Effects of emphasized plantarflexion stop and heel lift height on the plantarflexion resistive torque in healthy young adults wearing ankle foot orthoses Authors: Mikio Fujimoto, Shota Suzuki, Minori Masuda, Hirotaka Kuki Published: 2026 ABSTRACT [Purpose] This study aimed to investigate the effects of plantarflexion stop angle and heel-lift height on peak plantarflexion resistive torque (PFRT) of ankle-foot orthoses (AFOs), including the first and second peaks observed during the loading response and pre-swing phases, respectively. [Participants and Methods] Thirty healthy adults wore AFOs with stiff plantarflexion stops (20 Nm/°) and walked at a comfortable speed. The plantarflexion stop angle was set under three conditions: 10° plantarflexion, 0°, and 10° dorsiflexion. Heel-lift height was tested under four conditions: no heel lift, 1 cm, 2 cm, and 3 cm. Depending on the results of Levene’s test, two-way analysis of variance or Welch’s analysis of variance was performed using weight-corrected first or second peak PFRT values as dependent variables and plantarflexion stop angle and heel-lift height as independent variables. [Results] Increasing the plantarflexion stop angle toward dorsiflexion and increasing heel-lift height resulted in higher first peak PFRT values (range: 0.08–0.69 Nm/kg across conditions). The plantarflexion stop angle also affected the second peak PFRT (range: 0.07–0.44 Nm/kg), whereas the effect of heel-lift height was limited. [Conclusion] This study demonstrates that plantarflexion stop angle and heel-lift height influence peak PFRT values in AFOs. INTRODUCTION Foot clearance of the paretic limb is often impaired in hemiplegic gait, leading to increased reliance on compensatory movements1, 2). In clinical practice, ankle-foot orthoses (AFOs) are commonly prescribed for patients with hemiplegia to prevent equinus during stance and foot drop during the swing phase. In addition, AFOs improve foot clearance3, 4) and help prevent genu recurvatum by promoting forward tibial progression at the knee5). Studies have shown that the dorsiflexion angle and stiffness5,6,7,8) provided by an AFO influence gait parameters, including ankle and knee joint angles and movements during walking. Adjusting the strength of the plantarflexion resistive torque (PFRT) affects muscle activity around the ankle and can maximize the functional effect of the AFO9, 10). In contrast, Kobayashi et al.11) showed that excessive PFRT may hinder the generation of peak ankle power. AFOs may be manufactured to accommodate equinus and used in combination with heel lifts12). The combined use of an AFO and a heel lift in patients with post-stroke hemiplegia has been reported to improve knee moments and reduce genu recurvatum during the early stance phase13, 14). AFOs with varying levels of plantarflexion stiffness and different heel-lift heights are prescribed according to the characteristics and severity of gait disturbance, as well as the anticipated conditions of use for each individual. More severe equinus and/or genu recurvatum typically require greater plantarflexion resistance and/or a higher heel lift. For patients with severe hemiplegia, Klenzak joints are often prescribed, as they allow high plantarflexion resistance and provide an adjustable plantarflexion stop angle. Direct measurement of PFRT has been reported previously15,16,17,18). A commercial device allows PFRT to be measured conveniently17). In healthy individuals, PFRT exhibits two distinct peaks: the first peak occurs during the loading response phase, and the second peak occurs during the pre-swing phase17, 18). The first peak represents resistance to the ankle plantarflexion moment generated by ground reaction forces acting on the heel17, 18). Consequently, this first peak reduces braking forces at initial contact and facilitates progression of the first ankle rocker19). The second peak reflects resistance to the rapid plantarflexion that occurs during the pre-swing phase17, 18), a motion often described as a catapult-like mechanism without escapement20). This rapid plantarflexion results from a stretch–shortening cycle involving the triceps surae muscle and the Achilles tendon21). Because equinus limits ankle dorsiflexion before the pre-swing phase, the second peak often disappears in patients with stroke18). Peak PFRT serves as a quantitative index for evaluating equinus and/or genu recurvatum in patients with severe hemiplegia who require a stiff plantarflexion stop, such as those using Klenzak joints. The aim of this basic study was to clarify how plantarflexion stop angles, heel-lift heights, and their combined use influence peak PFRT in AFOs with stiff plantarflexion stops to obtain basic knowledge backed by quantitative indicators that can be applied to prescribing orthotics