SPORTS VISION AND SPORTS VISION TRAINING
Vision is central to success in nearly all sports. Hitting a baseball, spiking a volleyball, and executing a bicycle kick all rely on vision to guide precise and difficult motor actions. Such perception-action coupling is often regarded as a limiting factor for excellence across a wide range of sports,1 and it all starts with the eyes. Transduction of light into the neural code relies on optics and retinal sampling to produce high spatial acuity and oculomotor dynamics that can maintain focus as the body and the field of view move rapidly over space. Once transduced into the neural code, information is sent through the visual system to the brain where it is combined to form 3D representations. These representations interact with attention, cognitive control, memory, and motor systems in feedforward and feedback circuits to enable perception and action—ultimately allowing for the exceptional feats of athleticism we witness when athletes perform at their best on the field.
Just as all sports entail different movement dynamics, it is thought that different visual abilities are essential to success. For example, the visual needs of a softball batter are quite different from the visual needs of a football quarterback. In one case, a batter must be able to project the trajectory of a pitch that unfolds over milliseconds to decide whether to swing and optimally contact a softball moving up to 70 miles per hour. In the other case, a quarterback must monitor teammates and opponents as a play unfolds over several seconds. In all cases, the consequences can be remarkable, with success leading to tremendous reputational and financial gains, and failures leading to humbling physical and mental results. Past research has supported this notion, demonstrating that different sports draw on different visual skills, as shown with Olympic athletes from different sports2 and among athletes who play interceptive versus strategic sports.3
A core tenet of this emerging “sports vision” discipline is that better visual abilities underlie better athletic performance. Past research has approached this question by identifying cross-sectional differences, either comparing subjects at different levels of athletic achievement or testing for direct correlations between visual assessments and on-field performance. Evidence has accrued supporting both associations. For example, numerous studies have demonstrated that visual-perceptual and visual-cognitive abilities are enhanced in expert athletes4,5 with evidence for superior visual acuity,6,7 enhanced contrast sensitivity,8 and better visual tracking abilities9 in expert athletes. Such findings have been synthesized in two meta-analyses of the sports expertise literature,10,11 showing that higher achieving athletes are better at detecting perceptual cues, make more efficient eye movements, and have better attentional processing compared with less accomplished athletes or nonathletes.
Research linking baseline assessments to game performance has also emerged in recent years, which attempts to directly compare ocular or psychometric tests of vision to measures of game performance. These measures include both specific statistics that isolate the individual (e.g., plate discipline measures in baseball and softball) and nonspecific statistics that more broadly reflect productivity in the context of other teammates or opponents (e.g., batting average, which also depends on actions of fielders and base runners). Although conducted with a diverse range of sample sizes, assessments, and outcome measures, this literature provides preliminary evidence that better visual-perceptual,12–14 oculomotor,15,16 and visual-motor17–19 skills correlate with better performance statistics in competitive games and, in some cases, reduced injury instances.20 Because baseline assessments can be measured before game statistics are accrued, these findings may offer a prospective view that can be considered predictive and therefore potentially useful for player scouting.
Collectively, these studies demonstrating associations between better vision and better sporting performance have provided support for the notion that improving visual skills may also lead to better athletic outcomes. The growing practice of “sports vision training” relies on the notion that practice with demanding perceptual, cognitive, or oculomotor tasks can improve the ability to process and respond to what is seen, which may confer an advantage when such visual skills are marshaled in challenging sporting environments. Although there is considerable heterogeneity in the way that sports vision training interventions have been implemented, the common philosophy behind these programs stems from three assumptions: aspects of vision are important for particular sports, these abilities can be modified through training, and improvements in these visual abilities translate to better on-field performance.21
Early approaches to sports vision training entailed analog drills that imposed heavy oculomotor demands, requiring trainees to rapidly alter visual accommodation and convergence, or saccadic and smooth pursuit eye movements, to moving or physically separated visual targets. Although several empirical studies22 and consensus reviews23 have cast doubt on the effectiveness of such analog training techniques, a wide variety of new technologies have been developed in recent years to substantially extend the range, scope, and contexts of sports vision training applications.24 For example, digital devices, such as stroboscopic eyewear and mobile tablet-based devices, now allow for training during natural sporting activities.25–27 Through the use of augmented and virtual reality simulation, it is possible to recreate and augment sporting contexts to promote sports-specific training with high fidelity.28,29 As described in the recent “Modified Perceptual Training Framework,”30 the effectiveness of these approaches for improving actual competitive performance may depend on interacting factors that define the targeted perceptual function that is being trained, the correspondence of the training stimulus to the desired competitive skills, and the correspondence of the elicited response to competitive skills. As such, effective sports vision training approaches involve a multidimensional set of relationships that must fit into a training program that is logistically feasible and tolerable for athletes who generally have considerable demands for their time.
Research across the gamut of sports vision has led to widely discrepant findings regarding the reliability and effectiveness of these approaches. There are many possible reasons for these discrepancies, including the specificity of the assessment or the training, the match between the approach and the sport under consideration, and the scientific or statistical methodologies used in conducting the study. The next section of this article will briefly address some of the important concepts regarding evidence-based science that are important for interpreting the sports vision literature but pose challenges when working with athlete populations. This is followed by a summary of each of the considered publications and a critical review of the strengths and weaknesses of this literature along with recommendations that may lead to improvements in future research and practice.
