Target Box Stepping Accuracy and Variability

As can be seen from the Berg balance scale, FES-I and ABC results presented in Table 4-1, the participants of this study were relatively competent and confident walkers, and

would be considered to be at a low-risk of falling. Despite this, during the social evaluative threat trials, we observed a significant increase in mediolateral stepping error, and a reduction in anteroposterior step variability. Our findings that the higher anxiety trials produced more medial steps suggest an anxiety-based stiffening strategy which is characteristic of older adults who have experienced a fall. This provides support that state anxiety contributes to poor stepping performance (Young et al., 2011). We also observed more medial stepping in the more complex OO trials compared to TO. This could be due to the increased attentional demand necessary in OO trials, resulting in poorer stepping accuracy (Gage et al., 2003).

Our finding that increased anxiety reduced anteroposterior step variability suggests that social evaluative threat made participants step at more consistent lengths during the target step. This could have been a product of the intervention which explicitly told participants to step with greater consistency and precision. However, there was no main effect of session on A/P error. The mean values of mean A/P error showed minimal change from -22.6 ± 2.4mm during judged trials, compared to -21.9 ± 2.5mm during control. The reduction in A/P variability during judged trials suggests that they were stepping inaccurately with less variability under social evaluative threat. This might be due to an anxiety-mediated postural stiffening reducing the range of motion of the ankle and reducing acute variability of the foot placement in the direction of stepping (Brown et al., 2002), however we did not measure lower leg EMG to confirm this theory.

We observed a decrease in anteroposterior step accuracy as trial difficulty increased (Figure 4-3b), which shows the expected effects of increased attentional demand reducing stepping accuracy. Young et al. (2011) suggest that the effects of task

difficulty may be due to increased anxiety. Our results support this, but our anxiety measured lacked the clarity to show direct evidence of this.

There was a significant reduction in trailing toe clearance over the first obstacle from 14.8 ±1.8mm in judged trials, compared to 13.3 ±1.9mm in control. Although this difference is small, and unlikely to cause a fall in a real world situation, the finding that acute psychological anxiety reduces the clearance of the trailing toe could be detrimental to those with more severe anxiety, assuming that toe clearance might reduce further with greater anxiety. Trajectory of the trailing foot is important in planning the next step, and if it were held back unexpectedly the forward momentum might cause an individual to continue forwards without the trailing foot being able to prevent a fall. Di Fabio et al. (2004) demonstrated that older adults at a high-risk of falling showed reduced trailing foot clearance compared to low-risk older adults and young adults. They suggested the observed foot lift asymmetry could be due to limited hip extension or to reduced executive cognitive function. Here we have demonstrated a reduction in trailing toe clearance in low-risk older adults, and provided evidence of state anxiety being another contributing factor due to fear-of-falling being commonly present in high- risk older adults (Friedman et al., 2002). We speculate that anxiety drives this behaviour through attentional prioritisation of future tasks, much like that demonstrated during target stepping by Young et al. (2011).

There were task difficulty related changes in toe-obstacle clearance also for both leading and trailing feet. Participants cleared the obstacles with less space during BO compared to OO trials. The increased attentional demands of the BO condition could have reduced toe clearance, however this could also have been due to the layout of the course. There was a relatively small space between obstacles (100cm) in the BO

condition where participants had to step once with each foot in order to step over the second obstacle with a leading right step. This could have contributed to the decrease in toe clearance observed during BO trials, and therefore a direct comparison between task difficulties would be confounded.

The percentage of trials in which toe-first stepping occurred was much lower during TO trials than the other two task difficulties. To our knowledge this is the first study to look at step technique while stepping into a raised target. During regular gait, heel-first foot contact is the normal method of stance initiation, however the postural threat of the target box (see Chapter 2.2, Figure 2.2) interrupted the normal gait pattern and participants either chose to continue with the heel-first gait technique, or step with their toe first for reasons that are currently unclear. It is our assumption that participants felt there was a benefit to this toe-first stepping, probably due to an increased ability to visually judge the anteroposterior distance from the front inside edge of the target box to the front edge of their foot. The increase of its occurrence during more complex tasks suggests that there are attentional processes involved and that it might be a more cautious method of stepping when planning additional steps. The occurrence of this technique was also negatively correlated with stance duration in the target box indicating that stance duration was generally shorter when this technique was more common (Figure 4-4b). This could have been due to heel-first steps involving a transfer of pressure along the length of the foot, compared to maintaining pressure in the ball of the foot during steps using the toe-first technique.