In daily life, the environment presents extensive visual information, with the most relevant visual information being selected and obtained. Such information is extremely important when we perform various tasks or actions, especially in sports. Dynamic visual acuity (DVA) is defined as the ability to discriminate the fine parts of a moving object and involves accurately determining the relative motion between an object and the observer (8,22). DVA markedly affects performance in sports such as volleyball, basketball (24), baseball (26), softball (23), and motor sports (27). Baseball, tennis, and badminton players generally display superior DVA as compared with nonplayers in that they are able to recognize the gap in moving Landolt C rings, which is a standard symbol used for testing vision, at significantly higher velocities than nonplayers (13). Such excellent DVA in sport players has been suggested to be derived from appropriate saccadic eye movement in tracking moving objects (15,18).
In baseball, professional pitchers can throw a fastball at velocities exceeding 150 km·h−1. The angular velocity of eye movements required to track such fastballs surpasses the physiological limits of angular velocity in human eye movements. Therefore, even the eyes of the most proficient batters cannot track the ball (5,6). Yet professional batters do manage to hit the ball with surprising consistency. Bahill and Laritz (5) suggested that major league baseball players make anticipatory saccades when dealing with balls traveling faster than the upper limit of eye movement as the ball approaches the batter’s box.
Normal human saccadic eye movements, as observed in daily life, are seldom larger than 15° in amplitude (3). In experimental conditions, however, the velocity of saccadic eye movement increases with the amplitude of saccade and may even exceed 500°·s−1 at an amplitude of approximately 60° (4,9,10). In one study using Landolt C rings moving at a speed of 300°·s−1, eye movement velocity reached more than 600°·s−1 (17). However, most laboratory studies relating DVA to eye movement have used visual stimuli moving at a speed of 400°·s−1 or less (7,8,13), and there have been no systematic studies in which measurement of eye movement was performed with visual stimuli moving at a speed close to those experienced by baseball players. The first aim of the present study was to measure the peak velocity and latency of eye movement in response to visual stimuli moving at speeds up to 900°·s−1 in both baseball players and nonplayers.
We are able to recognize moving objects accurately when the error between “eye position” and “the position of the moving stimuli” is small, that is, when the distance between the image projected onto the retina and the central fovea (retinal error) is small (30). Visual perception decreases by an amount as great as 70% when the retinal error is 5° (1,21). However, there is a possibility that the excellent DVA exhibited by baseball players is derived from the ability to correctly perceive objects even in the presence of a large retinal error. The ability to perceive visual stimuli on the retina has not been compared between baseball players and nonplayers. Thus, the second aim of the present study is to determine the relationship between the ability to perceive moving visual stimuli and the retinal error in baseball players and nonplayers.
Sixteen men participated in this study. Eight belonged to a collegiate baseball team (baseball players, mean ± SD age = 20.5 ± 1.6 yr), and eight had no history of sporting activity (nonplayers, mean ± SD age = 20.8 ± 1.2 yr). All baseball players were fielders and were right-handed hitters. Their baseball experience was 12.4 ± 3.7 yr. All participants had good static visual acuity, equal to or better than 20/20. This study was approved by the ethical committee of the Faculty of Sports Sciences, Waseda University. The experiments were conducted following the Declaration of Helsinki. Written informed consent was obtained from the individual participants after a detailed explanation of the content of the experiment and the object of the study.
Participants were seated in front of a screen on which a moving target was projected. The semicircular screen extended 90° before the eyes. The distance from the participant to the screen was 90 cm (Fig. 1A). The size of the target had a visual angle of 4° that was determined based on the size of a baseball observed by a hitter. The size of a baseball is approximately 7 cm. When the ball reaches the home plate, the distance between the ball and the hitter is approximately 1 m, and the visual angle of the ball observed by the hitter is approximately 4°.