EVIDENCED-BASED SCIENCE AND BEST PRACTICES IN SPORTS VISION RESEARCH
As with any discipline, the practice of sports vision should be driven by the best available evidence. However, the opportunity to conduct carefully controlled and sufficiently powered research studies, especially in the case of elite athletes, is a challenge that requires considerable effort. Because of logistic hurdles, limits of available time, and the potential desire of athletes and teams to maintain a competitive advantage, there is a relative scarcity of opportunities to conduct well-designed studies that can be published for public consumption. Despite these challenges, the pursuit of well-designed and executed studies should remain the criterion standard in sports vision research. As an initial step, it is important to consider challenges to causal inference that may impact the sports vision literature. For a more complete treatment of human subjects research methods and sampling biases, see the text by Patten and Newhart.31
Causal inference refers to the process of drawing a conclusion about the connection between a condition and an effect. Good research designs are organized to allow for strong inference while minimizing sources of bias. Although randomized assignment of participants is considered among the criterion standards in causal research, studies with athletes pose a particular problem in this regard. For all intents and purposes, it is often not possible to randomly assign participants to treatment and placebo groups, or to different sports or levels of achievement, and therefore, there are selection processes that may be beyond experimental control. As such, it is important to consider selection biases that impact the interpretation and generalizability of findings. Broadly speaking, selection bias refers to the choice to analyze some data with regard to all possible data that may be available. Within the hierarchy of sports, one must also consider survivorship bias, a form of selection bias that specifically addresses concentration on people that have passed a selection criterion. Although these aspects of representative sampling pose challenges within the sports vision literature, it is important to encourage the interpretation of results regarding the specific sample under consideration and avoid, or acknowledge, potential sampling biases.
A second area of methodological challenge stems from the desire to test multiple relationships among the abundance of meaningful quantitative data generated in sporting activities. To infer relationships, null hypothesis tests are typically performed to determine the statistical likelihood of a result not having occurred by chance. Within this framework, type I errors and type II errors occur when incorrectly selected criteria are used to reject hypothesis. Type I errors are controlled by the level of statistical significance or certainty, typically 95% certainty or P < .05. With multiple analyses, however, the chance of a false-positive finding will be inflated, creating the appearance that a significant finding exists when indeed none is present. In such situations, it is necessary to adjust the significance level to account for the multiple tests being performed, so-called multiple comparison correction. Type II errors occur when one accepts a null hypothesis that is false. Such errors are related to the number of subjects enrolled in the study, with insufficient sample sizes making type II errors more likely. To have the best chance of rejecting the null hypothesis and of avoiding type II errors, studies must have enough subjects. Because of challenges in obtaining large sample sizes and the desire to test many aspects of performance simultaneously, studies addressing sports vision must take special precautions to avoid inflation of these statistical errors.
The biases listed previously largely reflect comparable challenges faced in other domains of human behavioral research32 and have sparked a movement toward “Open Science” principles that prioritize transparent, rigorous, and accountable research practices to promote accessible, verifiable, and valid findings. Driven partly by the tenets of clinical trials that permeate medical research, Open Science advocates for approaches such as pre-registration of hypotheses and methods before initiating studies, placebo control to isolate interventional effects, and open access to data and code to make the raw materials and outputs of studies available for others to replicate and build upon. Given the many methodological challenges noted previously, sports vision studies are advised to follow the Open Science principles. See the Appendix (available at https://links.lww.com/OPX/A496) for more on pre-registration and the Open Science approach.
In the following section, we review the literature that has tested the role of vision and/or vision training using sports performance statistics as outcome measures. To arrive at a representative sample of articles addressing vision assessment and vision training, we first included all references in the 2011 American Optometric Association, Sports Vision Section bibliography.33 These references were combined with results of keyword searches on PubMed for citations with the terms “sport” AND “vision,” “sports” AND “attention,” and “sports” AND “neuroscience,” performed on April 13, 2020. Forward and backward bibliographic searches using Google Scholar were performed on these articles and topical reviews in this field from Mann et al.,10 Voss et al.,11 Williams and Ford,34 and Appelbaum and Erickson.24
Articles were considered for review if they were written in English, described research with human subjects, and included statistical comparisons of quantitative visual and/or psychometric assessments with game performance measures obtained from competitive or simulated sporting events. Approximately 540 references were initially identified and screened. Title and abstract review eliminated most of these articles because they either did not include on-field sports performance metrics as outcome measures or did not include use of appropriate vision assessment or training interventions. The final reviewed sample included 13 articles that addressed correlations between visual assessments and game statistics, and 16 articles that addressed gains due to vision training using benchmarks from game, or game-like, statistical production. Although this process was designed to capture relevant articles, it was not intended to be exhaustive but instead create a representative sample of the relevant scientific literature.
Studies Linking Vision Assessment to Game Performance
A total of 13 publications that reported comparisons between visual assessments and sports performance were found. Of these articles, eight involved baseball athletes, two reported on basketball, and one each reported on hockey, marksmanship, and softball.
In an early study by Laby and colleagues35 published in 1998, 410 major and minor league players in the Los Angeles Dodgers baseball organization were assessed for ocular dominance using a modified Bryngelson technique (placing a tube over their eye to gaze at a distant target) and hand dominance. Eye-hand dominance was defined as uncrossed if the laterality matched for the eye and hand assessment or crossed if it did not. Dominance was then compared with batting and pitching performance using subsequent season game statistics. Results did not demonstrate significant associations between eye, hand, or eye-hand dominance and batting or pitching statistics in either the major or minor league samples.
The same year, Molia and colleagues36 published an article testing stereopsis in a sample of 23 collegiate baseball players that included 14 position players and 9 pitchers. In this study, a battery of functional vision tests was acquired, including Snellen acuity and eye alignment at distance, and three different near and far distance stereoacuity measures. These assessments were compared through separate Spearman correlations to the players' previous-year game statistics for batting average and slugging percentage among the position players and earned run average, out percentage, and strike-out percentage for the pitchers, including only batters with >50 at bats and pitchers with >9 innings pitched. Although results demonstrated better-than-average functional vision scores and correlations between stereoacuity measures, they did not demonstrate significant relationships with any of the game performance measures.
In 2011, Reichow and colleagues13 reported on results from a pilot study conducted with 20 members of the Pacific University baseball team to test if players' ability to correctly identify tachistoscopically presented pictures of a pitch was correlated with their batting averages from the previous season. In this study, players were presented with 30 randomly ordered slides depicting a pitcher throwing 1 of 4 different pitches, each presented for 200 milliseconds. A single Pearson correlation test was performed, which demonstrated a significant (P < .01) and strong (r = 0.648) positive correlation, with better average tachistoscope accuracy scores corresponding to higher season batting averages.