for patients with severe hemiplegia. PARTICIPANTS AND METHODS This study was conducted with approval from the institutional ethics committee, and all participants provided written informed consent before enrollment. Thirty healthy adults without gait abnormalities were included. Individuals with a history of allergy to leather or metal were excluded because of the potential for allergic reactions to the AFO materials. Because physique and walking characteristics differ by sex22, 23), participants were recruited in equal numbers of women and men to minimize sex-related bias. Three right-sided AFOs corresponding to shoe lengths of 23.5 cm, 25.5 cm, and 27 cm were selected for examination. Each AFO had bilateral metal uprights, bilateral double Klenzak ankle joints, and a plantarflexion stop consisting of a stiff metal rod without a spring on the lateral ankle joint. The AFOs were 2E width and featured flat, non-flared soles with a thickness of 1 cm, along with a 1 cm toe spring originating at the metatarsophalangeal joint17). A left shoe of the same size as each AFO was prepared to form a matched pair. Heel lifts were designed to attach to the AFO. For each AFO size, heel lifts were available in three heights: 1 cm, 2 cm, and 3 cm. The height of the contralateral shoe was raised horizontally by the same amount as the heel lift attached to the paired AFO. The AFO did not restrict dorsiflexion but limited plantarflexion using a rod equipped with a load cell to measure PFRT at the lateral ankle joint. The Gait Judge System (GJS; Pacific Supply Co., Ltd., Daito, Osaka, Japan) was used to automatically record PFRT, ankle joint angles, and walking videos (Fig. 1Fig. 1. Equipment used to measure plantarflexion resistive torque (PFRT).The Gait Judge System automatically records synchronized (A) walking videos, (B) PFRT, and (C) ankle joint angle. The system components include a (D) transmitter, (E) plantarflexion stop with an attached load cell, and (F) digital goniometer, all mounted on the lateral ankle joint of the AFO. (G) The heel lift is designed to be rigidly attached to the sole of the AFO. (H) PFRT and ankle joint angle data are transmitted to the transmitter via a lead wire. (I) The plantarflexion stop is rigidly fixed using a nut. (J) The plantarflexion stop angle can be adjusted by changing the depth of the rod. (K) A load cell is mounted on the plantarflexion stop rod. (L) A rigid metal bar, rather than a spring, was used as the plantarflexion stop in this study. (M) The AFO did not restrict dorsiflexion. AFO: ankle-foot orthosis.). Equipment used to measure plantarflexion resistive torque (PFRT). The Gait Judge System automatically records synchronized (A) walking videos, (B) PFRT, and (C) ankle joint angle. The system components include a (D) transmitter, (E) plantarflexion stop with an attached load cell, and (F) digital goniometer, all mounted on the lateral ankle joint of the AFO. (G) The heel lift is designed to be rigidly attached to the sole of the AFO. (H) PFRT and ankle joint angle data are transmitted to the transmitter via a lead wire. (I) The plantarflexion stop is rigidly fixed using a nut. (J) The plantarflexion stop angle can be adjusted by changing the depth of the rod. (K) A load cell is mounted on the plantarflexion stop rod. (L) A rigid metal bar, rather than a spring, was used as the plantarflexion stop in this study. (M) The AFO did not restrict dorsiflexion. AFO: ankle-foot orthosis. Preliminary experiments were conducted to clarify the mechanical properties of each AFO size used in this study. Participants wore an AFO set to a plantarflexion limit of 0° without a heel lift and walked 6 m after standing on tiptoes three times. During these trials, PFRT and ankle joint angles were measured using the GJS. For each AFO, scatter plots were generated with PFRT on the vertical axis and ankle joint angle on the horizontal axis. Regression equations describing the relationship between PFRT and ankle joint angle were calculated over the range in which PFRT exceeded baseline values. A series of measurements was obtained from the right lower limb. Participants mounted the GJS on the right AFO under various AFO settings, as described below, wore a left shoe of matching height, and walked along a 6 m flat walkway at a comfortable self-selected speed. The AFO plantarflexion stop angle was tested under three conditions: 10° dorsiflexion, 0°, and 10° plantarflexion. Heel-lift height was tested under four conditions: no heel lift and heel lifts of 1 cm, 2 cm, and 3 cm, resulting in a total of 12 experimental conditions. PFRT data (Fig. 2Fig. 2. Measurement of plantarflexion resistive torque and ankle joint angle using the Gait Judge System.The upper trace shows plantarflexion resistive torque (Nm), and the lower trace shows ankle