Information Sources

Video or field-based temporal occlusion, where the time course of visual information is controlled through a black video frame or glasses that switch from clear to opaque, has been predominantly used to understand anticipation in sport [6]. Alternatively, in video-based spatial occlusion, selected segments of the simulated scene such as an opponent’s tennis racquet are masked (occluded) in order to determine the sources of information used for anticipation in sport [6]. Through video-based temporal and spatial occlusion, it is well established that expert performers are superior to novices in their capability to use advance (pre-object flight) postural (kinematic) cues in the movements of an opponent in order to anticipate [8]. Advance information is used for initial body positioning to strike an object, save a shot on goal, or change direction when defending [9,10,11]. In the past decade, more fine-grained evidence has emerged concerning how experts use advance information. Studies have reported that experts are superior to lesser skilled players in their capability to pick-up both contextual and kinematic information to anticipate [12, 13]. Contextual information can be visual or non-visual such as information from memory of field-placings in cricket batting or opponent action tendencies in soccer, respectively [10, 14]. It has been reported that experts have a superior capability for integrated pick-up of contextual and kinematic information, and this is magnified dependent upon whether there is congruence (i.e., both sources of information suggest the same outcome), rather than incongruence (i.e., one source of information suggests an outcome, while the other does not), between these sources [14]. The capability to integrate contextual and kinematic information is a crucial mechanism to guide initial body positioning for achievement of the skill goal [1, 15]. There is evidence, however, that some experts rely solely upon contextual information even if it is congruent with kinematic information [16]. This can be a major concern as overreliance on contextual information such as opponent action tendencies can make the performer susceptible to being deceived by the opponent’s kinematics [10]. For example, in baseball, when a pitcher is behind the count, a batter may predict a fastball will be thrown, so the batter may place less emphasis on kinematic cues, but the pitcher may throw a curveball or change-up. Therefore, the capability to switch between contextual and kinematic information to anticipate is critical because opponents can intentionally or unintentionally alter kinematics relative to their skill goal [17].

Options to Better Simulate Advance Cues

As observed by Drew [2], some existing VR simulators do not present advance kinematic cues with temporal occlusion to challenge users’ early visual information pick-up. This is concerning as experts use postural (kinematic) information to gain time for initial body positioning in a range of sports skills including cricket and baseball batting, returns of serve in tennis, and goalkeeping in soccer and field hockey, to name a few [1, 17]. A possible reason that kinematic cues are not properly simulated in VR is due to the difficulty associated with accurately modelling and animating in a synthetic experience the intricate details of the fingers, hand, and arm connections such as of a baseball pitcher or cricket bowler. To overcome this limitation, 360° stereoscopic video footage of a pitcher could be captured and presented with temporal or spatial occlusion in an HMD. In addition, contextual information such as the pitch count in baseball could be presented to be congruent or incongruent with kinematic and ball flight information. The user can be required to make a verbal, button press or simulated motor response to advance cues (see perception–action coupling section). These improved content design features of VR attempt to maintain presence and immersion, but importantly exposure to crucial advance cue information sources (psychological fidelity). Accordingly, the capability of the user to pick-up crucial advance cues with control over their presentation through temporal occlusion will help determine in a controlled manner, whether performers can use these advance sources of information that are relevant to body positioning [17]. Thereafter, temporal and spatial occlusion can be applied in combination during object flight to challenge pick-up of this information for object interception [17]. Therefore, including these features into VR content design would take into consideration the mechanism of expert anticipation that maps the pick-up of temporally evolving sources of visual information to components of the performer’s action response.

Point-Light and Blurred Displays

Video simulation of an opponent(s) movements have been displayed as points-of-light placed on anatomical landmarks with a black background or where the display has been blurred [9, 18]. The purpose of these methods is to remove featural information such as colour, contour, or shape, but preserve configural (relation) information [19]. Several video-based studies have reported that experts, but not lesser skilled players, use the minimal (configural) information from kinematics presented in point-light and blurred displays to anticipate [18, 20, 21]. This low-resolution visual information is suggested to be related to the peripheral visual field and dorsal visual stream that processes visual information at a fast sub-conscious level [22]. In a related manner, pioneering work by Abernethy and Russell [8] using video temporal and spatial occlusion, reported that whilst expert badminton players verbalised use of racquet-shuttle contact information to predict shuttle landing location, they actually relied upon information from the arm holding the racquet to strike the shuttle. Therefore, it appears that there is an intricate interplay between pick-up of contextual information (mentioned in previous section) at a more conscious level [e.g., 14], and pick-up of kinematic information at a sub-conscious level [8] for visual anticipation.