We used moving Landolt C rings of the appropriate size and speed to simulate a baseball approaching a batter. There were four orientations of the targets (direction of the gap of the Landolt ring: top, bottom, right, and left), eight target speeds (200°·s−1, 300°·s−1, 400°·s−1, 500°·s−1, 600°·s−1, 700°·s−1, 800°·s−1, and 900°·s−1), and two moving directions (right and left). Among the total of 64 settings, the order was randomly determined. According to the selected setting, the target presented on the screen moved horizontally at a constant angle rate. The target appeared on the left or right side of the screen. The participant was informed before each trial which side of the screen the target would appear on and was asked to pursue the target immediately (Fig. 1B). The participants were required to judge the direction of the Landolt gap (up, down, right, or left) in a forced choice manner by pushing one of four buttons corresponding to each direction. The answer was directly transferred to a PC and recorded.
Movements of the dominant eye during target movement were recorded using an eye movement recording system (Eyelink II; SR Research Ltd., Mississauga, Canada; Fig. 1C) at a sampling rate of 500 Hz. Before the trial, the participant gazed at seven calibration points, during which time eye position was determined (Fig. 1A). The obtained data concerning horizontal eye position during target presentation on the screen were A/D converted and transferred to the PC. When the latency of eye movement exceeded the time of target movement on the screen, that is, when the target disappeared from the screen before the initiation of eye movement, the participants usually could not pursue the target. In such cases, the trial was excluded from analysis.
Data analysis and statistics
For an objective definition of DVA, we introduced the concept of “correct response rate (P),” which was obtained from the forced choice data. The correct response rate (P) is approximated by the reaction curve (psychophysical curve) defined with the following equation (31):
where v denotes the target speed, d is the horizontal shift of the curve, and s is the width of the slope. MATLAB (optimization toolbox) was used for the fitting to minimize Pearson’s chi-square statistic (19).
Because there were four gap directions of the target, the probability of correct answers by chance was one of four in an approximation using the psychological curve. Therefore, a correct response rate of 25% was defined as the chance level, and an appropriate psychophysical curve was determined for those data in a correct response rate range from 25% to 100%. As in a previous study (29), a correct response rate of 75% was used as the cutoff between “can visually recognize” and “cannot visually recognize.” The target speed (°·s−1) at the crossing of the psychophysical curve and the correct response rate of 75% was defined as the objective DVA value for each participant. A two-way MANOVA (target speed × group) was used for evaluating changes in the correct response rate with Pillai’s trace statistic that does not assume sphericity (16,25). When the interaction was significant, the simple main effects were examined to see the target speed at which the effect of group became significant.
The analysis items of eye movement were as follows: 1) the latency until the start of eye movement (Fig. 2A); 2) the peak eye movement velocity obtained as the time differential of the eye position (Fig. 2B); 3) the retinal error, which was defined as the positional error between the target and the eye position (Fig. 2C); 4) the retinal slip, which indicated the difference in relative velocity between target and eye movements (Fig. 2D); and 5) the minimum retinal error (MRE), which was defined as the retinal error when the retinal slip equaled zero (Fig. 2E). A two-way MANOVA (target speed × group) was used for evaluating changes in latency, velocity, and MRE. When the interaction was significant, the simple main effects were examined to see those when the effect of group became significant. All analyzes were performed using MATLAB (The MathWorks, Inc., Natick, MA) on the PC.
The mean correct response rate at each target speed in each subject group and its approximate curve are shown in Figure 3. There was a significant difference in interaction (target speed × group) (F7,98 = 2.722, P = 0.0127). In addition, significant differences were observed among the target speeds (F7,98 = 113, P < 0.0001) and between the subject groups (F1,14 = 9.176, P = 0.0090). Analysis of the simple main effects in the interaction showed a significantly higher correct response rate at 400°·s−1, 500°·s−1, and 600°·s−1 in the players group than that in the nonplayers group (400°·s−1, F1,112 = 4.19, P = 0.0430; 500°·s−1, F1,112 = 7.94, P = 0.0057; 600°·s−1, F1,112 = 0.14, P = 0.016). In addition, the value at the crossing of the approximate curve and the correct response rate of 75% was 486°·s−1 ± 72°·s−1 for the players group and 393°·s−1 ± 64°·s−1 for the nonplayers group. There was a significant difference between the two groups, confirming a higher DVA for the baseball players.
