How Eye Dominance Strength Modulates the Influence of a Distractor on Saccade Accuracy

Abstract

Purpose: Neuroimaging studies have shown that the dominant eye is linked preferentially to the ipsilateral primary visual cortex. However, its role in perception still is misunderstood. We examined the influence of eye dominance and eye dominance strength on saccadic parameters, contrasting stimulations presented in the two hemifields.

Methods: Participants with contrasted eye dominance (left or right) and eye dominance strength (strong or weak) were asked to make a saccade toward a target displayed at 5° or 7° left or right of a fixation cross. In some trials, a distractor at 3° of eccentricity also was displayed either in the same hemifield as the target (to induce a global effect on saccade amplitude) or in the opposite hemifield (to induce a remote distractor effect on saccade latency).

Results: Eye dominance did influence saccade amplitude as participants with strong eye dominance showed more accurate saccades toward the target (weaker global effect) in the hemifield contralateral to the dominant eye than in the ipsilateral one. Such asymmetry was not found in participants with weak eye dominance or when a remote distractor was used.

Conclusions: We show that eye dominance strength influences saccade target selection. We discuss several arguments supporting the view that such advantage may be linked to the relationship between the dominant eye and ipsilateral hemisphere.

French Abstract

The sighting dominant eye (DE) is the one chosen when performing a monocular task. It is determined classically based on the “hole-in-the-card test,”¹ which provides a binary measure: left or right DE, according to the eye chosen by the participant for sighting through the hole in a piece of cardboard. However, it has been suggested recently that eye dominance could be assessed more precisely with binocular recordings.² Participants are categorized according to eye dominance strength (i.e., strong or weak eye dominance) based on the analysis of the peak velocity of saccades directed toward an isolated target. Indeed, participants exhibiting higher peak velocities toward the hemifield ipsilateral to the DE regardless of which eye is being measured are considered as having strong eye dominance, while participants exhibiting higher peak velocities toward the left hemifield with the left eye and toward the right hemifield with the right eye (i.e., standard nasotemporal asymmetry³) are considered as having weak eye dominance.² 

The dominant eye also has been studied with neuroimaging data, showing that it activates a greater part of the primary visual cortex (V1) than the nondominant eye.⁴ Other evidence^5,6 suggests that the V1 ipsilateral to DE is larger⁵ and more activated⁶ than the V1 contralateral to DE, suggesting a privileged relationship between DE and ipsilateral V1. Due to the crossing of the optical pathways, the V1 ipsilateral to DE initially processes information presented to the hemifield contralateral to the DE. Recently, whether such a relationship could lead to differences in the visuomotor processing of information from the hemifield ipsilateral or contralateral to DE has been examined.⁷ Using the Poffenberger paradigm (manual response to a target presented either in the left or right hemifield, using either the right or left hand⁸), participants exhibited faster reaction times when the target was presented in the hemifield contralateral to the DE than in the ipsilateral hemifield.⁷ The investigators suggest that this advantage of the hemifield contralateral to the DE over the ipsilateral hemifield is linked to the relationship between DE and ipsilateral V1. Indeed, this relationship would lead to a better perceptual processing in the hemifield contralateral than ipsilateral to the DE. Interestingly, in a subsequent study, the investigators found this advantage of the hemifield contralateral to the DE only in participants with strong eye dominance⁹ according to the peak velocity criterion.² The participants with weak eye dominance exhibited the standard Poffenberger effect (i.e., faster reaction times when the stimulation and hand are on the same side⁸), suggesting that the relationship between DE and ipsilateral V1 and the induced perceptual advantage of the hemifield contralateral to the DE occur only when participants have strong eye dominance. 

