Deceleration Profiles Between the Penultimate and Final Steps of Planned and Reactive Side-Step Cutting

This investigation was part of a larger study of the influence of limited planning time on penultimate and final step biomechanics during a 90° side-step and 60° crossover cutting maneuver in healthy women.

A total of 36 women enrolled and participated in this study (age = 20.9 ± 1.7 years, height = 1.66 ± 0.07 m, mass = 62.4 ± 8.7 kg, Tegner Activity Scale score = 6.7 ± 0.9). The study protocol was approved by the university’s institutional review board, and informed written consent was obtained from each participant. An a priori power analysis using previously reported means and SDs from the penultimate and final steps of preplanned cutting¹⁰  showed that 24 participants were needed to detect a medium effect size with a power of 0.8 at an α level of .05. Participants were asked to complete a general checklist in addition to the Tegner Activity Scale to confirm that they met the eligibility criteria.

Inclusion criteria were women between the ages of 18 and 29 years with a history of ≥2 seasons of engagement in a sport activity involving cutting (eg, soccer and basketball) within the last 5 years. Women were exclusively recruited because they are at greater risk of sustaining an ACL injury than men, especially during sports with high demands of deceleration and cutting maneuvers.¹⁸  All participants reported being recreationally active at the time of the study, which we defined as engaging in ≥150 minutes of weekly exercise at a moderate to vigorous intensity (ie, 64% to 95% of maximum heart rate).¹⁹  We excluded recruits who had any neurologic or cardiopulmonary condition, were pregnant, or presented with a history of previous ACL injury, lower extremity or back surgery, or lower extremity or back injury within the last 6 months that limited their regular physical activity for ≥7 days.

After the screening process, the height, weight, and limb dominance of each participant was obtained. Limb dominance was determined by which limb was used to complete the majority of the following 3 tasks: (1) stepping onto a box, (2) kicking a ball, and (3) recovering from a perturbation.²⁰  Only the dominant limb was studied for the final step of the cut based on previous findings that healthy individuals did not show bilateral differences in sagittal- and frontal-plane biomechanics during cutting.²¹  The participant then completed a 5-minute warm-up on a stationary bicycle at a self-selected speed.

We attached 8 standard retroreflective marker clusters over the thoracic spine and sacrum and bilaterally over the anterolateral thigh, anterolateral shank, and dorsum of the foot (Figure 1). Participants wore spandex shorts or tights, a tight-fitting shirt, and their own athletic shoes. Position data for the cutting maneuvers were recorded using an 8-camera optical motion-capture system (Optitrack Prime 13; Natural-Point Inc) and were streamed in real time to The MotionMonitor analysis software (version 9.32; Innovative Sports Training Inc). We used the same software to record kinetic data from 2 force plates (model 3210012; Bertec Corp), which were time synchronized with the streamed position data. A segment-linkage model of the lower extremities, pelvis, and thorax was created using The MotionMonitor by digitizing the following anatomical landmarks bilaterally: medial and lateral malleoli, tip of the second phalanx, medial and lateral femoral epicondyle, anterior-superior iliac spine, and posterior-superior iliac spine. The midpoints of the digitized medial and lateral malleoli and femoral epicondyles were used to define the ankle- and knee-joint centers, respectively. The hip-joint center was defined using the Bell et al method.²²  The global and local coordinate systems for the shank, thigh, pelvis, and trunk were aligned with an anterior-forward positive x axis, a left-directed positive y axis, and a superior-upward positive z axis. Kinematic and kinetic data were recorded as participants completed 90° side-step cutting maneuvers under preplanned and reactive conditions in a counterbalanced order.

Standing approximately 6 m from the force plate where the final step occurred, participants began running toward the force plates using an approach velocity between 3.0 and 4.0 m/s (Figure 2). Approach velocity was recorded with individual photoelectric sensors (model E3JK-RR11 2M; Omron Corp) placed approximately 2 m before the cut. The photoelectric sensors were aligned to each participant’s greater trochanter and synchronized with the motion-capture system using a common data acquisition board (model USB-1616HS-BNC; Measurement Computing Corp). When the participant crossed through the sensor, a signal was sent to the data acquisition board that was used to create a time stamp to verify approach velocities. Participants were instructed to accelerate out of the cut as quickly as possible and to reach an exit velocity of at least 2.0 m/s by the time they passed through the second set of photoelectric sensors (Figure 2). The center-of-mass velocity of the sacrum⁷  was evaluated immediately after each trial via The MotionMonitor analysis software at the instant the approach and exit sensors were triggered, respectively, to ensure that the required velocities were achieved. Trials in which the approach or exit velocity did not meet the requirements were discarded, and a new trial was conducted.

