The goal of the current study was to determine how the balance between attentional control systems and the inhibition function impact a preplanned goal-directed motor task in individuals with high levels of trait anxiety. Three novel contributions emerged from these findings: 1) High anxiety was associated with attenuated performance efficiency but not performance effectiveness. 2) Threat stimuli lead to faster RTs in high and low anxiety groups. 3) Between group differences were evidenced in the hypothesized direction at relatively low but not at relatively high force levels. Each of these findings is addressed in detail below.
ACT (Eysenck et al., 2007
) postulates that high anxiety leads to a decreased ability to inhibit the stimulus-driven attentional system when the goal-directed attentional system is concerned with executing fast and accurate performance. This imbalance is hypothesized to manifest in slower RTs (Arnell et al., 2007
; Blair et al., 2007
), a prediction that drove our first hypothesis. RT findings with the target force set at 10% of MVC supported this hypothesis. This finding was noted in our primary analyses and was confirmed by the secondary analyses when controlling for between group differences in MVC.
We did not expect movement accuracy to differ between groups. However, the primary analysis revealed that although slower to initiate movements, the high anxiety group was more accurate, and therefore displayed the classic speed-accuracy trade-off (Fitts, 1954
). Such a classic speed-accuracy trade-off may have been driven by between group differences in strategic concerns/regulatory focus (Forster, Higgins, & Bianco, 2003
). Previous evidence suggests that highly anxious individuals have a greater tendency to implement a prevention oriented regulatory focus (Forster & Higgins, 2005
), and these initial findings support this assertion. The high anxiety group may have employed a prevention regulatory focus rather than a promotion regulatory focus, emphasizing movement accuracy at the sacrifice of movement speed.
Prior to embracing a regulatory focus interpretation, however, secondary analyses were conducted to ensure that MVC, and thus target force level between groups were not driving this speed-accuracy trade-off. Indeed, although MVC between groups was not statistically significant, the low anxiety group had higher MVCs across both target force levels. Relatively higher MVC levels led to higher target forces which in turn typically lead to greater error and an increased rate of change of force production which could have potentially driven between group differences in our primary analyses. This finding is exemplified in and , where decreases in absolute response accuracy and increases in absolute response vigor are evidenced when the target force was set at 35% of MVC as compared to 10% of MVC. Once MVC levels were controlled within our secondary analyses, between group accuracy scores were no longer different. In consequence, performance accuracy did not differ between groups which suggested that as theorized, performance efficiency is compromised in high as compared to low trait anxiety, but performance effectiveness is not. Importantly, this finding only held true at the low target force level (to be discussed below). As such, we interpret our data as supporting and extending the scope of ACT (Eysenck et al., 2007
), suggesting that following presentation of distractor visual cues, highly anxious individuals are less able to inhibit the influence of the stimulus-driven attentional system on the goal-directed attentional system during a discrete preplanned goal-directed motor task.
Rate of change of force production, reflecting performance vigor, remained unchanged across analyses, suggesting attenuated vigor in the high anxiety group as compared to the low anxiety group at the 10% MVC force level. Although this finding was not expected, the secondary analyses suggested that this effect was partially driven by the lower MVC scores of the high anxiety group. Additionally, this finding may have also been driven by differences in regulatory focus and a general increase in inhibition. Specifically, attenuated vigor is in line with a more prevention biased regulatory focus (Forster et al., 2003
) as well as an increase in inhibition (Eysenck et al., 2007
). Clearly, further research is needed to clarify the effect of anxiety on the rate of change of force production.
