از کنترل توجه تا گسترش توجه: تحقیقات سطح مهارت توجه، حرکت و نتایج عملکرد

Abstract Two experiments examined the impact of attention on the movement and putting accuracy of novice and experienced golfers. In Experiment 1, attentional control was manipulated via two different secondary tasks: (i) an extraneous condition in which participants judged the frequency of an auditory cue presented during their stroke and, (ii) a skill-focused condition in which participants judged whether the cue occurred closer to the starting or end point of the swing segment in which it was presented. For experts, putting performance was least accurate in the skill-focused condition and when the cue was presented earlier. This decline in accuracy was associated with a significant reduction in the relationship between downswing amplitude and distance. Novices showed the opposite pattern. In Experiment 2, we manipulated attentional control indirectly by introducing the possibility that participants would stop their swing mid-stroke in response to an auditory cue, thus pushing participants to exert added control over step-by-step execution. Stop-trials were interleaved with normal putting trials in which no instructions were given. Novices were better able to stop their putting stroke and putted more accurately on non-stop trials than experts. These findings are consistent with recent models of putting control.

Introduction What makes skilled athletes different from their novice counterparts? Although the answer to this question commonly revolves around skill-level differences in performance outcomes (e.g., the score of a round of golf or one’s baseball batting average), some researchers have argued that it is the underlying cognitive control structures supporting performance that truly distinguish highly skilled individuals from their less skilled counterparts (Abernethy, Maxwell, Masters, van der Kamp, & Jackson, 2007). These control structures rely on particular forms of memory, vary in the demands they place on attention, and are thought to change as practice accumulates and skill proficiency increases. But, it is not just the cognitive demands of performance that distinguish novice individuals from those more skilled, movement patterns have been shown to vary as a function of skill level as well. For example, in golf, the downswing amplitude distinguishes between novice and expert golfers (Delay, Nougier, Orliaguet, & Coello, 1997). While expert golfers regulate their downswing amplitude to appropriately control club head force for different putting distances, novices do not show this amplitude-distance relation (see also Sim & Kim, 2010). Despite work examining skill-level differences in attentional control (e.g., Beilock et al., 2004 and Beilock et al., 2002) and movement differences in novice and skilled perceptual-motor performance ( Delay et al., 1997 and Egret et al., 2004), relatively few studies have explored how the attentional demands of performance directly relate to movement – and how this might differ as a function of skill level. Insight into this relation is important for developing a comprehensive understanding of what makes a novice performer different from his/her highly-skilled counterpart, and may also shed light on how to optimize skill learning and prevent skill breakdown (e.g., in pressure-filled high-stakes situations) once high-level performance has been achieved. 1.1. Expertise and attention Theories of skill acquisition suggest that performance proceeds through identifiably different phases as learning progresses that are characterized by changes in the cognitive processes governing execution and changes in performance itself. Although a number of different frameworks have been proposed to capture these skill level differences, in general, novice performance is thought to be based on explicitly retrievable declarative knowledge that is held in working memory and consciously attended in real time (Anderson, 1983, Anderson, 1993, Fitts and Posner, 1967 and Proctor and Dutta, 1995). As learning progresses, information is restructured into “procedures” or “programs” (Brown and Carr, 1989 and Keele, 1968). This new “proceduralized” skill representation does not mandate the same degree of attention and control that was necessary at lower levels of practice, and is supported by different neural structures than were active early in learning (Milton, Solodkin, Hlustik, & Small, 2007). The notion that different cognitive processes underlie various stages of skill development – with a trend toward increased proceduralization at higher levels of proficiency – carries implications for the types of attentional manipulations that may influence performance. For example, Beilock et al. (2002) found that introducing a secondary task involving monitoring a stream of auditory tones hurt the putting performance of novice golfers but had no effect on experts. Conversely, introducing a secondary task that required participants to monitor the position of the putter head improved novices’ putting accuracy but hurt expert performance. The finding that high-level skills are disrupted by attention directed toward processes that normally run outside conscious awareness (Beilock and Carr, 2001, Beilock et al., 2002, Lewis and Linder, 1997 and Masters, 1992; Masters, Polman, & Hammond, 1993) has also been reported for baseball batting (Castaneda and Gray, 2007 and Gray, 2004), golf chip shots (Perkins-Ceccato, Passmore, & Lee, 2003), field hockey (Jackson, Ashford, & Norsworthy, 2006), and soccer (Beilock et al., 2002). Indeed, these negative effects of enhanced attention can not only be seen in complex skills such as golf chipping and baseball batting, but in more basic skills we use every day. For example, it has been suggested that directing performers’ attention to their movements through “internal focus” feedback on a dynamic balance task interferes with the automated control processes that usually control balance movements (Wulf & Prinz, 2001). 