The latency at each target speed in each subject group is shown in Figure 4A. Statistical analysis revealed that interaction (target speed × group) did not significantly differ between the two groups (F4,56 = 0.36, P = 0.8339). There was a significantly shorter latency in the players group than that in the nonplayers group for all target speeds from 200°·s−1 to 600°·s−1 (F1,14 = 27.6, P = 0.001). At a target speed of 700°·s−1, the mean latency and standard error in the players group was 105 ± 5.14 ms, still shorter than the target movement time on the screen (129 ms). However, the mean latency in the nonplayers group was longer than the target movement time. That is, when a participant from this group tried to track the target by eye movement, the target had already disappeared from the screen. Therefore, those eye movement data during target presentation that could not have been acquired were excluded from the analysis. At a target speed of 800°·s−1 and 900°·s−1, the mean latencies were longer than the target movement time on the screen (113 ms) in both groups and were similarly excluded from the analysis. The peak velocity of eye movement at each target speed in each group is shown in Fig. 4B. This interaction (target speed × group) showed a significant difference (F4,56 = 14.5, P < 0.0001). A significant difference was observed in the target speed (F4,56 = 7.14, P < 0.0001) between the two groups (F1,14 = 1.13, P = 0.0305). At target speeds of 500°·s−1 and 600°·s−1, the simple main effects significantly differed between the two groups (500°·s−1, F1,70 = 4.48, P = 0.0379; 600°·s−1, F1,70 = 7.94, P = 0.0063). On the other hand, at target speeds of 200°·s−1, 300°·s−1, and 400°·s−1, no significant difference was observed. For a rapidly moving target, the eye movement velocity in the nonplayers group peaked at approximately 600°·s−1, whereas in the players group, the peak was not reached until approximately 700°·s−1.
Statistical analysis of the MRE indicated significant differences in interaction (target speed × group) (F4,56 = 8.06, P < 0.0001), target speed (F4,56 = 239, P < 0.0001), and group (F1,14 = 16.4, P = 0.0012) (Fig. 4C). Analysis of the simple main effects in the interaction between the target speed and the group showed significant differences in the retinal error between the two groups at 400°·s−1, 500°·s−1, and 600°·s−1 (400°·s−1, F1,70 = 5.56, P = 0.0212; 400°·s−1, F1,70 = 11.4, P = 0.0012; 600°·s−1, F1,70 = 39.1, P < 0.0001). These results showed a smaller retinal error for rapidly moving targets in the players group as compared with the nonplayers group.
The relationship between the retinal error and the correct response rate is shown in Figure 5. The values in each group were approximated using the least squares method (players group: r = 0.85, P = 0.0021; nonplayers group: r = 0.89, P = 0.0010). The two approximate curves in the correct response rate range greater than 60% are almost identical. This indicates that there was no significantly difference in the ability to perceive a retinal image between the two groups.
In the present study, eye movement and visual perception were measured at target speeds in the range of 200°·s−1 to 900°·s−1. These measures were compared between baseball players, who are considered to have excellent DVA, and nonplayers. The results clearly showed that baseball players could recognize moving Landolt C rings better than nonplayers, especially at target speeds higher than 400°·s−1 (Fig. 3). We had initially hypothesized that the superior DVA of baseball players would partly be derived from a better perception of moving images on the retina, even in the presence of a large retinal error. However, there was no difference in the ability to perceive moving object’s images projected onto the retina between the two groups (Fig. 5). In fact, the mean retinal error was smaller in the players than that in the nonplayers, especially at high target speeds (Fig. 4C). Therefore, the baseball player’s high DVA was solely due to skill in tracking moving objects with their eyes rather than with an enhanced ability to perceive moving objects on the retina.
Baseball players were able to pursue objects with faster eye movement and at a shorter latency than nonplayers. In baseball, the speed of balls thrown by top level pitchers can exceed 150 km·h−1, in which case the angular velocity seen by a batter approximately 1 m from home plate is more than 1000°·s−1, which exceeds the limit of the eye movement velocity obtained in the present study (Fig. 6A and B). Therefore, even the eyes of the most proficient batters will not be able to pursue the ball strictly by following it with the eyes.