Options for Simulated Display Resolution

Drew [2] mentions the “low-resolution” avatar in a tennis VR simulator as a technological limitation. Indeed, the focus of several VR simulators is to present high-resolution displays such as the setting of an actual sports arena that create the performer’s presence relative to the real-world context [4, 23]. Whilst presence is important for user engagement in VR simulators [3], it is not crucial for anticipation performance and learning in sport. As mentioned above, it is important to note that experts rely upon low-resolution configural, rather than high-resolution featural, information, to anticipate. Accordingly, studies have compared visual-perception of handball thrower kinematics using point-light, wire-frame, no texture, and synthetic experiences in VR, with no differences reported in performance across the different experiences [24]. Therefore, VR simulators could be designed to incorporate a combination of high-resolution displays, as well as point-light and blurred displays to ensure that sub-conscious visual processes are targeted for the pick-up of advance and object flight information [24, 25]. Such manipulations applied to video simulation have shown improvements to anticipation in expert athletes [26] and novice [27] performers in sport, which can be manipulated in synthetic or video-based VR experiences. A combination of high- and low-resolution displays will ensure that presence, immersion, and interactivity are intricately balanced with psychological fidelity that targets the fast sub-conscious processes of anticipation.

Individual Differences

Whilst the studies discussed above have been useful, they have focused upon group-based comparisons, rather than individual differences within expert samples [28]. Drew [2] addresses this point effectively by discussing an individual differences study by Land and McLeod [29] that reported a professional cricket batsman made earlier predictive saccades to future ball landing positions, compared to amateurs, in order to strike balls. Predictive saccades position the eyes ahead of the ball to cope with time constraints for interception and further confirms that anticipatory control occurs in field-settings. More recently, using video temporal occlusion, it has been reported that there are differences between expert field hockey goalkeepers (of international and national level) in their capability to integrate congruent contextual and kinematic information to anticipate a drag-flick on goal [15]. This indicates that the mechanism to switch between contextual and kinematic information may not be well developed in all expert athletes. Accordingly, it is important to consider the individualised nature of anticipation skill, which can be dependent upon experience, expertise, and/or consistent exposure to practice environments that challenge the integrated pick-up of contextual and kinematic information [30]. Moreover, some athletes may have superior capability to pick-up visual information, but inferior motor response timing or vice versa. Therefore, an important goal of research is to identify why inter- and intra-individual differences exist, whilst for skill development it is important to target relative deficiency for training.

Level of Analysis in Simulated Contexts

Where possible, design of VR should incorporate nested individual comparisons for research and application in skill development. Due to the capability for body positional tracking in synthetic VR experiences, it is possible for anticipation outcomes such as response timing and response accuracy to be measured on a continuous scale. Whereas for video-based VR experiences, anticipation outcomes for response timing can still be measured on a continuous scale, with response accuracy on a dichotomous scale. This level of analysis in VR is valuable for three reasons. First, as competition skill level increases there is a smaller sample of genuine experts and it is challenging to recruit large samples for studies that include field-based assessments [31]. Adequate statistical power can be attained through sufficient individual experimental trials, which will allow investigation of individualised integration of contextual and kinematic information for anticipation [15]. Second, individual visual information pick-up and motor capability can be probed through combined temporal occlusion and response accuracy timing, respectively. This is useful to determine whether anticipation training should target visual-perceptual and/or motor responses for body positioning and interceptive phases of a skill. Third, practitioners including coaches, scouts, and sports scientists want to know the performance level of individual athletes for talent identification and/or skill development [28, 30]. Therefore, an individual differences approach can cater to research and skill development needs by targeting the underpinning mechanism of anticipation.

Perception–Action Coupling

There continues to be debate regarding the contribution of perception and/or overt action coupling to anticipation skill [32]. The literature is inconclusive in terms of whether visual-perception to overt action-coupled responses facilitate superior anticipation over perception-only responses. Some studies have reported superior anticipation based upon pick-up of advance kinematic information with an overt action response compared to a verbal (perception-only) response [33], whilst others have reported no advantage of perception to overt action coupling relative to perception-only responses [32, 34]. Moreover, studies have indicated that experts, but not less-skilled players, can predict above the guessing level based upon pick-up of kinematic information using a verbal response [35 see Experiment 1], and timing of visual information pick-up is consistent across verbal and overt action responses modes [36, 37 see introduction of latter]. Three important factors need to be considered to clarify how perception and/or overt action influences anticipation. First, the presentation of sports-specific kinematic information and its use relative to expertise is a key component of anticipatory skill [38]. Second, several neurophysiological studies have reported that sensory and motor regions of the brain are finely tuned, and in turn ‘coupled’, to sports-specific perceptual information when the performer is stationary [39, 40]. In a field setting of the motor skill, this neural coupling is, of course, necessary to prepare the body to initiate or inhibit action in a timely manner and for on-going action control [39]. Third, evidence indicates that the motor system is engaged (prior to action) to simulate the observed action in order to anticipate its outcome [41]. Therefore, perceptual (e.g., verbal) and/or motor responses are suitable for assessment and training anticipation skill [37 p2].

