Gaze stability in the elite athlete: A normative observational study

1 Introduction

The vestibulo-ocular reflex (VOR) is a physiological mechanism that allows stabilization of the visual field while the head is in motion due to involuntary movement of the eyes equal and counter to the movement of the head [3]. VOR impairment can be seen in many injuries and illnesses and can affect how clearly one can perceive objects when the head is in motion. Functional consequences of VOR impairment include difficulty focusing vision, dizziness, oscillopsia, and postural instability [24].

To determine presence and severity of VOR impairment, clinicians may use objective testing such as the dynamic visual acuity (DVA) test. DVA testing can be performed via manual clinical assessment with the use of a Snellen eye chart or through a computer-based system [4]. The dynamic visual acuity testing process begins by determining the smallest optotype the patient can see clearly while his or her head is still (static visual acuity, SVA). This value is then compared to the smallest optotype that can be perceived while the head is moving at a specified velocity. In the literature, the most common threshold for head velocity when testing DVA is 120 deg/second [5, 18]. Computerized DVA testing has been found to have good reliability, sensitivity, and specificity. One study that compared matched controls to patients with vestibular deficits found that DVA testing had a positive predictive value of 96.3%, a negative predictive value of 93%, and an overall accuracy of 94.8% [6]. Excellent sensitivity, specificity, and positive and negative predictive values were also found with horizontal DVA in children [15]. Good test-retest reliability has been established in an athlete population for head rotation in the yaw and pitch planes [10], and normative values across a variety of populations have been published for children [15] and older adults [1]. A computerized test of dynamic visual acuity was developed for use as part of the National Institutes of Health Toolbox and provides normative data for persons aged three to 85 years [14].

The gaze stabilization test (GST) is also used to evaluate VOR function, though does so in a slightly different manner. Where the DVA measures VOR function at a set head velocity (120 deg/second), the GST assessment measures the highest head velocity at which a patient can accurately see a specified target before losing visual acuity. With GST, the clinician first determines how clearly the patient can see with their head still (SVA), then establishes the minimum amount of time that the target of the size determined via SVA can be shown for the subject to correctly perceive its orientation (visual perception time test, VPT). When testing GST, the VPT must be 70 ms or faster to ensure accurate results when the head is moving during testing [23]. For gaze stability testing, the head is moved either actively or passively at incrementally increasing velocities until the patient can no longer perceive the target clearly. This velocity data can be captured in any axis of head rotation (yaw, pitch, or roll), and when testing in the yaw plane, the highest velocity can be measured in both right and left directional movements. Gaze stability testing has been found in prior research to have adequate test-retest reliability in the yaw plane in an athlete population [10] and in a general adult population [25].

Most studies reporting GST values in a healthy population have used a community-based sample. For example, Ward and others [25] examined GST values in an older adult (mean age = 76.3) and younger adult (mean age = 25.2) community sample. Analyses revealed GST values of 155 deg/sec in the yaw plane for adults under 60 and 118 deg/sec in the yaw plane for adults over 60. Honaker and Shepard [9] also found similar age-related differences where younger adults (ages 20–39 years) performed better than older adults (60–79 years) with GST values in the yaw plane of 155 and 115 deg/sec, respectively.

Very few studies on dynamic visual acuity or gaze stability have focused solely on an athletic population. Visual demands of athletes will vary by sport, position, and level of competition [10, 20]. The ability to quickly analyze complex visual information, precisely identify relevant targets, and filter irrelevant distractions lends itself to increased athletic success. These complex visual demands often occur in sport while both the athlete and his or her surroundings are in motion. Successful execution of such actions is at least partially dependent upon the visual-vestibular system. When playing sports, athletes are constantly moving their heads, scanning their environment and often watching moving targets (balls, hockey pucks, other athletes). The ability to do this at an elite level requires a high functioning vestibular ocular reflex. Because the environment in elite athletics is so dynamic, it is important to determine the function of the VOR at these higher levels of activities.

