing: 4.89 ± 1.9 mm). Reach precision was not significantly different between groups for different target locations across viewing conditions. 

There were no significant differences between participants with amblyopia across stereoacuity and visual acuity deficit levels for any other reach kinematic variables. 

Discussion

We examined the effects of impaired spatiotemporal vision on visually guided reaching movements in adults with strabismic amblyopia and compared them to adults with strabismus only and visually normal participants. The following are the major significant findings: (1) participants with strabismic amblyopia had longer reach latency during amblyopic eye viewing in comparison to viewing with both eyes and the fellow eye; (2) precision of reaching was reduced during amblyopic eye viewing in participants with strabismic amblyopia and negative stereopsis; (3) participants with strabismic amblyopia and those with strabismus only had lower peak acceleration and a longer duration of the acceleration phase in all viewing conditions, whereas peak deceleration and the duration of the deceleration phase were not significantly different; (4) participants with strabismic amblyopia reduced peak acceleration to attain better reach precision; and (5) in all participants, the absence of visual feedback of the target led to a small decrease in precision, a shorter total movement time, and a shorter duration of the deceleration phase. 

Effects of Amblyopia on Reach Outcome Measures

Latency.

We found that participants with strabismic amblyopia had longer reach latency during amblyopic eye viewing in comparison to both eyes and fellow eye viewing. This is in contrast to participants with anisometropic amblyopia who had reach latency comparable to visually normal participants across all viewing conditions. ⁴⁴ This difference is most likely due to a greater central suppression of the deviated (amblyopic) eye in strabismic amblyopia. ³⁶ It is likely that the suppressive mechanism is stronger in participants with strabismus so as to avoid confusion and diplopia. Because the visual axes of the two eyes do not align, different images fall onto the fovea of each eye, which causes visual rivalry and confusion. ⁵⁷ Over time, the signal from the deviated eye becomes strongly suppressed during normal binocular viewing. One of the consequences of this long-term suppression is that presenting a comparable visual stimulus to the deviated and nondeviated eye produces significantly reduced activity in the afferent sensory pathway. ^58,59 Another factor that may lead to increased response latency is positional uncertainty, which has been reported consistently in patients with strabismus during nondominant/amblyopic eye viewing. ^{33,35,37,38,60} In both cases (i.e., reduced activation of the afferent pathway and increased positional uncertainty), it takes the signal longer to reach the threshold for response initiation ⁶¹ ; hence, response latency is prolonged during monocular viewing with the amblyopic/nondominant eye. 

Our results are consistent with a previous report ⁶² that examined manual responses to a centrally presented visual stimulus in participants with strabismus with and without amblyopia. The authors ⁶² found that manual reaction time was significantly delayed in participants with amblyopia during amblyopic eye viewing. However, our results differed from those of this previous study ⁶² in two regards: (1) whereas the previous study showed a significant correlation (r = 0.68) between reaction time and amblyopic eye visual acuity, in our current study, the correlation between reach latency and visual acuity did not reach statistical significance; and (2) whereas the previous study found that participants with strabismus only had similar reaction times during dominant or nondominant eye viewing, our data showed a significantly prolonged latency during nondominant eye viewing only. The discrepancy between the current and a previous study might be related to the task: participants in the previous study made a manual response by pressing a button, whereas participants in our study performed a reaching movement. Thus, the task we used might provide a more sensitive measure of the effect of strabismus on goal-directed behavior. 

Accuracy and Precision.

We found that the overall accuracy (radial error) and precision were worse in participants with strabismic amblyopia during amblyopic eye viewing in comparison to participants with strabismus only or visually normal participants during nondominant eye viewing. However, further analysis showed that reach precision was reduced only in participants with amblyopia and negative stereopsis, independent of the visual acuity deficit. We have previously reported a similar pattern of deficits for saccadic eye movements. ⁴³ Together, these results suggest that participants with strabismic amblyopia and negative stereopsis have a global deficit in detecting and localizing visual targets during amblyopic eye viewing that affects both their saccadic and visually guided limb movements. 

Surprisingly, no differences in reach accuracy or precision were found in participants with strabismus only in any viewing conditions. There was also no difference between participants with strabismus only who had negative stereopsis (n = 9) and those who had residual stereopsis (n = 5). These results suggest that regardless of their binocularity status, participants with strabismus only are able to use the input from their nondominant eye to achieve normal reach precision. In contrast, the loss of stereopsis in participants with strabismic amblyopia has a significant impact on their reach precision: participants with strabismic amblyopia and negative stereopsis might experience more difficulty with visually guided reaching movements when viewing with the amblyopic eye despite having a relatively mild visual acuity loss. 