Concerning our second hypothesis, we anticipated that both groups would be faster following the presentation of threat as compared to non-threat images. With the target force set at 10% of MVC our data supported this hypothesis and in doing so corroborated previous evidence that threat images prime the motor system, leading to faster RTs (Coombes et al., 2007a
). Importantly, there was no effect of valence on performance accuracy, suggesting that the emotion driven priming of the motor system had no influence on the accuracy of force production. Moreover, in previous studies, increases in the speed of movement during exposure to unpleasant images have been evidenced only when the movement direction and emotional state are congruent (i.e., extension movement + unpleasant image). In the present case, movements were not direction specific (i.e., towards or away from the body), thus suggesting that when no conflict (or congruence) exists between the emotional state and the movement direction, unpleasant emotional states facilitate the speed of contractions. Interestingly, given that previous evidence shows that sustained pinch-grip tasks are equally altered by pleasant and unpleasant states as compared to neutral states (Coombes, Gamble, Cauraugh, & Janelle, 2008
), we argue that this specific unpleasant priming may only hold true for short duration discrete movements. Nevertheless, this finding further validates the notion that emotional states alter activity within the motor system (Coombes, Cauraugh, & Janelle, 2006
; Coombes, Janelle, & Duley, 2005
; de Gelder, Snyder, Greve, Gerard, & Hadjikhani, 2004
; Hajcak et al., 2007
; Marsh, Ambady, & Kleck, 2005
; Pessiglione et al., 2007
Finally, we found that group differences were only evidenced at low as compared to high target force levels. Three potential explanations may account for this finding. First, effects may not have been similar across target force levels because different subjects composed each group at each target force level. However, this would suggest that at the high target force level, either the low anxiety group performed as one would expect the high anxiety group to perform, or vice versa. While acknowledging this as a possible explanation we argue that it is highly unlikely given the polarization between trait anxiety scores between groups at each target force level. Dissimilar findings across target force level suggest that high anxiety will not always lead to an increase in reaction time and that anxiety driven changes in reaction time are modulated by task demands. Thus, a second possible explanation is task difficulty. Our data suggest that the target force level of 10% MVC was more difficult and required greater planning to achieve greater precision. Consequently, the increased demand on the working memory system was more susceptible to interference from the stimulus driven attentional system. Previous fMRI evidence supports this position. For instance, Ehrsson, Fagergren, and Forssberg (2001)
provided evidence that force production at relatively low precision grip forces, relative to larger grip forces, more strongly activated several sensory and motor related fronto-parietal areas involved in higher order sensorimotor integration during the planning and execution of goal-directed actions. Moreover, the average number of practice trials to reach the qualifying pre-experimental performance level suggested that the criterion was easier to reach when the target force level was set at 35% of MVC, and this pattern was stable across groups (see ). Finally, a third explanation is that limited power may have resulted in the null findings at 35% of MVC. Although power analyses were conducted prior to recruitment, they were based on pilot data which probed performance at the 10% target force level. Hence, the null findings at the 35% target force level may well have been the result of a small sample size and limited power.
Greater specification of performance differences as an interaction of trait anxiety and valence of distractor cues may require the manipulation of cue exposure length. Previous evidence suggests that conscious and unconscious processing of threatening information represents distinct operations that are thought to be associated with different neural and behavioral responses (van Honk, Schutter, d' Alfonso, Kessels, & de Haan, 2002
; Williams et al., 2006
). Indeed, changes in neural and behavioral function in response to unconscious threat appear to be most prominent in individuals with high levels of trait anxiety (Etkin et al., 2004
; Fox et al., 2001
), which supports the link between trait anxiety and negative affectivity. Therefore, future research should investigate the role that level of awareness (i.e., subliminal vs. supraliminal) of distractor cues may play in modulating the effect of trait anxiety on a ballistic goal-directed motor task. The current findings could also be extended by exploring the effect of emotional distractors on more complex and dynamic motor planning processes, which could be accomplished by alternating the target force level within the same protocol. Further, given the meaningful effects that we have evidenced in a relatively small sub-clinical sample, the replication of these findings in larger clinically anxious samples may accelerate the formulation of practical recommendations related to the clinical assessment of motor performance efficiency in highly trait-anxious individuals.
Additionally, future research efforts could extend our behavioral findings by integrating attentional, neurological, and self-report measures. Although we argue that high anxious individuals’ slower RTs resulted from an inability to inhibit the stimulus driven attentional system, we did not directly assess attention allocation to disparate image content, and we acknowledge this inferential limitation. While we do not believe this limitation compromises the current evidence, we encourage future researchers to verify our behavioral findings with the use of additional attention measures such as eyetracking (e.g., Calvo, Nummenmaa, & Hyona, 2007
Finally, as previously mentioned, performance differences could have been a product of disparate self regulatory focus between high and low anxiety groups. Although secondary analyses suggested this was not the case, the potential emergence of a speed-accuracy tradeoff which is driven by regulatory focus should not be ruled out in future work given that high anxiety is associated with a prevention regulatory focus (Forster & Higgins, 2005
). Thus, future experiments should include self regulatory focus indices if we are to (a) understand how self-regulatory focus impacts the balance between the attentional control systems, and (b) better understand how this (im)balance impacts the motor system.
In conclusion, guided by predictions derived from ACT (Eysenck et al., 2007
), the current study investigated how individual differences in trait anxiety alter the planning of force specific motor tasks under varying emotional states. The new findings presented show that motor efficiency, but not motor effectiveness was compromised in high anxious relative to low anxious individuals, but only when the motor task to be performed required greater precision and increased attentional resources. We argue that decreased performance efficiency was driven by attenuated inhibition in the high anxiety group which led to enhanced stimulus driven attentional control at the expense of goal driven attentional control. Finally, our data validate the hypothesis that unpleasant emotional states prime the motor system and do so irrespective of dispositional trait anxiety levels.