IntroductionPrevious research has documented the characteristics of both normal and pathological gait in unobstructed level walking (e.g., Grieve and Gear, 1966; Winter, 1984). Only recently has there been interest in modulations to gait in negotiating obstacles. This aspect of gait is important because in the real world our path is usually obstructed. Patla and co-workers (Patla, et al., 1991) have examined modulations to gait in stepping over obstacles using force platform and electromyographic variables. The focus of the experiment reported here was how individuals change stride phase characteristics and lower limb trajectories when stepping over obstacles of various heights. Obstacle clearance data were used to suggest the process of limb control and the conditions under which individuals might trip or falter.MethodologyTwelve unimpaired, healthy, adult volunteers (6 male and 6 female, aged 23yr to 59yr) served as subjects, all had normal or corrected visual acuity to 6. OPTOTRAK 3000 (Northern Digital Inc.) markers were attached to the obstacle, and to the heel and toe of the right and left shoes. The obstacle was a 10 mm diameter metal rod 825 mm in length clamped to a metal stand. The mean of three recordings of sitting height, standing height, and knee height were used to calculate leg length. Subjects walked a straight level path of approximately 9 m parallel to the apparatus at a natural cadence, ''... just as you would walking around the house''. The obstacle was positioned 5 m from the start of the path directly in front of the OPTOTRAK sensors. Following practice trials there were 16 experimental trials in each condition, first walking normally unobstructed and then traversing obstacles at 10%, 25%, and 40% of leg length. Presentation of the 3 heights was completely counterbalanced across subjects. Smoothed x,y,z coordinates of the markers were obtained for one complete lead and trail leg stride for each trial at each obstacle height. From these data the effect of obstacle height on obstacle clearance and stride phase measures was determined using ANOVA procedures.ResultsThe clearance data in Table 1 indicate vertical clearance by the heel of the lead and trail foot at bar crossing. Clerance is presented as a multiple of the bar height. Mean bar heights for males and females respectively were: 10%, 85.8 cm and 80.5 cm; 25%, 214.3 cm and 201.7 cm; and at 40% 342.5 cm and 322.2 cm. The crossing time and distance data indicate the time and position in the normalized stride at which the bar was crossed. While heel clearances for males and females were different, the ratio of trail to lead foot clearance was similar for each bar height. The clearance ratios for males and females respectively were; 10% 2.6 (4.6) and 2.6; 25% 2.3 and 2.1; and 40% 2.9 and 2.8. Obstacle crossing distance appeared uninfluenced by bar height but the crossing time systematically decreased for the lead foot and increased for the trial foot across conditions. To confirm these observations ANOVAs were conducted with Gender (male & female), Foot (lead & trail), and Condition (10%, 25%, & 40%) the factors.For clearance and crossing time all main effects were significant (p < .01) but there were also reliable 3-way interactions. Post-hoc tests (Tukey) were conducted to identify the reliable pairwise comparisons starred in Table 1. For crossing distance there were reliable Gender and Foot effects, but a lack of effect for Condition confirmed that crossing distance was invariant with increased obstacle height. The clearance data indicate that gender differences were associated with trail foot clearance only, except at the 10% bar height. Note that at 10% and 40% the trail foot clearance means are significantly lower for females. The crossing time data further support the observation of gender differences in trail foot control, with the female crossing time consistently earlier in the normalized heel-strike to heel-strike duration.DiscussionIn stepping over an obstacle the lead foot can be visually guided. The data presented above suggest a mechanism in which the non-visually guided trail foot is controlled via a constant, or ''gain'', based on lead foot clearance. These data suggest a gain of 2.1 to 2.9. This process could be viewed as a safety mechanism, whereby the unsighted trail foot is provided a greater margin for error in stepping over an obstacle. The finding that crossing distance is independent of obstacle height implies that foot placement is modified on approaching an obstacle to maintain crossing point constant. The crossing point might be that which confers economy of effort in limb control in addition to maintaining a safety margin. Increased obstacle height is also associated with a decrease in lead foot crossing time, whereas the trail foot is given more time because the obstacle is crossed later in the stride. The data also suggest gender differences in foot control, in particular control of the trail foot trajectory. Identifying the cause of this difference would be an interesting avenue for future research, as would the broader issue of control of limb trajectory in obstacle clearance. Examination of lead and trail foot trajectories in elderly persons might contribute to understanding the cause of falls.