joint angle (degrees, with negative values indicating plantarflexion). The upward arrow indicates the start of walking, and the horizontal arrows indicate the start and end of each gait cycle. Multiple peaks in plantarflexion resistive torque (PFRT) correspond to sequential gait cycles, with first and second peaks visible within each cycle.) were recorded by the GJS at a sampling frequency of 1,000 Hz. Peak PFRT values were extracted from the GJS output tables, and the mean values of the first and second PFRT peaks from the third to fifth gait cycles were calculated. Peak PFRT values were normalized to body weight. In this study, plantarflexion angles of the ankle were defined as negative. Measurement of plantarflexion resistive torque and ankle joint angle using the Gait Judge System. The upper trace shows plantarflexion resistive torque (Nm), and the lower trace shows ankle joint angle (degrees, with negative values indicating plantarflexion). The upward arrow indicates the start of walking, and the horizontal arrows indicate the start and end of each gait cycle. Multiple peaks in plantarflexion resistive torque (PFRT) correspond to sequential gait cycles, with first and second peaks visible within each cycle. Lines were marked on the floor at 1 m intervals from the start line, and a tape measure with red markings at 10 cm intervals, a primary scale at 1 cm intervals, and an auxiliary scale at 1 mm intervals were placed along the walking path. Using these markings, cadence, stride length, and walking speed were measured and calculated simultaneously during the experiments. All statistical analyses were performed using SPSS version 25.0 (IBM Corp., Armonk, NY, USA), with the level of statistical significance set at p<0.05. Data normality was confirmed using the Kolmogorov–Smirnov test, and homogeneity of variance was assessed using Levene’s test. Two-way repeated-measures analysis of variance (ANOVA) was conducted when assumptions were satisfied. If sphericity was violated, as determined by Mauchly’s test, degrees of freedom were adjusted using the Huynh–Feldt correction when ε>0.75 or the Greenhouse–Geisser correction when ε<0.75. When a significant effect was identified by two-way repeated-measures ANOVA, linear contrast analyses were performed to reconfirm statistical significance. If the linear contrast was significant, simple main effects were further examined with Bonferroni-adjusted post hoc comparisons. When homoscedasticity was not satisfied, analyses were limited to one-way comparisons using Welch’s ANOVA, followed by the Games–Howell post hoc test. To confirm that cadence, height-corrected stride length, and height-corrected walking speed were not influenced by plantarflexion stop angle, heel-lift height, or their interaction, two-way repeated-measures ANOVA was performed prior to the main analyses. Two-way repeated-measures ANOVA, followed by linear contrast and post hoc analyses of simple main effects, was conducted using weight-corrected first peak values ​​as the dependent variables and plantarflexion stop angle and heel-lift height as the independent variables. Welch’s ANOVA followed by the Games–Howell post hoc test was performed using weight-corrected second peak values ​​as the dependent variables and plantarflexion stop angle as the independent variable at each heel-lift height. Similarly, Welch’s ANOVA was conducted with heel-lift height as the independent variable at each plantarflexion stop angle. RESULTS The study included 15 male and 15 female participants. Mean age, height, and weight were 27.6 ± 4.0 years, 166.2 ± 7.0 cm, and 57.1 ± 8.2 kg, respectively. There were no significant differences in age between sexes (p>0.05), though males were significantly taller and heavier than females (males: 172.1 ± 4.2 cm, 63.4 ± 6.8 kg; females: 160.3 ± 4.1 cm, 50.8 ± 5.2 kg, p<0.001). The relationships between PFRT and ankle joint angle for each AFO are shown in Fig. 3Fig. 3. Relationship between plantarflexion resistive torque and ankle joint angle.Ankle-foot orthosis sizes are (A) 23.5 cm, (B) 25.5 cm, and (C) 27 cm.. The regression equations describing PFRT as a function of ankle joint angle for AFO sizes of 23.5 cm, 25.5 cm, and 27 cm were y= −19.1x + 2.74, y= −20.9x + 0.47, and y= −20.3x −0.86, respectively. Based on these equations, plantarflexion stiffness values in the preliminary experiments were 19.1 Nm/°, 20.9 Nm/°, and 20.3 Nm/°, respectively. Relationship between plantarflexion resistive torque and ankle joint angle. Ankle-foot orthosis sizes are (A) 23.5 cm, (B) 25.5 cm, and (C) 27 cm. No abnormal gait such as genu recurvatum was observed in any of the measurements. Mean cadence, height-corrected stride length, and height-corrected walking speed were 69.5 ± 6.1 steps/min, 0.87 ± 0.07, and 0.50 ± 0.07 