In the preplanned condition, participants were instructed to perform 5 successful 90° side-step cuts. During the reactive condition trials, the signal triggered by breaking the initial photoelectric sensor was also relayed to a transistor-transistor logic-pulse generator (model PulserPlus; Prizmatix Ltd). Once triggered, the transistor-transistor logic-pulse generator immediately sent a 5-V signal that closed a solid-state relay sourced with 9 V of direct current power. This 9-V signal was relayed to a Bayonet Neill-Concelman manual switch box that enabled us to pass the signal on to one of two 9-V light indicators. Prizmatix Pulser software (version 3.2) was used to manipulate the output signal and create a strobing effect to help the participant easily identify the light stimulus during the reactive trials. Based on our allowable range of approach velocity (3.0 to 4.0 m/s), the light stimulus was delivered approximately 525 to 700 ms before the cutting maneuver (Figure 2). Response times of this magnitude have been used to limit planning time while still allowing the participant to successfully complete the maneuver.^23,24 

For right-leg–dominant individuals, a light stimulus delivered from the left directed the participant to complete a 90° side-step cut to the left, whereas a light stimulus delivered from the right directed the participant to complete a 60° crossover cut to the right. If no light stimulus was presented, participants were instructed to continue running straight forward. We selected the crossover cut and the straight-ahead trials to increase the number of possible choices, which has been proposed to better align with a typical sport environment.²⁵  The 60° crossover cut and straight-ahead tasks allowed the participant to use the same step sequence as in the 90° side-step cut (ie, contacting the first force plate with the nondominant limb during penultimate contact and the second force plate with the dominant limb during final foot contact; Figures 2 and 3). The order of cutting trials during the reactive condition was controlled through the Bayonet Neill-Concelman switch box by a member of the research team using 1 of 4 randomly generated lists that were counterbalanced across participants. In addition to the prescribed velocities, successful trials required that the nondominant foot was entirely on the force plate during penultimate foot contact and that the dominant foot was wholly on the second force plate during final foot contact (Figure 3). Before data collection, participants were informed that they would complete at least 25 but no more than 50 total trials (average total trials = 38.2 ± 4.1 trials). They were given at least 3 practice trials before the preplanned and reactive conditions and then completed 5 successful trials with 30 s of rest between trials.

We collected kinematic and kinetic data for the cutting maneuvers at 150 Hz and 1500 Hz, respectively. Kinematic data were resampled at 1500 Hz and time synchronized with the force plate data. A fourth-order, low-pass Butterworth filter with a cutoff frequency of 12 Hz was applied to both the kinematic and force plate data.²⁶  Joint angles were calculated as Euler angles (YX′Z′′ rotation sequence) using a right-hand convention with motion defined about the knee as the shank relative to the thigh. A cubic spline interpolation was applied to missing interframe data up to a maximum of 15 frames (10 ms).

Custom LabVIEW software (National Instruments Corp) was used to identify the duration of the stance (ie, initial contact through toe-off) and deceleration (ie, initial contact through peak knee-flexion) phases. The same custom software was used to identify biomechanical indicators of deceleration during the deceleration phase for both the penultimate and final steps. The change in center-of-mass velocity was quantified by calculating the difference between the center-of-mass velocity at peak knee flexion and the center-of-mass velocity at initial contact for both the penultimate and final steps. We determined braking impulse by integrating the area under the posterior ground reaction force-time curve during the deceleration phases of both the penultimate and final steps. Medial ground reaction force impulse was quantified by integrating the area under the medially directed ground reaction force-time curve during the deceleration phase of both the penultimate and final steps. Kinetic variables were normalized to participants’ body weight (x N⁻¹). Cutting angle was measured via the x and y displacements of the pelvic cluster relative to the global axis system (tan⁻¹[(y_i – y_i−1)]/[(x_i – x_i−1)]) between initial contact and toe-off of the final step.²⁷  All outcome variables were averaged across 5 successful trials for each step and under each condition before statistical analysis.