A previous study indicated that a significant correlation exists between DVA and eye movement latency in athletes (17). This was also confirmed in the present study in that the players had a shorter latency than nonplayers. In addition, the present study clarified for the first time that peak velocity is also higher in baseball players as compared with nonplayers. Therefore, the high DVA seen in baseball players is produced by both short latency and high velocity of eye movement, which accordingly makes retinal error smaller.
Is the high DVA of baseball players obtained by learning/training or is it due to some congenital superiority? Because a high DVA would provide for early success and encourage the pursuit of baseball, there is the possibility that those people with superior eye movement would be more likely to become baseball players. However, good eye movement is just one of many abilities required of successful baseball players. The subjects in the present study were playing at the university level, which is not easy to reach. It is improbable that all of them had a higher innate ability in eye movement than the nonplayers. Previous studies have shown that eye movements can be improved with training (20). Furthermore, plasticity of the neural circuitry associated with latency and velocity of eye movements has also been neurophysiologically indicated in lower animals as well as humans (12,14). The visual tracking of a ball actually requires baseball batters to move their eyes much faster than the physiological limit (Fig. 6). According to the training principle of specificity, the use of a target muscle at a fast speed is necessary to train the muscle to contract faster (2). The daily practice of baseball is an optimal situation for training the eye muscles (extraocular muscles). Therefore, the good eye movement latency and velocity observed in baseball players might not be partially or totally innate. Rather, it could be the result of learning through daily practice, which involved the visual tracking of a ball. It is hoped that the relative contributions of innate abilities and learned skills will be clarified in future longitudinal studies.
As shown in Figure 5, there was no difference between the player and the nonplayer groups in the relationship between retinal error and correct response rate. That is, the ability to perceive a moving object’s images projected onto the retina did not differ. Rather, the present study showed that baseball batters could keep the ball images close to the fovea (small retinal error) until the very end of a ball’s trajectory (Fig. 6). Thus, they might not have the chance of improving their perception of retinal images in their daily training sessions. Another possibility might be that the number of times for batters to see balls coming toward them in their baseball training was not sufficient enough to influence their visual perception, although it was enough to improve their tracking ability. Thus, the present results do not completely exclude the possibility that ability in visual perception in the presence of a large retinal error might be improved.
In the present study, we arbitrarily defined the criterion of correct response rate at 75% in the determination of DVA. As can be seen in Figure 3, a plot of the approximate curves of correct response rate against target speed for players and nonplayers are parallel in the range of correct response rate from 50% to 80%. Therefore, within this range, changing the criterion would have no marked influences on the conclusion of this study.
The learning of visual perception is generally limited to the environment in which the visual stimuli are presented (11,28). Therefore, if learning effect on the DVA of baseball players exists, it will be most likely to be detected when visual stimuli are presented in a way that is similar to the way in which balls move in a game. With regard to this, it is interesting to consider whether there is direction specificity in the visual perception of baseball players and how direction specificity is related to handedness. For training of DVA, a system like the one used in the present study might be used to give trainees repetitive stimuli that would resemble the moving pattern of a ball in the visual field that occurs during a baseball game or practice.
This work was supported by a Grant-in-Aid for Young Scientists (B) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan (KAKENHI:23700685) to Y. Uchida. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
The authors thank Dr. Larry Crawshaw for English editing.
The authors declare no conflict of interest.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Anstis SM. Chart demonstrating variations in acuity with retinal position. Vision Res
. 1974; 14 (7): 589–92.
2. Baechle T, Earle R, Wathen D. Anaerobic exercise prescription. In: Baechle T, Earle R, editors. Essentials of Strength Training and Conditioning
. Champaign: Human Kinetics; 2008. p. 379–80.
3. Bahill AT, Adler D, Stark L. Most naturally occurring human saccades have magnitudes of 15 degrees or less. Invest Ophthalmol
. 1975; 14 (6): 468–9.
4. Bahill AT, Clark MR, Stark L. Dynamic overshoot in saccadic eye movements is caused by neurological control signal reversals. Exp Neurol
. 1975; 48 (1): 107–22.
5. Bahill AT, Laritz T. Why can’t batters keep their eyes on the ball? Am Sci
. 1984; 72 (3): 249–53.
6. Bahill AT, Stark L. Trajectories of saccadic eye movements. Sci Am
. 1979; 240 (1): 108–17.