Athletes of all ages, particularly those who participate in contact or high-risk sports, are at increased risk of injury relative to the general population. Injuries to the head and face may cause a transient disturbance in visual-vestibular function. In sports related concussion (SRC), it has been shown that vestibular symptoms are present in many athletes which are commonly caused by impairments of the function of the VOR [7, 11]. Dizziness and visual dysfunction after SRC can be pathognomonic for underlying vestibular and ocular motor system impairment and may delay recovery from an injury [12]. Thus, it is important to objectively assess potential impairment to these systems to better understand injury severity and monitor recovery of function. However, to understand recovery in the elite athlete population, it is important to first establish baseline measurements of how well they function prior to injury.

Searches of the literature revealed three studies that have examined gaze stability in an athlete population. A study by Kaufman et al. [10] found no differences between collegiate and high school football players on any DVA or GST measure, except for superior GST in the pitch plane for collegiate athletes. In this study, players in skilled positions outperformed players in unskilled positions for both DVA and GST in the yaw (but not pitch) planes. Another study that examined GST in football players [8] found no differences in GST between collegiate football players with and without a history of concussion. Similarly, a study comparing nine elite and nine less experienced motor sport athletes found a trend of better SVA, GST, and DVA performance in the elite group, though differences were not statistically significant [17].

2 Methods

3 Results

Descriptive statistics were used with continuous normally distributed variables reported as means and standard deviations and non-normal variables reported as medians and ranges. GST in the leftward direction compared with that in the rightward direction were analyzed using a paired t-test. GST among player positions was compared using ANOVA. All other comparisons between various groups were performed using either Mann-Whitney U tests or Kruskal-Wallis tests. For significant results, effect sizes were calculated. For normally distributed data, Cohen’s d was used. For non-normal data, r was calculated (Z/√N). In contrast with Cohen’s d, where effect sizes are small (0.2), medium (0.5) and large (0.8), r effect sizes are interpreted as small (0.1), medium (0.2) and large (0.3). A two-tailed p < 0.05 was considered significant. IBM SPSS Statistics for Macintosh, Version 25.0. Armonk, NY: IBM Corp. was used for the analysis.

All athletes were able to complete GST. Median SVA for the sample was – 0.1500 logMAR (range – 0.30– 0.20), and median VPT was 30.00 ms (range 30.0–52.5), with all athletes able to perceive the target faster than the 70 ms threshold required. The sample produced a mean GST score in the leftward direction of 208.62 (SD = 38.87, range = 130–335, Cohen’s d = 0.27) and in the rightward direction of 220.29 (SD = 44.42, range = 125–365). Athletes performed significantly better in the rightward direction (t = 3.173, p = 0.002). Table 1 offers percentile distributions for GST in both directions.

There was variability in the consistency index among the sample, with 63.5 % of athletes producing an optimal consistency (high CI) rating of 3 or 4 in the leftward direction and 57.4% in the rightward direction. CI ratings of 1 or 2 (low CI) suggesting suboptimal but valid performance were found in 36.5% and 42.5% of the sample in the leftward and rightward directions, respectively. Although the low CI group required more trials to establish the GST value in leftward (p = 0.004, r = 0.25) and rightward (p < 0.001, r = 0.49) directions, the difference between GST values for the low CI versus high CI groups was not significant for GST-Left (p = 0.545) or GST-Right (p = 0.295).

Comparison of athletes who identified as primary English speakers versus primary non-English speakers revealed no differences in leftward GST (GST-Left) (p = 0.804) or rightward GST (GST-Right) (p = 0.494). Further, no differences were observed for VPT (p = 0.526), SVA (p = 0.488), or number of trials completed left (p = 0.178) or right (p = 0.523) by language group.