The aim of the present study was to further examine the relationship between DE and ipsilateral V1 and its role in perception and action mechanisms. To do so, we assessed the respective influence of eye dominance (left or right) and of eye dominance strength (strong or weak) on a saccadic task. Participants were instructed to make a saccade toward a lateralized target with a distractor presented simultaneously in the same or opposite hemifield. It is now well established that a distractor being presented close to the target position modifies saccade amplitude by deviating the saccade to an intermediate position between the two stimuli (global effect [GE]), whereas a distractor remote from the target position increases saccade latency (remote distractor effect [RDE]).^10–12 Therefore, we hypothesize that a modulation of both effects, depending on the hemifield in which the distractor is displayed, will reflect the influence of eye dominance and of eye dominance strength on saccadic parameters. Indeed, in participants with strong eye dominance the perceptual advantage of the hemifield contralateral to the DE should result in a greater impact of the distractor presented in this hemifield compared to the ipsilateral one on saccade amplitude and latency. Conversely, we expected no differences between the two hemifields in participants with weak eye dominance, as found in previous studies based on manual reaction times.⁹ Finally, another manipulation involved varying distractor luminance. It was made either as bright as or brighter than the target. Indeed, this manipulation is known to provide greater GE when the distractor is brighter than the target.¹³ Greater perceptual weight given to the distractor should differentially modulate the effects of eye dominance and eye dominance strength. 

Methods

Subjects

We divided 92 right-handed participants into four groups according to their eye dominance (left or right) and eye dominance strength (weak or strong) as defined by the analysis of saccade peak velocity.² This classification was made a posteriori after recording eye movement data (see Figs. 1, 2). Hand preference was determined by using the Edinburgh Handedness Inventory¹⁴ and eye dominance by using the hole-in-the-card test¹ repeated three times. 

Figure 1

 

Average differences of peak velocities of saccades toward isolated targets in left and right visual fields indicating strong eye dominance. Participants were categorized into two groups according to their eye dominance (left or right) measured with the hole-in-the-card test. Negative values indicate that saccades toward the left visual field exhibit higher peak velocities than saccades toward the right visual field, and positive values indicate the opposite. Those differences have been calculated for saccades made toward isolated targets presented at 3°, 5°, or 7° of eccentricity for the right (R eye) and left (L eye) eyes All participants presented in this graph exhibit higher peak velocities toward a same visual field regardless of the eye being measured for at least two of the three eccentricities tested. Therefore, they have been categorized as having strong eye dominance.

Figure 2

 

Average differences of peak velocities of saccades toward isolated targets in left and right visual fields indicating weak eye dominance. Participants were categorized into two groups according to their eye dominance (left or right) measured with the hole-in-the-card test. Negative values indicate that saccades toward the left visual field exhibit higher peak velocities than saccades toward the right visual field, and positive values indicate the opposite. Those differences have been calculated for saccades made toward isolated targets presented at 3°, 5°, or 7° of eccentricity for the right eye (R eye) and left (L eye) eyes. All participants presented in this graph exhibit higher peak velocities toward the right visual field with the right eye and toward the left visual field with the left eye (i.e., nasotemporal asymmetry) for at least two of the three eccentricities tested. Therefore, they have been categorized as having weak eye dominance.

All participants had reported normal or corrected to normal vision: 22 had a strong right DE (4 male, 18 female; mean age, 22.6 years old; SD, 6.41; mean laterality score, 79%; SD, 22.9%), 35 had a weak right DE (7 male, 28 female; mean age, 21.3 years old; SD, 2.32; mean laterality score, 81%; SD, 16.2%), 10 had a strong left DE (1 male, 9 female; mean age, 21.9 years old; SD, 3.11; mean laterality score, 77%; SD, 18.8%), and 25 had a weak left DE (7 male, 18 female; mean age, 23.8 years old; SD, 6.7; mean laterality score, 83%; SD, 20.9%). 

Participants gave their informed consent after an explanation of the procedure. The study adhered to the principles of the Declaration of Helsinki and the procedure was approved by the ethics committee of Paris Descartes University (Comité d'Evaluation Ethique en Recherche Biomédicale, IRB number 20130500001072). 

Stimuli

The initial central fixation was a 0.5° × 0.5° white cross. The saccade target and the distractor were both a 0.5° × 0.5° white circle. All were displayed on a medium gray background with a luminance of 4.5 cd/m². The fixation cross and the saccade target had a luminance of 27 cd/m² and the distractor luminance was either 27 cd/m² or 54 cd/m². 