1.2. Attention and movement Very few studies have explicitly manipulated performers’ attentional focus and measured the impact on movement patterns especially in the types of complex perceptual-motor skills mentioned above. Mullen and Hardy (2000) investigated the effects of attention on movement in golf putting for high and low skill golfers (as defined by a median split based on baseline putting data). Golfers ranged in handicap from 12-18. Three attentional groups were compared: a task-relevant group which repeated coaching instructions aloud as they performed each phase of the putting action, a task-irrelevant group which generated a random number every second as they putted, and a control group which putted normally. Range of motion (downswing plus follow-through) was significantly greater in the two task conditions compared to the control condition. There were also significant effects for downswing movement time (MT) and time to peak speed (TTPS). For downswing MT, times were shortest in the control condition and longest in the task-irrelevant condition. For TTPS, times were shortest in the control condition and longest in the task-relevant condition. There were no significant Skill Level × Attention Condition interactions. There were also no significant differences in acceleration profiles or multi-joint dynamics (i.e., cross correlations between joints). However, it should be noted that the attentional manipulations did not produce significant changes in putting performance in this study. One limitation of this study is that it is difficult to compare the two experimental tasks in terms of attentional control because the task-relevant task both directed attention to movement and provided guidance on how to putt successfully while the task-irrelevant task presumably only shifted attention away from putting. Studies that have investigated movement changes associated with performance in high-pressure or in high-anxiety situations may also provide some insight into the attention-movement relation. In well-learned perceptual-motor skills, high-pressure situations are thought to harm performance by prompting individuals to allocate explicit attention to proceduralized performance processes that are typically outside of working memory (e.g., Baumeister, 1984; reviewed in Beilock & Gray, 2007). Therefore, kinematic changes associated with performance pressure may be related to changes in attentional control. Pijpers and colleagues (Pijpers et al., 2005 and Pijpers et al., 2003) investigated kinematic changes associated with anxiety in rock climbing. Anxiety was manipulated by having participants climb at two different heights on an indoor climbing wall. Consistent with a freezing of degrees of freedom theory (Bernstein, 1967) when climbing high on the wall participants exhibited movements that were more rigid and less-fluent as compared to climbers at the low level on the wall. Specifically, the cross-correlation between joint angles was significantly higher in the high-anxiety condition. Gray (2004) measured the batting kinematics of skilled baseball players performing a simulated hitting task under baseline and pressure conditions (in which there were monetary incentives and social pressures to perform well). Relative to baseline performance, baseball players had far fewer hits (by 32% on average) under pressure, i.e., they “choked”. In terms of swing movements, batters also exhibited an increased amount of variability in the timing of the different stages of their swing under pressure as compared to baseline conditions. In particular, the standard deviation of the ratio of “wind-up” to “swing” phases (Welch, Banks, Cook, & Draovitch, 1995) was significantly increased under pressure. Similar increases in movement variability under pressure have also been reported in weightlifting (Collins, Jones, Fairweather, Doolan, & Priestley, 2001). Why might movement variability increase under stress? In Gray (2004), it was found that skilled batters were better able to monitor the direction their bat was moving under pressure as compared to baseline conditions – suggesting they were attending more to the step-by-step components of skill execution under high-pressure as compared to low-pressure conditions. If increased attention to well-learned execution creates the opportunity to adjust the execution of one’s skill in a way one might not normally do, this could lead to increased movement variability. In summary, previous research has provided only equivocal and/or indirect evidence for links between the attentional focus adopted by a performer, movement patterns and performance outcomes for complex motor actions. It is also