m/s, respectively. Two-way repeated-measures ANOVA showed no significant main effects of plantarflexion stop angle, heel-lift height, or their interaction on height-corrected stride length or height-corrected walking speed. A significant main effect of plantarflexion stop angle and heel-lift height on cadence was observed (partial η2=0.086, p=0.020); however, this effect was not confirmed by linear contrast analysis (partial η2=0.000, p=0.937). Overall, cadence, height-corrected stride length, and height-corrected walking speed were not affected by plantarflexion stop angle, heel-lift height, or their interaction. Weight-corrected first peak values were recorded under all AFO settings (Table 1Table 1. Weight-corrected first peak plantarflexion resistive torque under all ankle-foot orthosis settings (mean ± SD, Nm/kg)Plantarflexion stop at 10° plantarflexionPlantarflexion stop at 0°Plantarflexion stop at 10° dorsiflexionHeel lift 0 cm0.08 ± 0.080.25 ± 0.130.47 ± 0.13Heel lift 1 cm0.10 ± 0.090.30 ± 0.120.53 ± 0.13Heel lift 2 cm0.15 ± 0.110.38 ± 0.130.61 ± 0.13Heel lift 3 cm0.21 ± 0.120.45 ± 0.130.69 ± 0.13). Table 2Table 2. Effects of plantarflexion stop angle, heel-lift height, and their interaction on the first peak plantarflexion resistive torqueTwo-way repeated-measures analysis of variancePost hoc linear contrastPlantarflexion stop angle0.94 (<0.01)0.97 (<0.01)Heel-lift height0.85 (<0.01)0.87 (<0.01)Interaction (synergetic effect)0.24 (<0.01)0.45 (<0.01)Data are presented as partial η2 values (p-values). Partial η2 values represent effect sizes, where 0.01=small effect, 0.06=medium effect, and 0.14=large effect. All observed effects in this study were large (partial η2>0.14), indicating substantial influences of plantarflexion stop angle and heel-lift height on first peak PFRT. summarizes the partial η2 values and two-sided significance levels for the effects of plantarflexion stop angle, heel-lift height, and their interaction on the first peak, as determined by two-way repeated-measures ANOVA and linear contrast analyses. All simple main effects on the first peak, adjusted using the Bonferroni correction, were significant across all combinations and conditions (p=0.000 for all). For the second peak analysis, Levene’s test indicated heterogeneity of variance; therefore, Welch’s ANOVA followed by Games–Howell post hoc tests were employed instead of two-way repeated-measures ANOVA. Table 3Table 3. Weight-corrected second peak plantarflexion resistive torque under all ankle-foot orthosis settings (mean ± SD, Nm/kg)Plantarflexion stop at 10° plantarflexionPlantarflexion stop at 0°*Plantarflexion stop at 10° dorsiflexion*Heel lift 0 cm*0.07 N,D0.16 P,D,40.34 P,N,4Heel lift 1 cm*0.08 N,D0.17 P,D0.38 P,NHeel lift 2 cm*0.09 N,D0.19 P,D0.40 P,NHeel lift 3 cm*0.11 N,D0.22 P,D,10.44 P,N,1* Asterisks in the table column and row labels indicate significant main effects identified by Welch’s analysis of variance. Superscript letters indicate significant difference in post hoc Games–Howell tests when compared with: P, plantarflexion stop at 10° plantarflexion; N, plantarflexion stop at 0°; and D, plantarflexion stop at 10° dorsiflexion. Numeric superscripts indicate significant differences in post hoc Games–Howell tests when compared with heel-lift heights of 0 cm, 1 cm, 2 cm, 3 cm respectively. presents weight-corrected second peak values under all AFO settings, together with the results of the statistical analyses. The plantarflexion stop angle significantly influenced the second peak, whereas the effect of heel-lift height was limited. Table 1. Weight-corrected first peak plantarflexion resistive torque under all ankle-foot orthosis settings (mean ± SD, Nm/kg)Plantarflexion stop at 10° plantarflexionPlantarflexion stop at 0°Plantarflexion stop at 10° dorsiflexionHeel lift 0 cm0.08 ± 0.080.25 ± 0.130.47 ± 0.13Heel lift 1 cm0.10 ± 0.090.30 ± 0.120.53 ± 0.13Heel lift 2 cm0.15 ± 0.110.38 ± 0.130.61 ± 0.13Heel lift 3 cm0.21 ± 0.120.45 ± 0.130.69 ± 0.13 Weight-corrected first peak plantarflexion resistive torque under all ankle-foot orthosis settings (mean ± SD, Nm/kg) Table 2. Effects of plantarflexion stop angle, heel-lift height, and their interaction on the first peak plantarflexion resistive torqueTwo-way repeated-measures analysis of variancePost hoc linear contrastPlantarflexion stop angle0.94 (<0.01)0.97 (<0.01)Heel-lift height0.85 (<0.01)0.87 (<0.01)Interaction (synergetic effect)0.24 (<0.01)0.45 (<0.01)Data are presented as partial η2 values (p-values). Partial η2 values represent effect sizes, where 0.01=small effect, 0.06=medium effect, and 0.14=large effect. All observed effects in this study were large (partial η2>0.14), indicating substantial influences of plantarflexion stop angle and heel-lift height on first peak PFRT. Effects of