In 2016, Müller and Fadde37 reported on findings from a temporal occlusion study performed with 34 minor league baseball batters with at least 100 at bats from which to draw statistics. In this study, the authors compared players' accuracy in judging the pitch type (fastball, curveball, change-up) when occluded at four different time points relative to ball release (at pitcher's front foot impact, shoulders squared, ball release, and a no-occlusion control) against their on-field batting performance statistics including batting average, on-base percentage, slugging percentage, base on balls, strikeouts, and walk-to-strikeout ratio. Pearson correlation coefficients were calculated for each combination of occlusion time point (4) and game statistic (6) using only results from the fastball and change-up pitches and a criterion for statistical significance of P < .05, without correction for multiple comparisons. Results demonstrated significant positive correlations between task performance at the front-foot release point and base on balls (r = 0.35, P = .04) and significant positive correlations in the ball release condition with base-on-ball percentage (r = 0.37, P = .03) and on-base percentage (r = 0.37, P = .03).
To evaluate the role of eye-hand coordination on batting performance, Laby and colleagues19 reported on a 2018 study with 450 professional baseball players (105 major league) in which performance on a commercially available assessment system (Sports Vision Trainer System; Sports Vision PTY, Sydney, Australia) was compared retrospectively with career plate discipline metrics. Using Bonferroni-corrected Pearson correlations, the authors compared eye-hand performance from three different assessment modes (proactive, reactive, go/no-go) to individual player's career plate discipline measures for out-of-zone chase percentage, fastball chase percentage, in-zone swing percentage, in-zone fastball swing percentage, and at bats per base on ball, as well as players' highest achieved league level and years of major league service. Results demonstrated statistically significant correlations between most of the task scores and plate discipline metrics (P < .001), as well as longer service in, and likelihood to achieve, the major league level.
The same year, Burris and colleagues17 made use of a naturalistic sample of data collected on the Nike (Beaverton, OR) SPARQ Sensory Station, a normative battery of nine validated tasks,38–40 to evaluate the links between visual-motor performance and batting performance in a sample of 252 professional baseball players (141 batters with >30 at bats and 111 pitchers with >30 innings pitched). Using a Bayesian hierarchical modeling approach that allowed for comparison across players from different leagues and contrast to a baseline model, task performance on the Sensory Station was compared with subsequent season game statistics for on-base, walk, strikeout, and slugging percentages for batters and fielder-independent pitching for pitchers. Results demonstrated that, compared with the base model including the player age and position as control variables, a full model including performance on the task battery produced probabilistically (outside 95% confidence interval) better prediction of on-base percentage, walk rate, and strikeout rate, with better performance on the task battery corresponding to better performance on the field.
In 2019, Laby and colleagues14 tested the relationship between performance on a functional vision screen (Enhanced Vision Testing System) that combines target size, target contrast, and presentation time with individual career plate discipline batting statistics in a sample of 475 professional baseball players. Results demonstrated statistically significant Spearman (rank order) correlations between performance on the visual evaluations and plate discipline measures including in-zone swing percentage, in-zone fastball swing percentage, chase percentage, fastball chase percentage, and at bats per base on ball. Further analyses comparing players in the top and bottom quintiles of vision screening scores showed significantly better plate discipline in metrics including at bats per base on ball and in-zone swing percentage, indicating that batters with better visual function are more likely to be successful when batting.
Most recently, among the baseball studies, Liu and colleagues16 evaluated the relationship between pre-season visual and oculomotor evaluations and pitch-by-pitch batting performance in the subsequent season from a sample of 71 professional baseball batters. In this study, eye tracking (RightEye, Bethesda, MD), visual-motor (Senaptec, Beaverton, OR), and optometric evaluations were compared through nested regression models with pitch-level data from Trackman 3D Doppler radar used to generate batting propensity scores including out-of-zone swing rate, in-zone swing rate, in-zone swing, and miss rate, as well as the batters' highest attained league levels during the season. Results indicated that visual evaluations relying on eye tracking (smooth pursuit accuracy and oculomotor processing speed) significantly predicted the highest attained league level and out-of-zone and in-zone swing propensity rates. These findings were taken as evidence that batters with superior visual and oculomotor abilities are more discerning at the plate.
Among the four identified studies that addressed sports other than baseball, the first report came in 1995 from Berg and Killian,41 who compared visual field size among a hybrid sample of 12 collegiate softball players and 12 collegiate nonathletes. In this study, visual field was measured through kinetic perimetry using a Topcon (Oakland, NJ) manual perimeter. Among the softball players, batting performance was assessed through both a noncompetition batting test done off of a pitching machine and batting averages from Division 1 competitive games (perimetry was assessed in the middle of the team's season). Results indicated that, although visual fields were significantly larger in the athletes than nonathletes, field size did not correlate with either competitive or noncompetitive batting averages.
To assess the role of eye movements in marksmanship, Causer and colleagues15 reported on a 2010 study measuring visual search behaviors and gun barrel kinematics in 24 elite and 24 subelite shooters. Point-of-gaze and gun barrel movements were recorded during skeet, trap, and double-trap events performed at the International Shooting Sport Federation shooting range during noncompetitive but Olympic-rules shooting events. The duration and onset of quiet eye, defined as the final fixation on a target before movement initiation,42 were calculated in relation to the scene camera on each shot and compared with gun motion profiles captured by two stationary external cameras. These were submitted to ANOVA to examine the effects of skill (elite/subelite) and shot outcome (hit/miss), with effect sizes calculated as partial η2. Results demonstrated that, in all shooting disciplines, elite shooters produced both earlier quiet eye onsets and longer durations than did subelite shooters. Moreover, in all disciplines, quiet eye duration was longer and onset earlier during successful, compared with unsuccessful, trials for all shooters, providing evidence that the stability of gaze before shot initiation is important for shooting success.
In 2014, Mangine and colleagues12 reported on results of a study testing the relationship between visual tracking speed and reaction times with game statistics in 12 professional basketball players. Visual tracking speed was obtained from 20 trials on the NeuroTracker (CogniSens Athletic, Inc., Montreal, Canada) multiple-object tracking test, whereas reaction times were measured with the Dynavision D2 light board (Dynavision International LLC, West Chester, OH). These assessments were compared with game statistics from the subsequent National Basketball Association season, including assists, turnovers, assist-to-turnover ratio, and steals. Results indicated positive relationships between visual tracking speeds and assists (r = 0.78, P = .003), steals (r = 0.77, P = .003), and assist-to-turnover ratios (r = 0.78, P = .003) but no relationship between reaction time and any of basketball performance measures.