not clear from previous studies whether this relationship is dependent on the performer’s level of expertise. Mullen and Hardy (2000) found an influence of attention for some movement variables but there were no associated changes in performance in their study. Furthermore, the effects of attention on kinematics did not differ as a function of skill level as would be predicted by theories of skill acquisition (e.g., Fitts & Posner, 1967). Since performance pressure is thought to cause a shift in the attentional focus (e.g., Baumeister, 1984 and Wulf and Prinz, 2001), one can infer that changes in kinematics resulting from pressure/anxiety manipulations are related to attention, however, this is only indirect evidence. In the current work we set out to directly test the relationship between attention and movement in golf putting and to investigate how this relationship varies as a function of expertise. Next we review research on motor control on golf putting and consider how attention may play a role in this action. 1.3. Motor control and attention in golf putting It has recently been proposed that expert golf putting involves both open loop and closed loop phases (Coello et al., 2000 and Craig et al., 2000). Specifically, a golfer is thought to control the putting stroke by specifying the downswing amplitude (i.e., the distance between the club head and ball at the end of the backswing) of the stroke prior to initiation based on the spatial parameters such as the distance to the hole, speed and slope of the green. The golfer thus executes a pre-programmed, open-loop backswing. During the downswing the movement of the club is continuously adjusted in response to the value of the optical variable τdeparture where this variable is defined as the optical angle between the current club head location and the location of the end of the swing (i.e., final follow-through position of the club head) divided by the rate of change of this angle. In other words, successful putting depends on continuously updating the movement of the club during the downswing and follow-through using visual feedback (though not necessarily an explicit process). The difference in control mode between the backswing and downswing phases of the golf putt suggests that their attentional demands may also be very different. In particular, it might be expected that attention to movement would interfere with the backswing (because it is pre-programmed) while not affecting the downswing (because it involves closed loop control). The control model of putting proposed by Delay et al. (1997) describes highly skilled performance. For novices, it has been proposed that skill execution is based on a set of un-integrated stimulus-response relationships. These relationships involve declarative knowledge about the action that is held in working memory (e.g., “when I start my backswing I need to make sure my wrists are not bent”, “when I begin my downswing I need to make sure the putter head is straight”, etc.). Performance thus requires constant online monitoring and attention to skill execution (for all phases of the swing) so that each piece of declarative knowledge can be utilized at the appropriate time during the action. Because these stimulus-response relationships are un-integrated, movements are highly variable and cannot be described by simple control laws. Consistent with this idea, Delay et al. (1997) reported that the relationship between downswing amplitude and distance was weaker in novices than experts. This description of novice control suggests that attention to movement should enhance all phases of the putting stroke (since they require attention) while attention to extraneous, external stimuli should interfere with both performance and movement. 1.4. Current work In Experiment 1, novice and skilled golfers took an series of putts under two different secondary task conditions: (i) an extraneous condition in which participants judged the frequency of an auditory cue presented during their stroke and, (ii) a skill-focused condition in which participants judged whether an auditory cue occurred closer to the starting or end point of a particular swing segment (e.g., backswing) in which it was presented. Both performance outcomes (i.e., putting accuracy) and movement variables were measured for expert and novice golfers. This experiment was designed to expand on the attention-movement findings of Mullen and Hardy (2000) by: (i) using more comparable secondary tasks, and (ii) comparing groups with a larger difference in skill level, namely true novices and experienced golfers. Gray (2004) used similar attention conditions for baseball batting. In the present study, we sought to extend this work to a different sport and different measures of movement. Experiment 1 was also designed to expand on this previous work by examining how the effect of the secondary task changes as a function of when the task stimulus is presented relative to the start of the movement. As described above, putting in experts is thought to involve distinct phases which utilize different control modes, therefore, it is likely that there would be large differences in the effect a secondary task depending on its timing. As discussed above, it has been reported that expert golfers primarily regulate the downswing amplitude (rather than