*2.2. Shoe Conditions*

All participants conducted this study in shoes with a NHS, normal shoes (NS), and PHS (Figure 1b). The NS were commercially available walking shoes. The NHS and the PHS were self-fabricated based on the NS in our laboratory. For the three conditions, the shoes were identical models and designs in the upper and outsole.

**Figure 1.** (**a**) Marking point paste location. (**b**) Experimental process, IDEEA position, and shoe conditions between NHS (negative 1.5 cm drop), NS (no drop), and PHS (1.5 cm drop). (**c**) Diagrammatic illustration of COM-COP inclination angles.

#### *2.3. Testing Procedure*

All participants walked with IDEEA (IDEEA, MiniSun, Fresno, CA, USA) on a 6.5 m walkway at their self-selected comfortable speed to present normal gait characters, striking their right foot on the force plate. Sensors were connected to a 32 Hz main recorder.

Previous studies have shown the reliability of IDEEA in measuring gait parameters [27–29]. Each footwear condition was collected with three successful trials for analysis. At the same time, an eight-camera Vicon motion capture system (Vicon Metrics Ltd., Oxford, UK) was used to capture the motion trajectory. The embedded AMTI force plates (AMTI, Watertown, MA, USA) recorded the GRF synchronously, with 200 Hz and 1000 Hz, respectively, as shown in Figure 1b. The camera system was calibrated to residual errors of 2.5 mm over a recording volume of approximately 6.5 m × 1.5 m × 1.80 m (L × W × H). The force plate was embedded in the middle of a 6.5-m walkway and covered with floor tiles to minimize participants' awareness of its presence. The original gait-2392 model in OpenSim was used for this study, with 23 degrees of freedom and 92 muscles (Figure 1a) [30].

#### *2.4. Data Processing*

Gait analyses were performed using a wearable intelligent analyzer (IDEEA, MiniSun, Fresno, CA, USA) equipped with accelerometers and gyroscopes, as shown in Figure 1b. The wearable intelligent analyzer consists of the main recorder and two secondary recorders. The gait data were collected and transmitted to the main recorder by the sensor affixed to the subject; each accelerometer used a proprietary algorithm [31]. The IDEEA was easy to wear and had almost no interference with normal walking. After the data acquisition was completed, the data were saved in the main recorder and downloaded to the computer. IDEEA Version 3.01 (IDEEA3, MiniSun, Fresno, CA, USA) was used for analysis [27]. The software equipped with the equipment can intercept the range of gait data needed and process it, and directly output walking speed, step frequency, stride length, and support time.

According to Winter's [32] description of the selected frequency for filtering biomechanical signals, the residual data analysis was carried out in subsets to determine the most appropriate signal-to-noise ratio. Marker trajectories and ground reaction forces were filtered by a zero-delay fourth-order Butterworth low-pass filter at 12 Hz and 30 Hz. A threshold of 20 N on the vertical GRF was applied to identify the initial foot contact and toe-off [33]. The magnitudes of each GRF component were normalized to the percentage of the participant's body weight, and the stance phase of each participant was normalized to 100% of their stance phase's duration [34]. The musculoskeletal model used was the generic OpenSim model Gait 2392 (Figure 1a), which has 23 degrees of freedom and 92 muscles [30] and calculates the center of mass (COM) in OpenSim.

#### *2.5. Outcome Measures*

The parameters evaluated in the study were: (1) Walking speed (m/s): the distance walked along the walkway per second. (2) Step frequency (steps/min): the number of steps per minute. (3) Stride length (m): the distance from one heel to the same heel touching the ground again during walking. (4) Double support time/single support time (%): double support time refers to the time taken by the use of biped support in a gait cycle, and single support time refers to the time spent using single foot support in a gait cycle. Double support time/single support time reflects the stability of the participants when walking, where the lower the ratio, the better the stability of the participants [27]. (5) Three-dimensional ground reaction forces (3D-GRF): GRF supports the body against gravity and accelerates the center of mass during walking. GRF is included in the vertical, anterior–posterior, and medial–lateral directions recorded from a three-dimensional force plate [35,36]. (6) The range of COP motion, including the medial–lateral range of the COP (ML-COP) and anterior–posterior range of the COP (AP-COP), were derived and averaged for all participants. (7) Center of mass (COM) and center of pressure (COP) inclination angles: we defined COM-COP inclination angles as the angle formed by the intersection of the line connecting the COP and COM with a vertical line through the COP [37], as shown in Figure 1c.

#### *2.6. Data Analysis*

Statistical analyses were performed using SPSS 16.0 (SPSS, Chicago, IL, USA) statistical analysis software. One-way repeated-measures analysis of variance (ANOVA) was performed to analyze the effects of different conditions on spatiotemporal parameters and peak COM-COP inclination angles. In the event of a significant main effect, post-hoc pairwise comparisons were conducted on all significant main effects, using a Bonferroni adjustment. Statistical parametric mapping based on the SPM1D package for Matlab (Mathworks, Natick, MA, USA) was used to compare the 3D GRF and COP statistically. In agreement with Patakt et al., SPM was implemented hierarchically, analogous to one-way repeated measures ANOVA (SPM F) with a post-hoc paired t-test [38]. The conditions NS vs. NHS, NS vs. PHS, and PHS vs. NHS were chosen to compare the 3D-GRF and COP waveforms [39,40]. The significance level was set at 0.05.

#### **3. Results**