We have previously found that reach precision was reduced in participants with anisometropic amblyopia with severe acuity loss (i.e., >20/100). ⁶³ Because most of the participants with severe anisometropic amblyopia (5/7) also had negative stereopsis in our previous study, it was not possible to determine whether reduced reach precision was due to deficits in visual acuity, stereopsis, or both. The results from our current study shed more light on this issue and suggest that reach precision is reduced in participants with amblyopia and negative stereopsis, regardless of whether the visual acuity loss is mild or severe. 

A previous study that examined localization errors in participants with strabismic amblyopia reported large between-subject variability in bias and variable errors. ³⁵ Most patients (6/8) in that study had negative stereopsis and five of these patients exhibited reduced precision when viewing with the amblyopic eye. The other two patients with residual stereopsis showed no reliable difference in localization precision between the eyes. These findings are in agreement with our results. Despite methodological differences between the previous and current studies, the overall data provide support that patients with negative stereopsis have larger deficits in localization precision. 

Effects of Amblyopia on Planning and Feedback Control

Reach peak acceleration is a kinematic marker related to movement planning. Because peak acceleration occurs very early during the trajectory, typically less than 100 ms after movement onset, it is unlikely to be influenced significantly by sensory feedback. ^64,65 On the other hand, the duration of the deceleration phase has been associated with feedback control. ⁶⁶ Previous studies have shown that reach trajectory can be adjusted by using sensory feedback during the deceleration phase as the hand is approaching the target. ^67–70  

Our data showed that both amblyopia and strabismus affect the planning of visually guided reaching movements. Participants with strabismus with or without amblyopia had significantly lower peak acceleration and a prolonged acceleration phase in all viewing conditions in comparison with visually normal participants. We also found that peak acceleration was similarly reduced in all participants with amblyopia, regardless of visual acuity or stereoacuity deficits. In contrast, our data show no significant difference in the duration of deceleration phase between participants with amblyopia or strabismus and visually normal participants. All of these findings are similar to the pattern we have previously reported in anisometropic amblyopia. ⁴⁴ Because this pattern of movement planning is not unique to participants with amblyopia or strabismus (i.e., it is similar in participants with strabismic amblyopia, in participants with anisometropic amblyopia, and in participants with strabismus only without amblyopia), the development of this strategy could be linked to a shared characteristic among these participants: impairment in binocular visual function. It is probable that normal binocular vision is essential for the optimal development of fine motor skills. The process of learning new motor skills, including relatively simple reaching and grasping movements, consists of the acquisition of an internal model; that is, an internal representation of action in a specific environment. ^71–74 Binocular vision may be necessary for the precise calibration of the internal model for optimal motor performance. If binocularity is disrupted during early development, either in association with amblyopia or strabismus, participants develop similar compensatory strategies that persist into adulthood. 

Another important finding from our study is that participants with strabismic amblyopia or strabismus only did not alter the reach deceleration phase. Our results showed that all participants reduced the duration of deceleration phase where visual feedback of the target was switched off at movement initiation. Consequently, all participants had reduced reach precision in comparison to the condition in which target feedback was present throughout the trajectory. These results indicate that despite the acuity and stereoacuity deficits, patients still rely on visual feedback for guiding their reaching movements. 

Our results suggest that the reduction in peak acceleration is the main strategy that participants with amblyopia use to attain good endpoint accuracy and precision during reaching. This conclusion is also supported by the correlation analysis, which showed that reduced peak acceleration, but not the duration of the deceleration phase, is associated with increased reach precision. In contrast, the opposite association was found in visually normal participants: better reach precision was associated with a longer deceleration phase. Generally, higher peak acceleration results in lower endpoint precision due to signal-dependent noise, unless these errors are amended during the deceleration phase. ^75–77 As expected, our data showed that people with normal binocular vision tend to have higher accelerations, which means that they can use online control effectively to correct the potential trajectory errors and achieve good endpoint precision. It is possible that participants with amblyopia lower their peak acceleration because their ability to engage in online control is impaired. Future studies should examine this hypothesis by using an experimental manipulation of visual input during the ongoing movement. 

Supplementary Materials

pdf S1 

Acknowledgments

Supported by Grants MOP 89763 and MOP 57853 from the Canadian Institutes of Health Research, Leaders Opportunity Fund from the Canadian Foundation for Innovation, and the Department of Ophthalmology and Vision Sciences and the Research Training Centre at The Hospital for Sick Children. 

Disclosure: E. Niechwiej-Szwedo, None; H.C. Goltz, None; M. Chandrakumar, None; A.M.F. Wong, None