Instruments and Eye Movement Recording

Stimuli were presented on an Iiyama (Nagano, Japan) HM240DT monitor with a refresh rate of 170 Hz and a resolution of 800 × 600 pixels. The experimental sessions took place in a dimly lit room. Subjects were seated 57 cm away from the screen and their heads kept stable with a chin and forehead rest. Movements of the two eyes were recorded with an Eyelink 1000 (SR Research, Ontario, Canada) sampled at 500 Hz and 0.25°. 

Each session began with a 9-point calibration filling the screen. Before each trial, a small circle was presented at the center of the screen to compare the actual eye position with the previous calibration. The participants had to fixate the circle and press a button on a pad. Trial began when comparison was detected successfully (see procedure). Online saccade detection corresponded to above-threshold velocity (30°/s) and acceleration (8000°/s²). 

Procedure and Design

Each participant ran four blocks of 165 trials for a total of 660 trials. The saccade target always was presented in the left hemifield in two blocks and in the right hemifield in the other two blocks. Thus, the uncertainty of target location was reduced by the hemifield blocked design to minimize the possible contribution to the distractor effect of decisional and strategy-based processes.¹¹ The order of the blocks was counterbalanced across subjects by alternating target side. 

Each trial of each session began with the presentation of a central fixation cross randomly displayed for 500, 600, 700, 800, or 900 msec. During this delay, the eye position was checked and if the distance between eye position and the center of the cross was greater than 0.75°, the trial was cancelled and repeated later in the session. The initial fixation cross disappeared simultaneously with the target appearance. In the no-distractor control conditions, the target was presented in isolation 3°, 5°, or 7° to the left or right of the fixation cross on the horizontal axis. In the target-distractor conditions, the target was presented at an eccentricity of 5° or 7° to the left or right of the fixation cross with a distractor presented at 3° of eccentricity in the same hemifield (testing the GE) or in the opposite one (testing the RDE). Participants were instructed to make a saccade either toward the isolated stimulus in case of no-distractor control condition or to the most eccentric stimulus in case of target-distractor condition. 

The whole experiment included 6 no-distractor control conditions (3 eccentricities * 2 hemifields) and 16 target-distractor conditions testing the GE or the remote distractor effect (2 target eccentricities * 2 target hemifields * 2 distractor sides * 2 distractor luminances). Each condition included 30 trials. In each block, the target hemifield was held constant, whereas all other conditions were intermixed. 

Data Analysis

Saccade latency and amplitude were measured. In the no-distractor control conditions, peak velocity of the rightward and leftward saccades also was measured to classify participants according to eye dominance strength.² In the target-distractor conditions, we also derived two standard additional measures to examine the effect of the distractor on saccadic behavior. The RDE corresponds to the average saccade latency difference between a given experimental condition and its corresponding control condition when the target is displayed at the same eccentricity with no distractor. The global effect percentage^15,16 (GEP) was used to examine the deviation of the saccade endpoint induced by the distractor. The GEP was calculated using the following formula: GEP = 100 * ([A_{3° + 5° or 7°} − A_3°]/[A_{5° or 7°} − A_3°]), where A_3° is the average saccade amplitude to targets presented in isolation at 3°, A_{5° or 7°} is the average saccade amplitude to targets presented in isolation at 5° or 7°, and A_{3° + 5° or 7°} is the average amplitude of saccades evoked by target-distractor pairs. A GEP of 0% means that the saccade landed on the distractor position (maximal GE) and a GEP of 100% that the saccade landed on the target position (no GE). In other words, the higher the GEP, the lower the GE. All analyses were run using data from both eyes separately. As our hypotheses do not involve any differences between saccadic parameters of the left and right eyes, and as we, indeed, do not report such differences, we here presented only the data from the right eye. 