Athletes were also classified by primary position (pitcher, catcher, infield, outfield) and by handedness (batting and throwing) for comparison. No significant differences by position were noted for GST-Left (p = 0.664), GST-Right (p = 0.951), VPT (p = 0.126), SVA (p = 0.110), number of trials for GST-Left (p = 0.220) or number of trials for GST-Right (p = 0.545). Twelve percent of the sample were left-handed throwers. For batting, 21.6 percent of players were left handed, 70.1% were right handed, and 5.2% were switch hitters. Four athletes (3%) had missing data for both batting and throwing data. Switch hitters demonstrated higher GST-Left scores (mean = 265.71), than did right-handed hitters (mean = 215.85, p < 0.001, r = 0.43), and right-handed hitters had higher GST-Left scores than did the left-handed hitters (mean = 157.93, p < 0.001, r = 0.72). As well, switch hitters demonstrated higher GST-Right scores (mean = 265.71), than did right-handed hitters (mean = 223.93, p = 0.023, r = 0.29), and right-handed hitters had higher GST-Right scores than did left-handed hitters (mean = 190.17, p < 0.001. Cohen’s d = 0.36). Differences between right- and left-handed throwers were also observed for both GST-Left (p < 0.001, Cohen’s d = 0.68) and GST-Right (p = 0.010, r = 0.23) scores. For both scores, right-handed throwers had higher GST scores. GST-Left values were 152.5 for left handed throwers and 213.07 for right handed throwers. GST-Right scores were 191.88 and 222.40 for left- and right-handed throwers, respectively.

Although it is intuitive that athletes should have better dynamic visual acuity skills than a general population, there are few studies that have examined this group and even fewer that have specifically examined performance of elite-level athletes. Of existing studies, most have limited sample sizes. This is the first study to our knowledge that examines gaze stability function in a relatively large sample of elite athletes, particularly professional baseball players.

The purpose of the study was to establish normative GST data for an elite athlete sample. Our data revealed GST leftward and rightward values that far exceed normative values previously provided in studies of GST in an adult community sample. Specifically, our means of 208.62 deg/sec leftward and 220.29 deg/sec rightward are much higher than those observed in Ward et al.’s [25] community sample of adults with a mean age of 25 (155 deg/sec) and in Honaker and Shepard’s [9] community sample of adults ages 20–39 years of age (155 deg/sec). Relative to other athlete samples, our study sample had higher values than those reported in Honaker et al.’s [8] sample of collegiate football players (147 and 155 deg/sec leftward in athletes with and without concussion history; 155 and 150 rightward in athlete with and without concussion history, respectively). Observed values in our study’s elite athlete sample more closely approximated those of elite motorsport athletes (202 deg/sec leftward, 212 deg/sec rightward) [17].

The differences between those GST values observed in the current study and in prior studies should be interpreted with caution, given differences in equipment used to provide GST measurements in this study. Our study is the first to utilize the Bertec® Vision Advantage^TM (BVA) computerized system for GST. All other studies described above have used the Neurocom^® InVision product. Unfortunately, there are no published studies comparing GST measurements across these two systems. The BVA system was chosen over other systems for baseline procedures because the equipment is highly portable and adaptable to smaller spaces, which is important in professional (and non-professional) multimodal baseline testing sessions when limited space must be shared between numerous providers involved in pre-season screening, and athletes must be able to easily move between stations in an efficient manner.

It is also important to note that differences in testing procedures (active versus passive head movement; distance of test taker from screen) may also account for observed differences in GST values for this study. Active versus passive head movement in GST testing has been studied by both Tian [22] and Lee [13]. Both studies reported passive head movement as a better test of vestibular function than active head movement. Tian studied both normal (N = 15) and abnormal (N = 11) subjects with unilateral vestibular loss using the DVA test with a magnetic search coil system, finding that manually imposed head movements diminished extravestibular compensatory strategies. Lee found better test/retest reliability when assessing GST via passive head movements in a healthy sample using Neurocom^® InVision equipment. Similarly, our decision to employ the passive head movement technique was based on the notion that doing so would allow the athlete to focus only on the orientation of the optotype (target task), rather than having to also monitor head speed and head rotation distance while trying to complete the task. Table 2 summarizes differences in testing procedures and equipment in the studies discussed above.