Poltavski and Biberdorff18 reported on results from a 2015 study comparing assessments on the Nike Sensory Station with game statistics from 38 Division 1 collegiate male and female ice hockey players (19 each). In this study, the authors performed multiple regression analyses with backward elimination to isolate which, if any, of the nine assessment tasks on the Sensory Station predicted each of three different game statistics, aggregated over two successive seasons. For offensive players, these included the percent of goals scored and the average number of points per game. For all players, the average number of penalty minutes per game was also calculated. Results demonstrated that 69% of variance in the goals made could be predicted by better performance on four for the assessment tasks, whereas 33% of variance in game points and 24% of variance in the duration of penalty time could be accounted for by other tasks.
Finally, Vickers and colleagues43 reported on the role of quiet eye analysis in basketball three-point shooting in a 2019 noninterventional descriptive report. Twelve university or semiprofessional basketball players, both male and female, were enrolled. Performance of three-point shots was recorded in both defended and undefended conditions. Both the accuracy of the shot and different visual fixations of the shooter were recorded and compared. The authors found that shooting accuracy was enhanced when optimal visual fixation patterns (i.e., quiet eye) were observed. They also note that this was not affected by defenders' actions.
Collectively, these studies stemming from highly variable assessment approaches, athlete samples, and scientific methodologies provide mixed evidence supporting the role of different visual skills toward sporting performance.
Studies Testing the Effects of Vision Training with Statistical Production during Competition
A total of 16 publications that addressed vision training intervention and their effects on sports performance were identified. Of these, eight involved baseball, two reported on cricket, and one each reported on volleyball, hockey, soccer, badminton, golf, and racquet sports.
Seven articles dealt with vision training and baseball performance metrics, all of which reported some level of improvement with vision training.
In 2006, Fadde44 reported on the effects of training perceptual decision making on baseball performance. All position players from a National Collegiate Athletic Association baseball team were rank ordered by their coaches based on overall hitting ability, and then adjacently ranked players were assigned to either a treatment group that underwent 2 weeks of pitch occlusion training or a control group that performed typical baseball activities. Batting statistics for the next 18 games were compared using the Mann-Whitney U test. The author reports that significantly better batting averages were observed in the trained group, as compared with the control group, during the post-training games (P < .05), whereas slugging percentage and on-base percentage were numerically superior but not statistically significant.
A 2012 article by Clark and colleagues45 reported on the use of a vision training program with the University of Cincinnati baseball team. All players on the team underwent a diverse vision training program consisting of circuit training with both analog (e.g., Brock String, near-far charts, and saccades) and digital (e.g., drills with stroboscopic eyewear, Dynavision light board, and tachistoscope) training activities for 6 weeks before the season and during the baseball season. Overall team results, such as cumulative batting average and slugging percentage, are reported and compared with the rest of the conference during that same season. The authors report a significant increase in batting average and slugging percentage for the Cincinnati team relative to the rest of the conference (P = .02), with similar findings from analyses of conference and nonconference games.
In a 2014, Deveau and colleagues46 report findings on the effects of a perceptual training program on on-field batting performance and generalized visual abilities. Nineteen collegiate batters from the University of California, Riverside baseball team completed 30 sessions, whereas 18 pitchers served as a control group. In this training program, individuals practiced a digital, near-threshold target detection task in which they searched for patterns of differing spatial frequencies and orientations that increased in contrast but decreased in point values over time. The authors made direct comparisons between the batters and pitchers on transfer visual acuity tests, which demonstrated significant improvements for the trained batters over the untrained pitchers. To assess baseball performance, the authors compared on-field game statistics before and after training for 11 of the batters in relation to 78 batters from the Big West conference who were matched for age and position. Results demonstrate that the vision training group had significantly fewer strikeouts (P = .03) and more runs created than did the comparison group, with aggregated improvements equaling four to five projected wins.
Belling and Ward47 reported in 2015 on the effects of video-based training for pitch recognition and pitch location in nine National Collegiate Athletic Association Division 1 players who had on-field performance data. Training consisted of temporal occlusion training of pitch recognition and location on a 65-in touch screen display under an adaptive procedure with progressively earlier occlusion after successful responses. ANOVA, uncorrected for multiple comparisons, and calculation of Hedges g effects sizes, without the benefit of a control comparison group, were performed. The authors reported significant improvements in the number of home runs (P = .008), runs scored (P = .04), and slugging percentage (P = .04) with moderate or large effect sizes. Analysis of walks, batting average, and on-base percentage were numerically larger for the intervention group, although these differences did not reach statistical significance.
In 2016, Fadde48 reported on the effects of pitch recognition training in a team of 18 collegiate baseball players. Analysis was performed on total team statistics rather than individual player performance, and comparison was made against batting performance of other teams in the conference, without the benefit of a matched control group. Eight different batting metrics were used in comparing the study population with the control teams. Using a Mann-Whitney U test that compared the likelihood of rank ordering from the intervention group being higher than the control group, the authors reported a significant gain in walk-to-strikeout ratio between years for the treatment group over the control (d = 0.953, P = .02), although none of the other statistics reached significance and the tests were not adjusted for multiple comparisons.
In 2017, Gray49 reported on the use of a virtual environment to assess potential improvements in batting performance. In this relatively large interventional study, 80 high school batters were assigned to 4 different groups that trained with adaptive hitting training in the virtual environment, extra sessions of batting practice in the virtual environment, extra sessions of real batting practice, or a control condition involving no additional training to the players' regular practice. Eight different virtual and real batting metrics, including in-zone swing percentage and chase rate, were analyzed with Bonferroni correction performed for multiple comparisons. Results showed that trained players in the adaptive virtual environment group demonstrated significantly greater improvement from pre-training to post-training compared with the other control groups on most of the on-field measures using independent-samples t tests at the Bonferroni-corrected critical P value of .006. These players also had superior batting statistics in league play and reached higher levels of competition after training (χ2 test of proportions, χ2 = 7.9, P = .05).