Results

Preliminary Analyses

Trials with blinks (less than 0.01%) and with latency, amplitude, or peak velocity outliers diverging from individual distributions (0.06%) were discarded from further analyses. A preliminary analysis showed there was an effect of the block rank on saccade latency (F_[3,315] = 8.86, P < 0.001). Latency was longer on the first block than on the other three blocks. Therefore, to keep the number of saccades executed to the left and right hemifields constant, the first and fourth blocks were removed from following analyses. 

Before analyzing the data on the derived measures to examine the effect of eye dominance and eye dominance strength on the GEP and RDE, we conducted a 2-fold preliminary analysis. We checked that the distractor presented in the hemifield opposite the target increased saccade latency. We also checked that the distractor presented in the same hemifield as the target deviated saccade amplitude. Average saccade latency and amplitude obtained for each condition are presented, respectively, in Tables 1 and 2. 

Table 1

 

Average Latencies (and Standard Deviations) in Milliseconds

 

Table 2

 

Average Amplitudes (and Standard Deviations) in Degrees

 

We ran ANOVAs on the average saccade latency (Table 1) and on the average saccade amplitude (Table 2) with eye dominance and eye dominance strength as between-subject factors and distractor condition (no distractor, distractor as bright as the target, brighter distractor), saccade target eccentricity (5° or 7°) and target presentation hemifield (left or right) as within-subject factors. Note that regarding the average saccade latency, the no-distractor control conditions were compared to the opposite-hemifield distractor conditions. Regarding the average saccade amplitude, the no-distractor control conditions were compared to the same-hemifield distractor conditions. 

A main effect of the distractor was found on saccade latency (F_[2,176] = 255.52, P < 0.001) as well as on saccade amplitude (F_[2,176] = 505.67, P < 0.001). As expected, compared to the no-distractor condition, the presentation of a distractor in the opposite hemifield simultaneously with the target induced an increase of 19 ms of saccade latency (F_[1,88] = 341.23, P < 0.001), whereas the presentation of a distractor within the same hemifield induced a deviation of the saccade of 0.8° closer to the distractor (F_[1,88] = 341.23, P < 0.001). Thus, RDE and GE were well observed in our experiment. 

Main Results

We then conducted several analyses on the derived measures to examine whether the effects of the distractor could be modulated by eye dominance and eye dominance strength. We expected greater effects of the distractor located in the hemifield contralateral to the DE in the case of strong eye dominance and no difference between the two hemifields in the case of weak eye dominance. Average saccade latency difference (remote distractor effect) and average global effect percentage are presented, respectively, in Tables 3 and 4. 

Table 3

 

Average Remote Distractor Effect (and Standard Deviations) in Milliseconds

 

Table 4

 

Average Global Effect Percentage (and Standard Deviations)

 

Two ANOVAs were conducted on the average saccade latency difference (Table 3) and the GEP (Table 4) with eye dominance (left or right) and eye dominance strength (strong or weak) as between-subject factors and saccade target eccentricity (5° or 7°), hemifield of target presentation (left or right), and distractor luminance (same or brighter than the target) as within-subject factors. Concerning the average saccade latency difference (Table 3), we found no main effect or interaction between factors (all P > 0.10) with the exception of a significant effect of the distractor luminance (F_[1,88] = 4.57, P < 0.001). Saccade latency increased very slightly with a brighter distractor (1.5 ms on average). Data on the GEP (Table 4) indicated a significant effect of target eccentricity (F_[1,88] = 131.95, P < 0.001): GE was higher with shorter target-distractor distance (difference of 11.5%). Distractor luminance also significantly affected GE (F_[1,88] = 9.82, P < 0.005): GE was higher with a brighter distractor with a very slight difference (1.4%). We found no main effect either of eye dominance (F < 1) or eye dominance strength (F_[1,88] = 1.05, ns). However, a main effect of the hemifield of presentation was found (F_[1,88] = 7.73, P < 0.01), the deviation of the saccade toward the distractor being greater in the left (69.1%) than in the right (71.6%) hemifield. More interestingly for our purpose, such an effect interacted with eye dominance and eye dominance strength (F_[1,88] = 8.86, P < 0.005). Figure 3 presents this interaction between eye dominance (left or right) and hemifield (left or right) in participants with strong (Fig. 3a) and weak (Fig. 3b) eye dominance. The effect of the hemifield of presentation did not reach the significance threshold for people with weak eye dominance (F_[1,58] = 2.975, P < 0.10) regardless of their DE (F < 1), whereas it was amplified in participants with a strong left DE, the saccade being more deviated toward the distractor presented in the left than in the right hemifield (62.4% vs. 75.4%; F_[1,9] = 11.92, P < 0.007). Participants with a strong right DE seemed to show the reverse effect, with a distractor impact greater in the right than in the left hemifield, but the difference failed to reach the significance threshold (Fig. 3a; F_[1,21] = 2.92, P < 0.10). However, it should be noted that an effect of eye dominance strength was found in the right hemifield in participants with a right DE (F_[1,55] = 3.87, P < 0.05) with the distractor effect being greater in participants with strong (66.9%) than with weak (73.4%) eye dominance. 