Flexibility in the Use of Perceptual and Motor Responses

Based upon the above, VR simulator content design can incorporate verbal, button press, and/or motor responses to sports-specific perceptual information. This may be necessary for two reasons. First, a verbal or button press response will allow easy use of commercial video-based VR experiences that can be cost-efficient and re-create sports-specific information, but are not sufficiently advanced like more expensive synthetic experiences to measure perceptual-motor positional or accuracy responses at high sampling frequencies [42, 43]. Second, task difficulty and challenge may require progression from perception-only to perception–action responses relative to the individual and skill level [44]. Synthetic VR experiences, however, provide greater control over visual stimuli with opportunity for tighter coupling of motor responses [42]. This is invaluable to several sports such as baseball or cricket, where batters can be assessed and trained against a variety of simulated opponents. It is noteworthy, however, that in other domains such as military and law enforcement, VR simulators indeed incorporate a combination of verbal, button press, or simplified motor responses [4]. The crucial design feature is to ensure that key visual-perceptual information is displayed (as mentioned earlier) to ensure that presence and immersion, as well as interaction (where relevant) are maintained so that a variety of instructional methods can be used to draw sub-conscious attention to these information sources for improved anticipation (see below).

Early and Late Visual Information for Action

The previously discussed content has considered the pick-up of early visual information for anticipation. Some field-based studies have also used spatial occlusion, which involves masked advance or object flight cues, with performance decrements relative to the unoccluded control condition indicating the importance of the masked (occluded) scene for anticipation [45]. Using spatial occlusion, studies have reported that experts require opponent kinematic information to make initial action responses to position their body [45]. Related field-based temporal occlusion studies have reported that experts are superior to lesser skilled players in their capability to use earlier occurring contextual [46] and kinematic [11] information to position their body. Together, these studies have confirmed that anticipatory control of gross body positioning occurs in the field-setting of the sport. Field-based temporal and spatial occlusion studies have also reported that experts use later occurring object flight information to fine-tune action during the interception phase [47]. This finding indicates that experts also use a prospective (pick-up of ongoing information) control strategy when time permits to adjust later occurring motor responses. Collectively, the findings of these field-based studies are in line with expert use of advance and object flight information to make object depth/direction (related to body positioning) and location (fine-tuning action) predictions in video-based temporal occlusion tasks, respectively [14, 38]. Therefore, it appears that expert performance involves an intricate blend of predictive and prospective control strategies [1]. Nonetheless, it is important to consider that in high time-constrained skills, if predictive control is not well developed, it is unlikely that there will be time for fine-tuning of action through prospective control [48].

Simulation of Early and Later Visual Information

To target both predictive and prospective control, VR simulators can be designed where temporal occlusion is applied immediately after presentation of advances cues, but also during object flight. This would ensure that performers must attend to both early visual information that maps onto gross body positioning and later visual information that maps onto fine-tuning of action [1]. Synthetic VR simulators offer a range of options where temporal and spatial occlusion can be applied throughout the course of an opponent’s action and object flight. For example, spatial occlusion can be applied to a segment of the pitcher’s body (early cues), with temporal occlusion applied to sections of only the ball’s flight (late information), whilst the surrounding synthetic experience is maintained [49]. This can maintain immersion, interaction, and presence within VR, whilst selectively manipulating early and late information used for anticipation. Currently, VR simulators are not consistently designed to probe this early and late visual-motor mapping [2], which can draw the user’s focus towards late object flight information to inform action that is a characteristic of lower skilled performers.

Learning and Transfer

A variety of instructional approaches have been used to improve anticipation skill. These include; temporal occlusion with an unoccluded replay [50, 51], verbal cueing to key kinematic cues [52], desensitisation through training under anxiety [53], use of point-light and blurred displays with an unoccluded replay [26, 27], above real-time training [54], visual-perceptual and motor practice of the observed action [26,