Most recently, in a 2020 report, Liu and colleagues50 reported on the results of a dynamic visual skills training program versus placebo training in 24 Division I collegiate baseball batters. Unlike all other studies summarized thus far, this study was pre-registered to report all the methods and a priori hypotheses before the study (https://doi.org/10.17605/OSF.IO/JXH8U) and included randomization into either an active dynamic vision training group or a matched placebo training control group. Three different types of metrics were used in comparing the treatment and control groups: visuomotor evaluations, instrumented batting practice measures, and National Collegiate Athletic Association game statistics. The authors found no difference between the groups in the visuomotor evaluations or in the game statistics. Significant differences were, however, observed for launch angle (P = .002) and hit distance (P < .001) in the instrumented batting practice tasks that each constituted medium effect sizes according to Cohen's standards (d′ = 0.74 and 0.70, respectively). As such, this pre-registered study provides evidence that vision training improved batting practice performance on metrics showing longer hit distances and higher flight angles, but further studies with larger samples are needed to corroborate and extend these findings.
In 2008, Balasaheb and colleagues51 reported on the effects of visual skills training on cricket batting. Vision training lasted 6 weeks and occurred 3 days each week. Thirty subjects were divided into three groups: a treatment group that received visual training, a placebo group given a reading task and a video task not believed to improve vision, and a control group that performed their typical cricket training with no added visual tasks. The authors found that all three groups showed statistically significant improvement in batting performance, but only the treatment group demonstrated improved visual skills after training in all skills, whereas the placebo and control groups showed improvement in some, but not all, visual abilities. As such, these findings do not indicate selective improvements in cricket performance because of vision training over the control interventions.
Hopwood and colleagues52 published results in a 2011 study with 12 senior international cricket players in which they compared pre-versus-post visual training abilities on a video-based decision test and on-field performance metrics. Training lasted 6 weeks, and seven players performed an on-field training program in addition to perceptual training, whereas five players underwent only on-field training. The authors report that the trained group demonstrated a statistically significant improvement in video-based decision making when compared with their pre-test ability and compared with the control group. The decision accuracy of the control group showed a large decrease over the course of the study, despite an expectation for this to remain relatively stable. In addition, although the mean fielding success score showed a statistically significant improvement in the trained group, the mean movement initiation time did not show any difference.
In 2001, Abernethy and Wood22 reported on the effects of a vision training program on racquet sports. Subjects were divided into four groups with 10 subjects in each group and underwent either an analog oculomotor training program based on drills originally described by Revien and Gabor,53 a video-package-based “eyerobics” training program, placebo training, or no training. Training lasted for 4 weeks, and pre-performance and post-performance measures were recorded, including tennis forehand drive accuracy. The authors found no evidence of either improved vision or sports-specific performance in any of the groups.
In 2011, Vine and colleagues54 described results from quiet eye training in golf putting with treatment and control groups, each consisting of 11 elite golfers. The treatment group received golf putting quiet eye training, whereas both groups received video-based gaze feedback. Putting was scored both before the study and after training. The authors found no difference between the groups at the outset but reported a significant difference after training, with the intervention group requiring 1.9 fewer putts per round under game conditions than the control group.
Ryu and colleagues55 published a 2018 report in which 36 novice badminton players were divided into 3 groups that received different perceptual training during a 3-day period. In addition to the control group that was provided normal video of badminton shots, the training groups viewed either low or high spatial frequency–filtered videos of badminton shots. In all cases, videos were occluded at the moment of contact between the racquet and the shuttle, and subjects were asked to identify the direction of the shuttle movement. The authors noted that there was a significant difference in accuracy between the control group and that of the low spatial frequency–trained group (P = .005). This was theorized to result from longer gaze at the point of racquet-shuttle contact in the low-spatial-frequency group.
In 2016, Romeas and colleagues56 studied the effects of multiple-object tracking training on soccer performance. They describe the use of a 3D training program on university-level athletes, with nine participants in the active multiple-object tracking training group, seven in the active control group who watched 3D soccer videos without multiple object tracking, and seven subjects in a passive, no intervention, control group. Results indicated improved decision-making ability for passing in the actively trained group compared with the combined control groups. There was no difference in dribbling or shooting decision-making accuracy between the groups.
In 2013, Mitroff and colleagues57 reported on the results of a pilot study conducted with 11 National Hockey League players, performed during pre-season training camp, as players were attempting to qualify for a professional team. Here, six players underwent a 14-day training program in which they did on-ice hockey drills while wearing Nike Vapor Strobe eyewear (Nike Inc., Beaverton, OR), whereas a control group of five players did comparable drills with normal vision. Before and after training, players were scored on puck placement drills receiving higher points for better accuracy when shooting at distant targets. Using analysis of covariance that controlled for pre-training performance, the authors found a significant gain for the intervention group over the controls, which amounted to an 18% improvement for strobe training, but no change for the control group.
Jenerou et al.58 studied the effect of a vision training program on the performance of a collegiate level (Division I) ice hockey team. In their 2015 study, they enrolled 22 male players who underwent a 6-week, generic and nonpersonalized binocular, accommodative, and dynamic visual skills training program. Although it was not possible to determine which training tasks had no effect, some effect, or a significant effect on performance, the authors note that, subjectively, players reported a positive impact on their game performance. In addition, there was a statistically significant improvement in goals (t = −3.778, P < .003), shots on goal (t = −3.262, P < .01), and shooting percentage (t = −2.598, P < .03) when compared individually from pre-training to post-training.
Finally, Formenti and colleagues59 reported in 2019 on the effects of perceptual vision training in volleyball. This study enrolled 51 female volleyball players who were randomly assigned to either a 6-week context sport-specific training group that performed repetitive volleyball drills without analytic vision exercises, a vision training group that performed analytic vision exercises during non-volleyball training, or a vision training sport-specific group that performed analytic vision exercises during volleyball training activities. Each player underwent pre-training and post-training assessment of both volleyball-specific abilities (accuracy of passing, setting, and serving) and more general cognitive abilities (reaction time, executive control, perceptual speed). Results indicated that, although perceptual vision training improved the cognitive abilities of the perceptually trained players as compared with the conventionally trained players, the conventionally trained players demonstrated superior volleyball-specific abilities as compared with the perpetually trained players.