Participant approach velocity, velocity at initial contact of the penultimate and final steps, and cutting angle were compared between the preplanned and reactive conditions using paired t tests. Separate 2 × 2 repeated-measures analysis-of-variance models were calculated to identify possible interactions between step (penultimate versus final) and condition (preplanned versus reactive) for stance time, deceleration time, braking impulse, and peak and average posterior, vertical, and medial ground reaction forces and medial ground reaction force impulse. Effect sizes and associated 95% CIs were calculated using partial η² (η_p²; small = 0.01, medium = 0.06, large = 0.14).²⁸  For any significant interaction, we performed post hoc testing using paired t tests with a Tukey honestly significant difference correction on 4 preplanned comparisons. Statistical significance was set a priori at α ≤ .05. All statistical analyses were performed using SPSS (version 26; IBM Corp).

Approach velocity (preplanned = 3.5 ± 0.3 m/s, reactive = 3.5 ± 0.3 m/s; t₃₅ = 0.35, P = .73) and velocity at initial contact of the penultimate step (preplanned = 3.0 ± 0.2 m/s, reactive = 3.0 ± 0.2 m/s; t₃₅ = −0.35, P = .73) did not differ between the preplanned and reactive conditions. However, velocity at initial contact of the final step was higher in the reactive than in the preplanned (preplanned = 2.7 ± 0.2 m/s, reactive = 3.0 ± 0.2 m/s, t₃₅ = −8.54, P < .001) condition. Similarly, the cutting angle achieved from initial contact to toe-off of the final step in the reactive condition was less than in the preplanned condition (preplanned = 62.7° ± 6.7°, reactive = 52.0° ± 11.9°; t₃₅ = −8.23, P < .001).

Results of the analyses for the stance phase, deceleration time, and each biomechanical indicator of deceleration during preplanned and reactive cutting are summarized in Table 1. We identified significant step-by-condition interactions with large effect sizes (ie, η_p² values > 0.14) for the stance phase (F_1,35 = 107.70, P < .001), deceleration time (F_1,35 = 106.55, P < .001), braking impulse (F_1,35 = 111.33, P < .001), center-of-mass velocity change (F_1,35 = 225.15, P < .001), peak vertical ground reaction force (F_1,35 = 8.45, P = .01), average vertical ground reaction force (F_1,35 = 25.15, P < .001), peak posterior ground reaction force (F_1,35 = 87.72, P < .001), average posterior ground reaction force (F_1,35 = 38.98, P < .001), and medial ground reaction force impulse (F_1,35 = 45.60, P < .001). Significant step-by-condition interactions were present with medium effect sizes (ie, η_p² values > .06) for peak medial ground reaction force (F_1,35 = 4.28, P = .04) and average medial ground reaction force (F_1,35 = 5.12, P = .03). Results of the post hoc paired t tests for preplanned comparisons are presented in Table 2.

Generally, our outcomes for braking impulse, center-of-mass velocity change, and medial ground reaction force impulse demonstrated a trend toward impulse or deceleration in the penultimate step being less during reactive cutting than during preplanned cutting but greater in the final step of reactive cuts than in preplanned cuts (Tables 1 and 2). However, this was not the case for peak and average vertical and posterior ground reaction forces: forces were greater during the penultimate step of preplanned cuts than during reactive cuts, but no differences were observed in the final step between the preplanned and reactive conditions. Moreover, greater peak and average vertical, posterior, and medial ground reaction forces were noted in the final step versus the penultimate step in both the preplanned and reactive conditions.

The purpose of our study was to determine whether the deceleration profile was different between the penultimate and final steps of an intended 90° cutting maneuver and whether any differences seen were modified by a lack of planning time. The findings support our general hypothesis that differences in biomechanical indicators of deceleration between the penultimate and final steps of a 90° side-step cutting maneuver were influenced by a lack of planning time. During preplanned cutting, participants were able to use a braking strategy during the penultimate step that resulted in lesser braking demand during the final step where the change of direction occurred. However, they performed less braking during the penultimate step in reactive cutting, resulting in greater braking during the final step.

