Figure 1.
(A) Configuration of stopped escalator and gait pattern shown by gray ellipses (the case of gait initiation with right leg). The number beside the ellipse indicates the order of gait steps in one trial. The physical dimensions of four steps (i.e. the shorter first step) and approach was duplicated by wooden stairs and plastic material. The stairs have no hand rail. (B) Back and lateral views of participants. Markers were placed C7 of the spine, basis ossis sacri (BOS), and right and left heels (gray points). As indices of postural sway, we used tilting angle (TA) defined as the angle made by the line C7 connecting with the BOS and vertical axis and tilting angle velocity (TAV) obtained by differentiating the TA. (C) The experimental sessions. In session A, five consecutive moving escalator trials (ME1–ME5) were followed by three consecutive trials of stepping onto stopped escalator (SE1–SE3) as one block. In session B, participants performed five moving escalator trials, two wooden stairs trials (WS1, WS2), and finally a stopped escalator trial (SE3-B) sequentially as one block. A total of 16 blocks were done for each session.
Figure 2.
Significant differences between SE1 and WS1 conditions and between SE2 and WS2 conditions suggest participants surely felt the odd sensation when they stepped onto a stopped escalator. Mean scores in WS1 and WS2 conditions were nearly 1, suggesting the structural step nonuniformity (the shorter first one) was not essential for the perception of the odd sensation. Mean score in the SE3 condition was significantly lower than in SE1, suggesting that the more we experience stepping onto a stopped escalator, the less strongly we feel the odd sensation. There was also a significant difference between SE3 and SE3-B conditions. These observations suggest that the visual context (escalator vs. wooden stairs) in which the action is taken, rather than the irregular step-height itself, is essential for the perception of the odd sensation. Asterisks represent p<0.05.
Figure 3.
Temporal variations of walking velocity (BOS), TA and TAV aligned by the time of each heel strike.
Mean 0.2 s time window horizontal velocity of basis ossis sacri (BOS) (as walking velocity) in three conditions (blue, red, green lines indicate ME5, SE1, WS1 conditions, respectively) were aligned by the heel strike of the (A) second, (B) third, and (C) fourth steps. Time 0 indicates the time of the heel strike of each step, each time (abscissa axis) ±0.1 s represents a 0.2 s time window. Note that the 0.6 in Fig. 3A and the 0 in Fig. 3B and the 0.5 in Fig. 3B and the 0 in Fig. 3C were temporally overlapped. Blue dotted lines in Figs. 3B and C indicate the actual BOS velocities in the ME5 condition after stepping which were calculated by subtracting escalator speed itself (0.5 m/s) from measured BOS velocity. Tilting angle (TA) and tilting angle velocity (TAV) temporally averaged in each 0.2 s time window, aligned by the heel strike of the (D) third and (E) fourth steps. As in Fig. 3B and C, time 0 indicates the time of the heel strike of each step, and 0.5 in Fig. 3D and 0 in Fig. 3E were temporally-overlapped. Action sequences are shown by stick figures above each figure. 2nd HS, 3rd HS, 4th HS, and 5th HS in Figs. 3A–3E indicate the approximate time of the heel strikes of the second, third, fourth, and fifth steps in each figure.
Figure 4.
Behavioral properties of lower limbs in moving-escalator, stopped-escalator and wooden-stairs conditions.
(A) The representative temporal profiles of heel height, heel vertical velocity aligned by the maximal heel velocity of the fourth step. The magenta symbols (inverted triangle, circle, and square) above the stick figure indicate the time point shown by the same symbol in the temporal profile figure. Each stick figure shows the action sequences to the step elevation in the ME5 condition with those without step elevation in SE1 condition (left) and to the heel's downward approach to the strike (right). (B) Schematic profile of heel vertical velocity in stopped escalator (SE) and wooden stairs (WS) before the heel strike of the fourth step. The time window [0.3–0.5] corresponds to the gray area in Fig. 4A. The temporal profile in stopped-escalator situation showed double decelerations, which may reflect corrective movement before the heel strike (the area ASE indicated by the hatched lines diagonally right downward), while that in wooden-stairs condition showed a single deceleration (the area AWS indicated by the hatched lines diagonally right upward). On the basis of the data aligned by the maximal heel velocity of the fourth step (see black dotted vertical line in Fig. 4A), we calculated area ASE as heel's downward approach to strike (HDAS) for the stopped escalator and area AWS as that for the wooden stairs. (C) The HDAS index in the SE1 and WS1 conditions (see also the stick figure in Fig. 4A). The HDAS in the SE1 condition showed significantly larger than that in the WS1 condition. Asterisks represent p<0.05.
Figure 5.
The causal path model and behavioral changes in different conditions.
The adopted behavioral properties are as follows: Mean tilting angle velocity (TAV) at 0.1 aligned by the heel strike of the third step (TAV3), The heel's downward approach to strike (HDAS), mean TAV at 0.1 aligned by the heel strike of the fourth step (TAV4), mean horizontal velocity of the basis ossis sacri (BOS) at 0.1 aligned by the heel strike of the fourth step (BOS4). Stick figures show the action sequence and the adopted motor actions schematically. Bar graphs show the mean values of each behavioral index in each condition. Blue lines indicate paths from the behavioral index to the odd sensation score, and red lines indicate paths between behavioral indices, and the width of the path indicates the strength of relationship (the number of participants who showed significant path coefficient, see also Tables 1 and 2). OS stands for the odd sensation score, and error 1–4 is the error term. The inappropriate lower limb movements (HDAS) did not directly induce the perception of the odd sensation (except for one participant) but upper body movements (mainly TAV4) induced it, although there is a kinematics chain between lower limb and upper body movements (from HDAS to BOS4, and from HDAS to TAV4).
Table 1.
Standardized path coefficients in the SE condition from each kinematics variable to the odd sensation in the model in Fig. 5.
Table 2.
Standardized path coefficients in the SE condition between each kinematics variable in the model in Fig. 5.