Men wore spandex shorts with no shirt and women wore a dance leotard with open back and performed two different tasks with bare feet: level ground walking and stair descent. Participants were in bare feet to minimize the effect shoes have on the kinematics. The task order was randomly selected for each subject, and the total task duration was not extensive enough to induce fatigue.²²

The walking task required participants to walk along a 10 m long walkway (Fig. 1(A)). During stair descent, participants were instructed to initiate from the back end of an elevated walkway, descend a three-step staircase, and continue walking for several steps (Fig. 1(B)). Participants were instructed to perform both tasks at their normal walking speed.

Two different force plate configurations were utilized to detect gait events such as heel strike (HS) and toe off (TO), as quantified through measures using vertical ground reaction force (GRFv). HS was determined to occur when the GRFv was greater than 10% of the maximum GRFv, and TO was determined to occur when the GRFv was less than 10% of the maximum GRFv.^23–25 Three force plates (FPs) (Advanced Mechanical Technologies, Inc., Watertown, MA) were placed in series and embedded level into the laboratory floor for the walking task. The first two FPs were immediately adjacent to one another, and the third plate was separated by a distance of 15 cm (Fig. 1(A)). Four FPs were used during stair descent trials. Two FPs were embedded level into the laboratory floor and two made up the steps²⁶ (Fig. 1(B)).

Three-dimensional marker position data were collected at 60 Hz with a ten-camera Eagle motion analysis system (Motion Analysis Corporation, Santa Rosa, CA). Sixty-two retro-reflective markers (diameter = 14 mm) were placed on the subject as described by Breloff and Chou⁵ (Fig. 2).

Five successful trials from each task were analyzed. Walking trials were defined as the time interval between two consecutive ipsilateral HSs (FP 3 to FP 1; Fig. 1(A)) and stair descent trials were consecutive ipsilateral HSs following first step down toward the ground level (FP 4 to FP 2; Fig. 1(B)). A successful trial consisted of clean FP striking from both participants’ feet as well as the right foot striking the first FP. To ensure successful trials, several practice trials were completed — in which the starting point was altered to accommodate each participant’s gait. A MATLAB® (Mathworks, Natick, MA) program calculated segmental trunk angles using adjacent trunk levels⁵ — sacrum-to-lower lumbar, lower lumbar-to-upper lumbar and upper lumbar-to-lower thorax (Fig. 2).

Outcome measures for the current study were: peak trunk angle, peak angle index and ROM. Peak trunk angles, timing of the peak trunk angle and trunk ROM have been previously used to describe spinal motion during ambulatory ADL.^27–34 Additionally, ROM has been used to detect aging effects in the spine during gait.^27,35–37 All three outcome measures were recorded for the three different planes of motion. Peak trunk flexion — forward/anterior bending of the trunk — was calculated in the sagittal plane. Frontal plane motion was represented by peak trunk ipsilateral bending — which indicates the trunk is bending toward the initial contact leg, or the ipsilateral side. Peak trunk contralateral axial rotation is presented for the trans-verse plane, which designates following the initial HS; the spine will twist away from the striking leg or toward the contralateral leg.

Each outcome measure was analyzed using three-way — 4(age) × 3(trunk level) × 2(task) — mixed-effects analysis of variance for each plane. If the three-way interaction was not significant, then two-way ANOVA’s explored differences between factors. The level of significance for these statistical tests was set at 0.05. The statistical software SPSS (version 18, IBM., New York, NY) was used for all statistical analyses.

Flexion and contralateral rotation maximum trunk angles during gait changed as a result of the different tasks. Lower lumbar-to-upper lumbar (4.14° ± 0.82°) was significantly larger than upper lumbar-to-lower thorax (1.59° ± 0.59°) during stair descent in the course of contralateral rotation. Introduction of a task other than walking did alter peak trunk angles in the sagittal and transverse plane (Table 3; Fig. 3).

The timing of the peak flexion angle was used to determine what gait events were occurring when segmental trunk peak flexion was produced (Fig. 4). In all age groups (20’s, 30’s, 40’s and 50’s) the peak flexion angle at the sacrum-to-lower lumbar joint during walking occurred before contralateral HS into the weight acceptance phase on the contralateral limb (approximately 45% to 51% of the gait cycle [GC]). The lower lumbarto-upper lumbar joint produced maximum angle during the GC for all age group cohorts during weight acceptance and into mid-stance of the contralateral foot (approximately 55% to 65% of the GC). In the upper lumbar-to-lower thorax joint during walking, the 20, 30 and 40 year old clusters of individuals found the maximum angle during the GC flexion angle to occur at the end of the heel rocker into the ankle rocker of the ipsilateral leg (approximately 57% of the GC). The 50 year old cohort had the maximum angle during the GC flexion angle during mid-stance (foot-flat) of the contralateral leg.