plantarflexion stop angle, heel-lift height, and their interaction on the first peak plantarflexion resistive torque Data are presented as partial η2 values (p-values). Partial η2 values represent effect sizes, where 0.01=small effect, 0.06=medium effect, and 0.14=large effect. All observed effects in this study were large (partial η2>0.14), indicating substantial influences of plantarflexion stop angle and heel-lift height on first peak PFRT. Table 3. Weight-corrected second peak plantarflexion resistive torque under all ankle-foot orthosis settings (mean ± SD, Nm/kg)Plantarflexion stop at 10° plantarflexionPlantarflexion stop at 0°*Plantarflexion stop at 10° dorsiflexion*Heel lift 0 cm*0.07 N,D0.16 P,D,40.34 P,N,4Heel lift 1 cm*0.08 N,D0.17 P,D0.38 P,NHeel lift 2 cm*0.09 N,D0.19 P,D0.40 P,NHeel lift 3 cm*0.11 N,D0.22 P,D,10.44 P,N,1* Asterisks in the table column and row labels indicate significant main effects identified by Welch’s analysis of variance. Superscript letters indicate significant difference in post hoc Games–Howell tests when compared with: P, plantarflexion stop at 10° plantarflexion; N, plantarflexion stop at 0°; and D, plantarflexion stop at 10° dorsiflexion. Numeric superscripts indicate significant differences in post hoc Games–Howell tests when compared with heel-lift heights of 0 cm, 1 cm, 2 cm, 3 cm respectively. Weight-corrected second peak plantarflexion resistive torque under all ankle-foot orthosis settings (mean ± SD, Nm/kg) * Asterisks in the table column and row labels indicate significant main effects identified by Welch’s analysis of variance. Superscript letters indicate significant difference in post hoc Games–Howell tests when compared with: P, plantarflexion stop at 10° plantarflexion; N, plantarflexion stop at 0°; and D, plantarflexion stop at 10° dorsiflexion. Numeric superscripts indicate significant differences in post hoc Games–Howell tests when compared with heel-lift heights of 0 cm, 1 cm, 2 cm, 3 cm respectively. DISCUSSION Increasing the plantarflexion stop angle toward dorsiflexion and increasing heel-lift height were associated with greater increases in the first peak during the loading response phase. A synergistic effect of plantarflexion stop angle and heel-lift height on the first peak was also observed. In contrast, the second peak was primarily influenced by plantarflexion stop angle, whereas the effect of heel-lift height was limited. The plantarflexion stop stiffness used in this study was approximately 20 Nm/°, which is substantially greater than that reported in previous studies of AFOs with plantarflexion resistance5,6,7,8,9, 11, 15, 16, 18, 24, 25), such as those employing springs or oil dampers. Minor differences were considered to be attributable to variation in the tightening force applied by the therapist or technician to secure the plantarflexion stop bolt within the screw hole. This level of stiffness was selected to enable measurement of PFRT in patients with severe spastic hemiplegia who cannot adequately control genu recurvatum or achieve heel contact without a very stiff plantarflexion stop or a high heel lift. While data were collected from healthy adults to establish baseline mechanical relationships, these findings provide fundamental insight into the biomechanical demands that AFO materials and structures must withstand when prescribed for patients with severe hemiplegia. The primary aim of this study was to clarify the effects of plantarflexion stop angle and heel-lift height on the first peak. These findings have direct implications for the clinical prescription of AFOs, suggesting that careful consideration of material strength and structural design is required to achieve appropriate plantarflexion control. In addition to considering patient physique, AFO settings must be tailored to optimize function. Furthermore, reinforcement of the AFO ankle joint should be considered when increasing the plantarflexion stop angle or adding a heel lift to prevent mechanical failure. No effects of plantarflexion stop angle or heel-lift height on cadence, stride length, or walking speed were observed in this study. This finding does not necessarily indicate that these AFO settings have no influence on gait parameters. Because participants were aware that the study focused on the effects of plantarflexion stop angle and heel-lift height on PFRT, they may have consciously or unconsciously maintained consistent gait patterns, thereby indirectly influencing the results. Further studies are needed to clarify how plantarflexion stop angle and heel-lift height affect gait parameters and how any indirect changes in gait subsequently influence PFRT. The main effects of plantarflexion stop angle and heel-lift height on the first peak can be explained by the positional relationship between the