We did not find performance differences of any kind between athletes who identify as English-speaking and those who identified as non- or limited-English speaking. Given that we relied on bilingual teammates to explain the task to their non-English speaking counterparts, then modified the responses by having the athlete point in the direction of the E (rather than stating the direction orally), we analyzed the data between the language groups to ensure that our modified procedures during the baseline session did not produce discrepant results. No differences were observed; thus, the combined normative data should be useful in evaluating and interpreting gaze stability in both populations. Normative data provided as percentile scores (see Table 1) should help the clinician understand how the athlete performs in relation to other elite athletes in situations where rehabilitation and/or vision performance training may be indicated.

In this sample, number of trials required varied by consistency index rating for both leftward and rightward directions. Perusal of the mean values across CI indicates that those athletes whose performance across trials was more consistent (high CI) required fewer trials to achieve valid GST scores. Because data were still considered valid, and were similar, across CI levels, CI may not be a useful indicator of test validity in an elite athlete population. However, this data provides the clinician with awareness that GST values with a low CI can still be useful in understanding gaze stability, but it would be expected that a greater number of trials may have to be performed by the clinician to achieve the valid GST score.

The relation of handedness to GST was examined for exploratory purposes only. We found that switch-hitting batters had superior GST scores to right- or left-handed batters, and that right-handed throwers demonstrated superior GST scores when compared to left-handed throwers. While this finding is interesting, it is important to note the small sample size for switch hitters and left-handed athletes. Thus, additional research would be needed in order to make any assumptions about generalizability or about etiology of these differences.

The current study has several limitations. First, it is a retrospective chart analysis; thus, many additional variables that might be acquired in prospective research were not available for analysis (e.g., detailed history of sport participation, history of vision problems, concussion, etc.). Although our sample size is the largest examining GST in athletes to date, the sample is restricted by age, gender, and sport. Given that dynamic vision has been noted to change with age in a community sample, it is likely that older athletes, even at the elite level, would perform differently on GST. Although gender differences have not been reported for GST in prior studies, it is important to understand gaze stability function in elite athletes of both genders. Further, the sample was restricted to baseball players competing at an elite level, which restricts generalizability to other elite sports. While other test platforms have established reliability of gaze stability testing, reliability data have not been published for GST testing on this specific equipment. As stated above, earlier studies have reported better reliability and reduced extravestibular compensatory strategy use for GST using passive versus active methodology, but no comparative studies have been completed using the Bertec® equipment.

Improving our understanding of the visual skills of elite athletes and the demands of competitive sports is an important step toward being able to provide excellent care to this group, whether designing a rehabilitation plan or sport performance enhancement protocol. Consistent with prior studies of dynamic vision in healthy athletes, our GST study results suggest that rehabilitating or training an athlete to “normal” by community sample standards is likely insufficient. Following an injury which affects gaze stability (e.g., concussion, ocular injury), a practitioner should be aware that achieving community based normative performance may mean the athlete is still unable to process visual stimuli at rate that is necessary to perform well in sport and to avoid further injury.

Although there is scant literature about improving vision performance and subsequently sport performance in athletes, there is some evidence that vision enhancement training may be beneficial [19], particularly in baseball athletes [2, 21]. However, evidence is mixed as to whether these learned vision skills transfer into other skills that may lead to better sport performance [19].

This study represents a first step in more clearly understanding high-level dynamic visual function in elite baseball athletes. Consistent with prior research, this group of athletes exhibited superior performance relative to normative scores for community samples. It is important for a clinician who provides care to injured elite athletes to understand these differences when tasked with helping to determine their readiness to safely return to sport activity. Understanding sport-specific normative data in elite athletes should also allow providers to optimize treatment strategies and goals if a player is injured, and potentially provide more accurate information to be used in multidisciplinary return to play decisions. Future research should expand our understanding of normal performance of elite athletes in additional sports and should explore potential sex differences in GST among elite athletes.