CRITICAL REVIEW AND FUTURE DIRECTIONS
The representative collection of sports vision studies summarized previously includes a diverse range of research questions, experimental approaches, and findings. They offer a heterogeneous set of conclusions: many support the hypothesis that assessments correlate with game performance, and others do not; some support the hypothesis that vision training improves game performance, and others do not. Reviewing these publications reveals several commonalities that can be used to gain a greater appreciation of the role of vision and visual training in sports performance, as well as the strengths and limitations of research in this field. In the following sections, we review issues related to the sample sizes, assessments and interventions, outcome measures, statistical approaches, and conclusions. We end by summarizing improvements that may move the field forward.
Review of Studies Addressing Visual Assessment and Sports Performance
As noted previously, the power of any statistical inference depends on the effect size and the number of observations. In the case of the 13 articles addressing visual assessments, 4 studies14,17,19,35 included large samples with between 252 and 475 professional baseball players, whereas the others had 71 or fewer individuals. Several, but not all, studies included players with only a large number of plate appearances, whereas others included low numbers or did not report this information. Among the non-baseball studies, the average sample size was 19.6 athletes.
Access to research with large athlete cohorts is not common, and therefore, it is important not only to take advantage of opportunities when they are presented but also to scale research questions to fit the sample that is available. For instance, as reported by Reichow and colleagues,13 the research sample consisted of only 20 collegiate batters, but the study addressed a focused question and single statistical test that did not require multiple comparison correction. In contrast, Laby and colleagues19 included a sample of 450 professional batters while testing a range of relationships between hand-eye coordination and batting performance while correcting for multiple comparisons. These two studies presented appropriately scaled research questions, but this was not always the case across the whole literature.
Although visual perception is unitary, the visual system is made up of many different, interrelated, fundamental processes, including static and dynamic acuity, binocular vision, ocular-motor function, attention, visual working memory, and higher-level cognition. The range of assessments used in the reviewed studies reflects this heterogeneity. These include optometric measures of ocular function, such as ocular dominance, perimetry, eye movements, and measures of refractive error, using both sport-specific stimuli and non–sports-specific stimuli. Although these assessments and stimuli are generally well motivated in the context of each study, and many individual studies show positive relationships between visual assessments and game performance, collectively, they use widely different assessments, making it difficult to build consensus or demonstrate reproducibility around certain relationships.
Another challenge in this literature stems from the variable and nonspecific game performance measures that are available for athletes who play different sports or positions. Among the 13 assessment studies reviewed, 8 tested baseball and 1 tested softball. These sports are generally favorable for research because common metrics of batting and/or pitching are used that broadly capture production among batters or among pitchers. Nonetheless, certain statistics are more suitable for isolating production of individual athletes and even decision making during each at bat occurrence. With the advent of radar-based tracking systems, these “advanced analytics” have grown in popularity, and this is reflected in the reviewed studies. The five articles published before 2018 all used counting statistics such as batting averages, home runs, and earned run averages, which depend on the actions of fielders and base runners. Three of the four studies published since14,16,19 used plate discipline and swing propensity measures that isolate the actions of the batter, without regard to other actions on the field. Continued use of such plate discipline metrics should be encouraged, as they are most able to directly link vision and performance in an individual athlete.
Across the studies addressing links between assessment and game production, several different statistical approaches were used. The most common was application of Pearson (linear) or Spearman (rank order) correlations, which attempt to measure the dependent relationships between assessment scores and game performance scores, across individuals. In most cases, multiple outcome measures were tested; however, multiple comparison corrections were not implemented to account for these tests. Some studies, however, did apply corrections.14,19 At least four studies16–18,43 implemented regression analyses, which allow for estimation of multiple predictor variables, including one that performed a Bayesian regression approach.17 As such, these studies offered a deeper accounting of multiple interrelated factors that may contribute to game statistical performance. Two studies15,43 also provided information about effect sizes, which go beyond statistical inference to provide information about the strength of relationships between variables.
In conclusion, these studies offer promising but incomplete evidence that performance on visual assessments may correlate with game performance in competitive situations. Despite such promise, strong conclusions cannot be drawn across this literature because of the considerable heterogeneity in assessment and outcome measures, sample sizes, and statistical approaches.
Review of Studies Addressing Visual Training and Sports Performance
In addition to the attributes noted previously, sufficiently powered subject sample sizes, randomization, and the use of matched-control groups are additional important components of a well-designed interventional study. Although it is often difficult to recruit athletes for such studies, especially at the elite and professional levels, the quality and validity of study findings are critically dependent on optimal pre-study planning and study execution. For training studies, authors should consider the anticipated effect of their intervention on the sports performance metric they wish to study and calculate sample sizes required to have a sufficiently powered study that delivers meaningful results. Software packages, such as G*Power (University of Dusseldorf, Düsseldorf, Germany) are available to help with this estimation, whereas statistician colleagues can help guide optimal study design.
In the intervention studies summarized previously, there was an average of 30 subjects, split approximately evenly between treatment and control groups. Considering that the influence of various visual functions on sports performance is likely relatively small, when compared with other factors, it is somewhat surprising that, despite these small sample sizes, most reports showed significant gains due to the training intervention. Some studies49,51,59 were slightly larger in scope, offering a better chance to demonstrate the effects of the intervention, whereas the remaining studies generally had a low number of subjects, leading to questions about the validity and replicability of these results. Despite challenges in obtaining access to larger sample sizes, especially among professional and collegiate athletes, future studies should attempt to achieve large samples, possibly by performing cooperative multisite studies.