ankle joint axis and the ground reaction force during the heel rocker, from initial contact to the foot-flat position. During this phase, the ground reaction force is applied from the heel and passes posterior to the ankle joint axis, generating a plantarflexion moment. PFRT arises when this plantarflexion moment is resisted by the AFO. After the foot-flat position, PFRT decreases and disappears as the heel rocker transitions to the ankle rocker. When the plantarflexion stop is set closer to dorsiflexion and heel-lift height is increased, the angle between the sole and the floor at initial contact increases. This configuration prolongs the interval from initial contact to foot-flat, thereby enhancing PFRT. A similar mechanism was reported by Kobayashi et al.24), who demonstrated that higher plantarflexion resistance prolongs the time from initial contact to foot-flat, resulting in increased PFRT. The effect of plantarflexion stop angle on the second peak may be explained by earlier engagement of the plantarflexion stop while the triceps surae muscle–tendon unit remains stretched. In contrast, the effect of heel-lift height on the second peak was limited, and the data from the present study alone are insufficient to fully elucidate the underlying mechanism. One possible explanation is that heel lifts primarily influence the geometric relationship between the foot and ground during the early stance phase, whereas the second peak occurs during pre-swing when foot-ground contact is minimal and the rapid plantarflexion motion is governed more by plantarflexion stop engagement than by heel position. Further investigation using electromyography and motion capture analysis is warranted to clarify this mechanism. This study has several limitations. First, participants were recruited based on the absence of observable gait abnormalities as assessed by a physician and physical therapists, without the use of objective gait measurements. Consequently, the presence of subtle gait abnormalities may not have been detected during visual assessment. Second, the position of the AFO ankle joint could not be individually adjusted for each participant. However, this limitation was mitigated by the use of three AFO sizes, which likely reduced the degree of mismatch. Previous work by Gao et al.25) reported that deviation between the AFO ankle joint center and the anatomical axis of rotation affects plantarflexion stiffness. Third, the experiments were conducted in healthy young adults. Therefore, orthopedic factors commonly observed in patients with hemiplegia, such as ankle or knee contractures and increased muscle tone, were not reflected in the measured PFRT. Further studies are required to determine how these clinical factors influence PFRT. Fourth, during normal gait, the tibialis anterior muscle contracts eccentrically during the heel rocker phase to resist rapid plantarflexion, which is expected to reduce the first peak. The activity of the tibialis anterior muscle was not assessed in the present study. In patients with hemiplegia who have impaired tibialis anterior function, the first peak PFRT values may be higher than those observed in our healthy participants, as the AFO would need to compensate for the absent or reduced eccentric muscle control. Future studies should incorporate EMG measurements to quantify this relationship. Finally, in this study, PFRT was measured by mounting the GJS on an AFO with a plantarflexion stop. In clinical practice, however, AFOs are not limited to rigid plantarflexion stops; some designs permit controlled plantarflexion while providing resistive torque. When such AFOs are used, PFRT may differ from that observed in the present study. In the daily life of patients with hemiplegia, PFRT at the AFO ankle joint is influenced by various factors, including environmental conditions, surface unevenness and slope, walking speed, and the need to respond to surrounding demands. Therefore, when prescribing and fitting an AFO in clinical settings, a broad range of factors, including patient lifestyle and environment, should be considered. In this study, among the many conditions that impose mechanical stress on an AFO during daily use, only brace settings were examined. Nevertheless, the findings provide fundamental insight into the functional settings and mechanical strength required for AFO ankle joints. Further research is warranted to clarify how AFO biomechanics change under different conditions of use. Funding This study was partially supported by the Research Foundation for Rehabilitation Medicine, Koshigaya, Japan. The funding body had no role in the study design, conduct, data analysis, interpretation of results, or manuscript preparation and publication. Conflict of interest The authors declare no conflicts of interest.