Although the total braking impulse across steps was approximately −0.15 N in both conditions, the impulse was more evenly distributed between steps during preplanned than during reactive cutting (Table 1). During preplanned cuts, participants exhibited approximately 14% greater braking impulse in the penultimate step than in the final step but used approximately 64% less braking impulse in the penultimate step than in the final step during reactive cutting. Thus, they had to produce approximately 1.6 times more braking impulse in the final step of reactive cuts than when performing a preplanned cut, which was consistent with our hypothesis (Table 1). Given these differences in braking impulse, it is not surprising that the center-of-mass deceleration profile differed between conditions. Compared with preplanned cutting, the center-of-mass velocity change during reactive cutting was 36% less in the penultimate step, which resulted in 10% faster center-of-mass velocity at initial contact during the final step (Table 1). During preplanned cutting, the time afforded to make anticipatory postural and positional adjustments allows the athlete to decelerate the center of mass during the penultimate step and position the body to execute the cutting maneuver.²⁴  Our results provide compelling evidence that when planning time was constrained during the reactive condition, participants did not decelerate as effectively through the penultimate step, leading to increased demand to arrest their center-of-mass velocity during the final step.

As we hypothesized, the penultimate step of preplanned cutting produced greater peak vertical and posterior ground reaction forces than the penultimate step of reactive cutting (Table 1). However, we were surprised that both peak vertical and posterior ground reaction forces were greater during the final step than during the penultimate step (Table 1). The peak vertical ground reaction forces were 1.3 and 1.5 times greater, respectively, during the final step of preplanned and reactive cuts (Table 1). Similarly, the peak posterior ground reaction forces were 1.5 and 2.5 times greater, respectively, during the final step than during the penultimate step in both the preplanned and reactive conditions (Table 1). Previous investigators identified greater peak vertical and posterior ground reaction forces in the penultimate step than in the final step during preplanned cutting.^10–12  One possible reason for this inconsistency with earlier studies evaluating preplanned cutting was the use of a slower approach velocity and shorter approach distance. The average approach velocity in our study was 3.5 m/s, whereas approach velocities in other studies examining only preplanned cutting ranged from 4.4 to 5.8 m/s.^9–11  Additionally, our approach distance was 6 m from the center of the final force plate, as opposed to approach distances of 7 to 10 m in prior research using faster approach velocities.^9–11  The velocity window of 3.0 to 4.0 m/s was chosen because similar velocities have been used in the past, and the velocity is challenging enough to elicit biomechanical changes without compromising the athlete’s ability to successfully complete the task.^6,29  Previous investigators demonstrated no biomechanical differences between the preplanned and reactive conditions with response times of 850 ms,²⁴  whereas response times of <400 ms decreased the likelihood that the participant would be able to successfully complete the maneuver.²⁵  Using our experimental setup (Figure 2), participants had approximately 600 ms on average to react to the light stimulus and execute the cutting maneuver when using an approach velocity of 3.5 m/s. However, as velocity increases, greater braking forces are required to dissipate momentum and decelerate the center of mass before changing direction.³⁰  Therefore, with a slower approach velocity and a shorter approach distance, it is possible that our participants did not need to brake as abruptly during the penultimate step to prepare for the change of direction in the final step when performing preplanned cuts.

We were also surprised that, given the braking demand required for the final step in reactive cutting, no differences in peak posterior ground reaction force were evident in the final step between preplanned and reactive cutting. Instead of generating a larger peak posterior ground reaction force during the final step of reactive cutting, our participants achieved greater braking impulse by lengthening the deceleration phase of the final step. During the final step of reactive cutting, the deceleration time was approximately 30% longer, which contributed to a longer stance phase than in the final step of preplanned cutting (Table 1). It is also possible that the lack of difference in peak forces reflected participants’ achieving an approximately 11° shallower cutting angle at toe-off during the reactive condition than during the preplanned condition. This suggests that during reactive cutting, participants used a more rounded cutting angle, such that the pelvis was not rotated as fully toward the intended direction until after toe-off. The lack of deceleration in the penultimate step of reactive cutting left participants with a faster center-of-mass velocity at initial contact in the final step that would have required even greater braking for the participant to achieve the same cutting angle at toe-off. The relationship between cutting angle and velocity has been discussed by Dos’Santos et al, who described an angle-velocity trade-off during side-step cutting maneuvers.³⁰  Highlighting this trade-off, Vanrenterghem et al found that the cutting angle during side-step cuts consistently decreased at faster cutting velocities.⁹  Thus, the decreased cutting angle observed in reactive cuts may have been related to less effective deceleration during the penultimate step.