In designing interventional trials, the correspondence between treatments and outcome assessments plays a central role in the study's success. The reviewed studies explored visual training interventions including stroboscopic training, ocular-motor training, eye-hand coordination training, anticipatory timing training, and adaptive perceptual learning. Nine studies implemented training in naturalistic interventions, whereas seven used general interventions that seek to transfer foundational skills and trained out of context. Among the reviewed studies, seven contained on-field performance metrics, and nine contained competitive game metrics. Although several studies made straightforward links between treatments and outcomes, others failed to demonstrate good matches, for example, using whole team performance statistics to detect individually directed training interventions.48 Most of the reviewed studies used control groups, although in some cases these may not have been matched and/or randomized optimally. Lastly, studies did not always apply a personalized sports-specific approach to the interventions used, instead opted for a generalized wide-ranging set of interventions performed equally by all participants.58
Within this literature, careful consideration needs to be paid to the statistical analysis plan, particularly in situations where multiple analyses can be performed. Many of the interventional studies summarized previously may have failed to correct for multiple comparisons, inflating the possibility of type I (false-positive) error and misinterpreting a finding as being hypothesis driven. In several cases, findings were described as exploratory, inviting follow-up research, and a single study was labeled as a pilot study.57 Some studies, including that by Gray,49 appropriately corrected for the multiple analyses, whereas only Liu and colleagues50 pre-registered hypotheses and analyses, further strengthening the evidence. Encouragingly, these studies both showed positive effects, using specific outcome measures and pointing to the promise of virtual reality and digital training augmentations of vision training. To move the field forward, future studies should adopt these best practices.
Suggestions to Improve the Field Going Forward
To understand the role of vision in sports, develop and test interventions to improve performance, and provide honest evaluation, the field of sports vision needs to continue to evolve. The combination of vision assessments, interventions, and game performance data makes this a multidisciplinary endeavor that spans the traditional bounds of science. Increased demonstrations from pre-registered, appropriately powered, and controlled studies will continue to bolster the value of these techniques, whereas reporting of negative findings will aid in determining reproducibility. Ultimately, adopting these approaches will better demonstrate the value of these services to teams while growing the specialty and attracting more high-quality practitioners and researchers.
1. Le Runigo C, Benguigui N, Bardy BG. Perception-action Coupling and Expertise in Interceptive Actions. Hum Mov Sci 2005;24:429–45.
2. Laby DM, Kirschen DG, Pantall P. The Visual Function of Olympic-level Athletes—An Initial Report. Eye Contact Lens 2011;37:116–22.
3. Burris K, Liu S, Appelbaum L. Visual-motor Expertise in Athletes: Insights from Semiparametric Modelling of 2317 Athletes Tested on the Nike Sparq Sensory Station. J Sports Sci 2020;38:320–9.
4. Williams AM, Ford PR, Eccles DW, et al. Perceptual-cognitive Expertise in Sport and Its Acquisition: Implications for Applied Cognitive Psychology. Appl Cogn Psychol 2011;25:432–42.
5. Starkes JL, Ericsson KA. Expert Performance in Sports: Advances in Research on Sport Expertise. Champaign, IL: Human Kinetics; 2003.
6. Uchida Y, Kudoh D, Higuchi T, et al. Dynamic Visual Acuity in Baseball Players Is Due to Superior Tracking Abilities. Med Sci Sports Exerc 2013;45:319–25.
7. Laby DM, Rosenbaum AL, Kirschen DG, et al. The Visual Function of Professional Baseball Players. Am J Ophthalmol 1996;122:476–85.
8. Hoffman LG, Polan G, Powell J. The Relationship of Contrast Sensitivity Functions to Sports Vision. J Am Optom Assoc 1984;55:747–52.
9. Vickers JN, Adolphe R. Gaze Behavior during a Ball Tracking and Aiming Skill. Int J Sports Vis 1997;4:18–27.
10. Mann DT, Williams AM, Ward P, et al. Perceptual-cognitive Expertise in Sport: A Meta-analysis. J Sport Exerc Psychol 2007;29:457–78.
11. Voss MW, Kramer AF, Basak C, et al. Are Expert Athletes Expert in the Cognitive Laboratory? A Meta-analytic Review of Cognition and Sport Expertise. Appl Cogn Psychol 2010;24:812–26.
12. Mangine GT, Hoffman JR, Wells AJ, et al. Visual Tracking Speed Is Related to Basketball-specific Measures of Performance in NBA Players. J Strength Cond Res 2014;28:2406–14.
13. Reichow AW, Garchow KE, Baird RY. Do Scores on a Tachistoscope Test Correlate with Baseball Batting Averages? Eye Contact Lens 2011;37:123–6.
14. Laby DM, Kirschen DG, Govindarajulu U, et al. The Effect of Visual Function on the Batting Performance of Professional Baseball Players. Sci Rep 2019;9:16847.
15. Causer J, Bennett SJ, Holmes PS, et al. Quiet Eye Duration and Gun Motion in Elite Shotgun Shooting. Med Sci Sports Exerc 2010;42:1599–608.
16. Liu SC, Edmunds FR, Burris K, et al. Visual and Oculomotor Abilities Predict Professional Baseball Batting Performance. Int J Perf Anal Sport 2020;20:683–700.
17. Burris K, Vittetoe K, Ramger B, et al. Sensorimotor Abilities Predict On-field Performance in Professional Baseball. Sci Rep 2018;8:116.
18. Poltavski D, Biberdorff D. The Role of Visual Perception Measures Used in Sports Vision Programmes in Predicting Actual Game Performance in Division I Collegiate Hockey Players. J Sports Sci 2015;33:597–608.
19. Laby DM, Kirschen DG, Govindarajulu U, et al. The Hand-eye Coordination of Professional Baseball Players: The Relationship to Batting. Optom Vis Sci 2018;95:557–67.
20. Harpham JA, Mihalik JP, Littleton AC, et al. The Effect of Visual and Sensory Performance on Head Impact Biomechanics in College Football Players. Ann Biomed Eng 2014;42:1–10.
21. Erickson GB. Sports Vision: Vision Care for the Enhancement of Sports Performance. Philadelphia, PA: Butterworth-Heinemann; 2007.
22. Abernethy B, Wood JM. Do Generalized Visual Training Programmes for Sport Really Work? An Experimental Investigation. J Sports Sci 2001;19:203–22.
23. Page J, Causer J, Wilson M, et al. The BASES Expert Statement on the Effectiveness of Vision Training Programmes. Sports Exerc Sci 2013;12–3.
24. Appelbaum LG, Erickson G. Sports Vision Training: A Review of the State-of-the-art in Digital Training Techniques. Int Rev Sport Exerc Psychol 2018;11:160–89.