7. Brown B. Dynamic visual acuity, eye movements and peripheral acuity for moving targets. Vision Res
. 1972; 12 (2): 305–21.
8. Burg A, Hulbert S. Dynamic visual acuity as related to age, sex, and static acuity. J Appl Psychol
. 1961; 45 (2): 111–6.
9. Collewijn H, Erkelens CJ, Steinman RM. Binocular co-ordination of human horizontal saccadic eye movements. J Physiol
. 1988a; 404: 157–82.
10. Collewijn H, Erkelens CJ, Steinman RM. Binocular co-ordination of human vertical saccadic eye movements. J Physiol
. 1988b; 404: 183–97.
11. Gibson EJ. Perceptual learning. Annu Rev Psychol
. 1963; 14: 29–56.
12. Hopp JJ, Fuchs AF. The characteristics and neuronal substrate of saccadic eye movement plasticity. Prog Neurobiol
. 2004; 72 (1): 27–53.
13. Ishigaki H, Miyao M. Differences in dynamic visual acuity between athletes and nonathletes. Percept Mot Skills
. 1993; 77 (3): 835–9.
14. Iwamoto Y, Kaku Y. Saccade
adaptation as a model of learning in voluntary movements. Exp Brain Res
. 2010; 204 (2): 145–62.
15. Jacob R, Lillakas L, Irving EL. Dynamics of saccadic adaptation: differences between athletes and nonathletes. Optom Vis Sci
. 2005; 82 (12): 1060–5.
16. Keselman HJ. Testing treatment effects in repeated measures designs: an update for psychophysiological researchers. Psychophysiology
. 1998; 35 (4): 470–8.
17. Kohmura Y, Aoki K, Honda K, Yoshigi H, Sakuraba K. The relationship between dynamic visual acuity and saccadic eye movement. Hum Perform Meas
. 2008; 5: 23–30.
18. Land MF, McLeod P. From eye movements to actions: how batsmen hit the ball. Nat Neurosci
. 2000; 3 (12): 1340–5.
19. Linhart H, Zucchini W. Model Selection
Oxford, England: John Wiley & Sons. 1986. p. 189–91.
20. Luna B, Velanova K, Geier CF. Development of eye-movement control. Brain Cogn
. 2008; 68 (3): 293–308.
21. Meinecke C. Retinal eccentricity and the detection of targets. Psychol. Res.
1989; 51 (3): 107–16.
22. Miller JW, Ludvigh E. The effect of relative motion on visual acuity. Surv Ophthalmol
. 1962; 7: 83–116.
23. Millslagle DG. Dynamic visual acuity and coincidence-anticipation timing by experienced and inexperienced women players of fast pitch softball. Percept Mot Skills
. 2000; 90 (2): 498–504.
24. Morris GSD, Kreighbaum E. Dynamic visual acuity of varsity women volleyball and basketball players. Res Quart
. 1977; 48 (2): 480–3.
25. Olson CL. On choosing a test statistic in multivariate analysis of variance. Psychol Bull
. 1976; 83 (4): 579–86.
26. Rouse MW, DeLand P, Christian R, Hawley J. A comparison study of dynamic visual acuity between athletes and nonathletes. J Am Optom Assoc
. 1988; 59 (12): 946–50.
27. Schneiders AG, Sullivan SJ, Rathbone EJ, Thayer AL, Wallis LM, Wilson AE. Visual acuity in young elite motorsport athletes: a preliminary report. Phys Ther Sport
. 2010; 11 (2): 47–9.
28. Schoups A, Vogels R, Qian N, Orban G. Practising orientation identification improves orientation coding in V1 neurons. Nature
. 2001; 412 (6846): 549–53.
29. Uchida Y, Kudoh D, Honda M. A new quantitative method for the assessment of dynamic visual acuity. Hum Perform Meas
. 2010; 7: 1–6.
30. Wertheim T, Dunsky IL. Peripheral visual acuity. Am J Optom Physiol Opt
. 1980; 57 (12): 915–24.
31. Yamamoto S, Kitazawa S. Reversal of subjective temporal order due to arm crossing. Nat Neurosci
. 2001; 4 (12): 759–65.