movement time or club head speed) to appropriately control club head force for different distances, with differences in the downswing amplitude (as a function of hole distance) as the main kinematic variable that distinguishes novice and expert golfers (Delay et al., 1997). Therefore, our analysis focused on downswing amplitude (as defined as the distance the putter head traveled between the highest position of the club during the backswing and the highest position of club after contact with the ball), although we also analyzed other movement variables as described below. On the basis of previous research we sought to test the following hypotheses: (i) For experts, putting accuracy would be significantly higher in the extraneous condition than in the skill-focus condition. (ii) For experts, the relationship between downswing amplitude and putting distance would be significantly stronger in the extraneous condition as compared to the skill-focused condition. (iii) For experts, the effects of the skill-focus condition on putting accuracy and the DS amplitude-distance relationship would be significantly larger when the task stimulus was presented in the pre-programmed part of the movement (i.e., the backswing) than in the continuously controlled phase of the movement (i.e., the downswing). (iv) For novices, putting accuracy would be significantly higher in the skill-focus condition than in the extraneous condition. (v) For novices, the relationship between DS amplitude and putting distance would not be significantly different in the two attention conditions. (vi) For novices, putting accuracy and the DS amplitude-distance relationship would not be significantly different for conditions in which the auditory cue was presented in the backswing versus when it was presented in the downswing. In Experiment 2, we explored the attentional demands of skill execution and its relation to movement in a different way – specifically, by comparing the relative capability of novice and skilled golfers to stop their putt mid-stroke in response to an auditory cue. If highly-skilled performances are controlled by proceduralized processes that operate largely outside of attentional control in a way that novice performance is not, then skilled golfers may actually be worse than novices at stopping their swing mid-stroke. Moreover, as mentioned above, if attention to one instance of performance carries implications for another, then it may be that being told to stop one’s swing on certain putting trials may impact performance on other trials where this does not occur. Such a result would suggest that attentional control during a particular instance of performance not only impacts execution of a skill, but can spillover onto other performances as well. In Experiment 2 we were also again interested in the effect the timing of the stop cue would have on movement. Previous research on simple motor actions (e.g., reaching) has shown that the ability to stop an action after it has been initiated is highly dependent on the stimulus onset asynchrony (SOA) between the start and stop signals; inhibition is less successful when the stop signal is presented later in the movement. This effect has been modeled as a race between independent stochastic processes responsible for producing and inhibiting the action (Logan, Cowan, & Davis, 1984). Gray (2009) recently extended this race model to explain inhibiting (“checking”) a swing in baseball batting. Therefore, on one hand, it might be expected that stopping a golf putt in mid-stroke may show a similar effect. Namely, the distance required to stop the putter movement would be longer when the stop signal is presented later. On the other, the effect of SOA on stopping success may be determined by differences in modes of control for the different swing segments: experts may be more effective at stopping the putter when the signal is presented in the closed-loop downswing as opposed to the open-loop backswing. On the basis of previous research we sought to test the following hypotheses: (i) The distance required to stop the putting stroke would be significantly shorter for novices than experts. (ii) For experts, that ability to stop the putting stroke (as measured by stopping distance, described below) would be significantly better when the auditory cue was presented in the downswing in comparison to when it was presented in the backswing. (iii) For experts, introducing the task of trying to stop the putting stroke on some trials would significantly impair performance for putts in which stopping was not required (relative to baseline) because it would effectively induce skill-focused attention. (iv) For experts, the relationship between DS amplitude and putting distance for non-stop trials would be significantly weaker relative to baseline. (v) For novices, the ability to stop the putting stroke would not be significantly different when the auditory cue was presented in the backswing versus the downswing. (vi) For novices, introducing the task of trying to stop the putting stroke on some trials would significantly improve performance for putts in which stopping was not required (relative to baseline). (vii) For novices, the relationship between downswing amplitude and putting distance would not be significantly different for non-stop trials versus baseline putting.