The stair descent task incurred different peak indices than walking — though these were not statistically significant. For the 20 and 30 year old cohort, the maximum angle during the GC flexion angle occurred during mid-swing to just before HS of the contralateral limb (approximately 40% of the GC). The older cohort of participants (40 and 50 years) had a maximum angle during the GC flexion in the sacrum-to-lower lumbar joint during mid-stance of the ipsilateral foot (approximately 23% to 28% of the GC). All age groups had a maximum angle during the GC just before the contralateral foot contacted the step to mid-stance (approximately 37% of the GC) in the lower lumbar-to-upper lumbar joint. The maximum angle during the GC for upper lumbar-to-lower thorax angle for the 20’s 40’s and 50 year old-groups occurred during weight acceptance until just before foot-flat on the contralateral leg (approximately 56% of the GC). However, the 30 year old group produced a maximum flexion angle during the GC during mid-stance (approximately 38% of the GC) of the ipsilateral leg (Fig. 4).

All age groups had a maximum peak ipsilateral bending angle during terminal swing into the loading phase on the contralateral foot at the sacrum-to-lower lumbar joint during the walking task (approximately 54% of the GC). In the lower lumbar-to-upper lumbar joint, all age cohorts presented with a maximum ipsilateral bending angle during terminal swing and into initial contact loading on the contralateral limb (approximately 53% of the GC). The upper lumbar-to-lower thorax joint produced a peak ipsilateral bending angle for all age groups between initial HS of the contralateral foot until mid-stance of the contralateral foot.

All age cohorts had a maximum ipsilateral bending angle during gait between the mid-swing and terminal swing of the contralateral leg during the stair descent task in the sacrum-to-lower lumbar joint. The 20 and 40 year old cohorts at the lower lumbar-to-upper lumbar joint had a maximum ipsilateral bending angle during terminal swing of the contralateral limb (approximately 43% of the GC). The 30 and 40 year old clusters had a maximum ipsilateral bending angle during mid-swing of the contralateral leg (approximately 35% of the GC). The upper lumbar-to-lower thorax joint produced a peak ipsilateral bending angle between mid-swing to foot-flat on the contralateral leg in all age groups (Fig. 4).

During the walking task, the 20 and 30 year old cohorts had a maximum contralateral axial rotation angle in terminal swing of the contralateral limb at the sacrum-to-lower lumbar joint for all age groups (approximately 45% of the GC). The 20 and 30 year old groups had a maximum contralateral axial rotation angle during initial contact of the contralateral leg in the lower lumbar-to-upper lumbar and upper lumbarto-lower thorax joints (approximately 52% of the GC). The 40 year old cohort’s maximum contralateral axial rotation angle was found during terminal swing of the contralateral limb in the lower lumbar-to-upper lumbar and upper lumbar-to-lower thorax joints (approximately 47% of the GC). The 50 year old group’s maximum contralateral axial rotation angle in the lower lumbarto-upper lumbar and upper lumbar-to-lower thorax joints was during the heel rocker phase following contralateral HS (approximately 56% of the GC).

The 20, 30, 40 and 50 year old cohorts had a maximum contralateral axial rotation angle during the terminal swing to foot-flat of the contralateral leg at the sacrum-to-lower lumbar joint in the stair descent task (approximately 50% of the GC). In the lower lumbar-to-upper lumbar joint during stair descent, a maximum contralateral axial rotation angle during mid-swing to terminal swing of the contralateral limb was produced in all age groups (approximately 44% of the GC). The upper lumbar-to-lower thorax joint produced a peak ipsilateral bending angle between terminal swing to foot-flat on the contralateral limb (Fig. 4).

Age group did not have any effect on the ROM of trunk kinematics (flexion, ipsilateral bending, and contralateral rotation) — indicating an individual’s age does not alter the amount of motion in their trunk.

The differences in trunk ROM were present at different trunk levels (sacrum-to-lower lumbar, lower lumbar-to-upper lumbar, or upper lumbar-to-lower thorax) in the different planes of motion. Pairwise comparisons on contralateral rotation ROM revealed that sacrum-to-lower lumbar (11.74° ± 3.01°) had larger ROM during walking when compared to stair descent (6.47° ± 1.19°, as in Fig. 5). Trunk level had a main effect in trunk flexion ROM (p < 0:001) and trunk ipsilateral bending ROM (p = 0:003). Trunk flexion ROM follow up pairwise comparisons of the trunk level marginal means found sacrum-to-lower lumbar (11.78° ± 10.95°) to have significantly more ROM than upper lumbar-to-lower thorax (4.93° ± 2.17°). Additionally, it was found that sacrum-to-lower lumbar had significantly more ROM than lower lumbar-to-upper lumbar (9.42° ± 7.56°, as in Fig. 5(B)). Trunk ipsilateral bending ROM post hoc comparisons of the trunk level marginal means revealed sacrum-to-lower lumbar (10.81° ± 2.47°) was significantly larger than both lower lumbar-to-upper lumbar (6.23° ± 1.69°) and upper lumbar-to-lower thorax (2.91° ± 0.24°, as in Fig. 5(B)).

Task (walking or stair descent) only changed trunk flexion ROM. There was a significant main effect of task on trunk flexion ROM (p = 006). When compared to walking (10.58° ± 3:05°) the stair descent task was significantly smaller (6.71° ± 1.14°).

Some significant differences between specific tasks, spine level and age group were present, however a noticeable difference in the flexion angle — in both tasks — is noticeably higher in the 50 year old cohort compared to the younger (Fig. 6).