Figure 3

 

Interaction between eye dominance strength, eye dominance, and visual field on GEP. (a) Interaction between eye dominance (L, left DE; R, right DE) and visual field (LVF, left visual field; RVF, right visual field) in participants with strong eye dominance. (b) The same interaction in participants with weak eye dominance. *Significant (P < .05). ≈Differences that failed to reach significance (0.05 < P < 0.10). Error bars: Standard errors.

Discussion

Measuring Eye Dominance Strength: The Peak Velocity Criterion

Analyses of saccade peak velocities have been shown useful to estimate eye dominance strength based on binocular recording of eye movements made toward an isolated target.² Accordingly, participants exhibit higher peak velocities toward the hemifield ipsilateral to the DE in case of strong eye dominance and exhibit a nasotemporal asymmetry³ in case of weak eye dominance.² However, note that 2 of the 18 participants in the 2012 study exhibited higher peak velocities toward the hemifield contralateral to DE whichever eye they used. In the present study, when we categorized the 92 participants according to eye dominance strength, we noticed that those with strong eye dominance also did not systematically exhibit higher peak velocities toward the hemifield ipsilateral to the DE (see Fig. 1). Indeed, 37.5% (12/32) exhibited higher peak velocities toward the hemifield contralateral to the DE. However, the results on the GEP for those 12 participants matched the patterns observed in their eye dominance groups as defined by the hole-in-the-card test, with lower GE (i.e., higher GEP) in the hemifield contralateral than ipsilateral to the DE. Therefore, these results on GEP as well as the different patterns of peak velocities in the two studies suggest that the criterion for strong eye dominance should finally be to exhibit higher peak velocities toward the same hemifield (left or right) with both eyes, and not only toward the hemifield ipsilateral to the DE. 

Distractor Luminance

To manipulate the perceptual weight of the distractor, the distractor was either as bright as or brighter than the target. We did not find a strong modulation of the distractor effect, neither for the remote distractor effect nor for the global effect. In remote distractor effect conditions, we observed an only very slight effect of distractor luminance on saccade latency, but no interaction with eye dominance, eye dominance strength, or hemifield. A very slight effect of distractor luminance also was found on the GEP. Overall, the manipulation of distractor luminance we used appeared not to be important enough to modify the pattern of results depending on eye dominance and eye dominance strength. 

Remote Distractor Effect

A distractor displayed in the hemifield opposite the target hemifield produced a RDE. However, neither eye dominance nor eye dominance strength modulated this effect. Unlike saccade amplitude or saccadic peak velocity,² the presence of asymmetries on saccade latency is unclear in the literature: some studies reported average left–right asymmetries,^17,18 while others did not.^19,20 However, these studies never took into account eye dominance or even manual laterality. Very few studies have evaluated the effect of eye dominance but again without consistent results.^2,21,22 The present study tested these asymmetries on a large sample of participants, and failed to find any left–right asymmetries on saccade latency. The fact that RDE did not differ between the two hemifields in participants with strong eye dominance suggests that eye dominance does not influence saccade latency, at least in the conditions we tested. 