25. Appelbaum LG, Cain MS, Schroeder JE, et al. Stroboscopic Visual Training Improves Information Encoding in Short-term Memory. Atten Percept Psychophys 2012;74:1681–91.
26. Appelbaum LG, Lu Y, Khanna R, et al. The Effects of Sports Vision Training on Sensorimotor Abilities in Collegiate Softball Athletes. Athl Train Sports Health Care 2016;8:154–63.
27. Appelbaum LG, Schroeder JE, Cain MS, et al. Improved Visual Cognition through Stroboscopic Training. Front Psychol 2011;2:276.
28. Akbas A, Marszalek W, Kamieniarz A, et al. Application of Virtual Reality in Competitive Athletes—A Review. J Hum Kinet 2019;69:5–16.
29. Michalski SC, Szpak A, Loetscher T. Using Virtual Environments to Improve Real-world Motor Skills in Sports: A Systematic Review. Front Psychol 2019;10:2159.
30. Hadlow SM, Panchuk D, Mann DL, et al. Modified Perceptual Training in Sport: A New Classification Framework. J Sci Med Sport 2018;21:950–8.
31. Patten ML, Newhart M. Understanding Research Methods: An Overview of the Essentials. New York, NY: Routledge; 2017.
32. Maxwell SE, Lau MY, Howard GS. Is Psychology Suffering from a Replication Crisis? What Does “Failure to Replicate” Really Mean? Am Psychol 2015;70:487–98.
33. American Optometric Association (AOA). Sports Vision Section Bibliographic Resource. St. Louis, MO: American Optometric Association; 2011.
34. Williams AM, Ford PR. Expertise and Expert Performance in Sport. Int Rev Sport Exerc Psychol 2008;1:4–18.
35. Laby DM, Kirschen DG, Rosenbaum AL, et al. The Effect of Ocular Dominance on the Performance of Professional Baseball Players. Ophthalmology 1998;105:864–6.
36. Molia LM, Rubin SE, Kohn N. Assessment of Stereopsis in College Baseball Pitchers and Batters. J AAPOS 1998;2:86–90.
37. Müller S, Fadde PJ. The Relationship between Visual Anticipation and Baseball Batting Game Statistics. J Appl Sport Psychol 2016;28:49–61.
38. Erickson GB, Citek K, Cove M, et al. Reliability of a Computer-based System for Measuring Visual Performance Skills. Optometry 2011;82:528–42.
39. Wang L, Krasich K, Bel-Bahar T, et al. Mapping the Structure of Perceptual and Visual-motor Abilities in Healthy Young Adults. Acta Psychol (Amst) 2015;157:74–84.
40. Krasich K, Ramger B, Holton L, et al. Sensorimotor Learning in a Computerized Athletic Training Battery. J Mot Behav 2016;48:401–12.
41. Berg WP, Killian SM. Size of the Visual Field in Collegiate Fast-pitch Softball Players and Nonathletes. Percept Motor Skills 1995;81:1307–12.
42. Lebeau JC, Liu S, Sáenz-Moncaleano C, et al. Quiet Eye and Performance in Sport: A Meta-analysis. J Sport Exerc Psychol 2016;38:441–57.
43. Vickers JN, Causer J, Vanhooren D. The Role of Quiet Eye Timing and Location in the Basketball Three-point Shot: A New Research Paradigm. Front Psychol 2019;10:2424.
44. Fadde PJ. Interactive Video Training of Perceptual Decision-making in the Sport of Baseball. Tech Inst Cogn Learn 2006;4:237–56.
45. Clark JF, Ellis JK, Bench J, et al. High-performance Vision Training Improves Batting Statistics for University of Cincinnati Baseball Players. PLoS One 2012;7:e29109.
46. Deveau J, Ozer DJ, Seitz AR. Improved Vision and On-field Performance in Baseball through Perceptual Learning. Curr Biol 2014;24:R146–7.
47. Belling P, Ward P. Time to Start Training: A Review of Cognitive Research in Sport and Bridging the Gap from Academia to the Field. Procedia Manufac 2015;1219–24.
48. Fadde PJ. Instructional Design for Accelerated Macrocognitive Expertise in the Baseball Workplace. Front Psychol 2016;7:292.
49. Gray R. Transfer of Training from Virtual to Real Baseball Batting. Front Psychol 2017;8:2183.
50. Liu S, Ferris L, Hilbig S, et al. Dynamic Vision Training Transfers Positively to Batting Performance among Collegiate Baseball Batters. Psychol Sports Exerc 2020;51.
51. Balasaheb T, Maman P, JS S. The Impact of Visual Skills Training on Batting Performance in Cricketers. Serb J Sports Sci 2008;2:17–23.
52. Hopwood M, Mann D, Farrow D, et al. Does Visual-perceptual Training Augment the Fielding Performance of Skilled Cricketers? Int J Sports Sci 2011;6:523–35.
53. Revien L, Gabour M. Sports Vision: Dr Revien's Vision Training Exercises for Athletes. New York, NY: New York Publishing; 1981.
54. Vine SJ, Moore LJ, Wilson MR. Quiet Eye Training Facilitates Competitive Putting Performance in Elite Golfers. Front Psychol 2011;2:8.
55. Ryu D, Abernethy B, Park SH, et al. The Perception of Deceptive Information Can Be Enhanced by Training That Removes Superficial Visual Information. Front Psychol 2018;9:1132.
56. Romeas T, Guldner A, Faubert J. 3DMultiple Object Tracking Task Performance Improves Passing DecisionMaking Accuracy in Soccer Players. Psychol Sport Exerc 2016;1–37.
57. Mitroff SR, Friesen P, Bennett D, et al. Enhancing Ice Hockey Skills through Stroboscopic Visual Training. A Pilot Study. Athl Train Sports Health Care 2013;5:261–4.
58. Jenerou A, Morgan B, Buckingham RS. A Vision Training Program's Impact on Ice Hockey Performance. Optom Vis Perform 2015;3:139–49.
59. Formenti D, Duca M, Trecroci A, et al. Perceptual Vision Training in Non–sport-specific Context: Effect on Performance Skills and Cognition in Young Females. Sci Rep 2019;9:18671.