Global Effect

When the distractor was in the same hemifield as the target, our results showed that the distractor had more impact on saccade amplitude (GE) when presented in the hemifield ipsilateral than contralateral to the DE. This was true only in participants with strong eye dominance. However, this contrasts with our assumption based on findings involving the presentation of a unique stimulus.^2,7,9 Interestingly, the presentation of two stimuli, one of which is the saccade target as used in the present study, specifies the perceptual processing advantage of the hemifield contralateral to the DE, which would finally not occur in the overall hemifield, but would be restricted to the saccade target location. Therefore, in a saccadic task we suggest that the relationship between DE and ipsilateral V1 would lead to a more accurate selection of the saccadic target in this hemifield (i.e., smaller effect of the distractor on saccade amplitude) than in the ipsilateral hemifield. 

Note that the accurate selection of the saccadic target in the hemifield contralateral to DE for participants with strong eye dominance is hypothesized in light of the relationship between DE and ipsilateral V1, but V1 is only the starting point of the sensorimotor transformation. The signals then are transmitted to the parietal eye fields in the posterior parietal cortex and to the frontal eye fields, close to the precentral sulcus.^23,24 So, it remains open whether the relationship between DE and ipsilateral V1 then will lead to left–right asymmetries in parietal eye fields and frontal eye fields activations. Future neuroimaging studies could help to clarify this point, contrasting participants with left and right dominant eye, strong and weak eye dominance. 

However, our results showed a clear difference between the two hemifields in participants with a strong left DE, but the difference was slighter and did not reach the significance threshold in participants with a strong right DE. To explain this difference between those two groups we proposed that two phenomena are involved: On one hand, in participants with strong eye dominance, the relationship between DE and ipsilateral V1 would induce a more accurate selection of the saccadic target in the hemifield contralateral than ipsilateral to the DE. On the other hand, there would be an attentional bias toward the left hemifield giving more weight to the distractor due to the specialization of the right hemisphere for visuospatial attention.^25–28 Note that this attentional bias is hypothesized for all participants, and may explain that the distractor deviated saccade amplitude more when presented in the left than in the right hemifield in participants with weak eye dominance (see Fig. 3b). 

Figure 4 separately summarizes those two phenomena in participants with a strong left and right DE. In participants with a strong left DE, each phenomenon occurs separately in one hemifield and does not counteract the other one, leading to a great GEP difference between the two hemifields. Conversely, those two phenomena occur in the same hemifield in participants with a strong right DE. Moreover, the attentional bias that gives more weight to the distractor counteracts the accurate selection of the saccadic target. Accordingly, a smaller GEP difference between the two hemifields was found in this population. 

Figure 4

 

Illustration of the two phenomena inferred from our results on the GEP. Black indicates the relationship between DE and ipsilateral V1, leading to a more accurate saccadic selection in the visual field contralateral than ipsilateral to the DE. This phenomenon occurs in opposite visual fields in participants with a strong left DE (a) and with a strong right DE (b). Gray indicates the second phenomenon, an attentional bias toward the left visual field due to the right hemisphere specialization for visuospatial attention, giving more weight to the distractor in this visual field than in the right one.

Conclusions

Researchers now can measure precisely participants' handedness based on questionnaires assessing a percentage of handedness. However, eye dominance still is evaluated based on binary measures. Much research has been done to develop a more graduated measure of eye dominance.^2,9,29–35 We here show different visuomotor influences of eye dominance according to eye dominance strength. Moreover, the use of two stimuli helped to specify the link between DE and ipsilateral V1 (previous studies used a simple target stimulus^2,7,9). Indeed, the better processing that it involves in the hemifield contralateral to the DE seems not to operate in the whole hemifield, but seems restricted to the saccade target location. These findings point out the importance of taking into account participants' eye dominance and eye dominance strength in further visual or visuomotor studies. 

Acknowledgments

The authors thank the editorial board member and two anonymous reviewers that helped them to improve their paper. 

Disclosure: J. Tagu, None; K. Doré-Mazars, None; C. Lemoine-Lardennois, None; D. Vergilino-Perez, None