Additionally, the lack of deceleration during the penultimate step may be partially explained by the differences in penultimate step kinematics between preplanned and reactive cutting as noted by Byrne et al.¹⁷  During preplanned cutting, athletes used a foot position across the midline directed away from the intended direction of travel and positioned their trunk more toward the intended direction of travel compared with the reactive condition.¹⁷  These preparatory movement strategies used during the penultimate step may allow for more effective deceleration during preplanned cutting than during reactive cutting, but this notion would need to be evaluated in a future study. With a less effective deceleration strategy during the penultimate step, our participants may not have been able to dissipate the momentum necessary to complete the reactive cut at the same angle as in preplanned cuts, thereby changing direction at a shallower angle. However, even when modifying their cutting angle by approximately 11°, participants still generated greater braking impulse during the final step of reactive cutting than during preplanned cutting.

In addition to generating a posteriorly directed braking impulse to dissipate linear momentum, athletes must generate a medial ground reaction force impulse to achieve translation away from the original direction and begin orienting the center of mass in its new direction.¹¹  Consistent with the findings of Havens and Sigward,¹¹  our participants produced greater medially directed impulse as well as greater peak and average medial ground reaction forces in the final step than in the penultimate step of preplanned cutting (Table 2). These findings were consistent in the reactive condition, yet the magnitudes of the differences between the penultimate and final steps were more pronounced in the reactive condition. Participants generated approximately 33% more medial ground reaction force impulse during the final step of reactive cutting than during the final step of preplanned cutting. Nonetheless, during the penultimate step, participants consistently exhibited lesser peak and average medial ground reaction force during reactive cuts than during preplanned cuts (Table 2). Our results suggest that when planning time is constrained, participants are left with greater braking and translation demands in the final step, which may also explain the differences in stance time and cutting angle between conditions. Without an effective deceleration strategy in the penultimate step of reactive cutting, participants in our study may have countered the increased braking and translation demand by lengthening the stance phase and changing direction at a shallower angle. In sport activities that are performed at faster velocities in a constrained area of space, it may not always be possible for athletes to lengthen the stance phase to the same extent or to cut at shallower angles to successfully evade defenders or execute sport-specific tasks. Instead, in various reactive scenarios, the athlete may have to decelerate and redirect the center of mass more abruptly, potentially leading to greater peak forces during deceleration. With increased braking during the final step, where redirection also occurs, the lower extremity would need to counter potentially increased forces that may result in injury.

Similar to any biological tissue, the ACL is injured when internal stresses exceed the ligament’s physiological capacity to withstand stress, resulting in deformation and tissue failure.³¹  At faster velocities or when cutting at a shallower angle is not feasible, a lack of braking during the penultimate step to arrest the center-of-mass velocity could result in the unfavorable deceleration mechanics often observed during noncontact ACL injury.^3,32  Without an effective deceleration strategy during the penultimate step of reactive cutting, our athletes produced approximately 60% more braking impulse during the final step of reactive cutting than during preplanned cutting. Even with the increased braking demand in the final step of reactive cutting, we did not identify elevated peak forces that would heighten the ACL injury risk. Although we believe these unexpected findings were related to approach velocity, achieved cutting angle, or the athletes’ ability to lengthen the stance phase (or a combination of these), future researchers should evaluate how athletes decelerate across the final 2 steps of a reactive cutting maneuver in a more sport-specific environment.

From a task execution perspective, if the individual is unable to use an effective braking strategy through the penultimate step, this may result in less effective attempts to complete a sport-specific skill such as cutting to evade a defender.¹¹  Additionally, the increased stance times we observed during the final step of reactive cutting may result in slower changes of direction and time to completion, as longer ground contact times have been associated with slower agility and T-test times in female basketball players.³³  We instructed our participants to accelerate out of each cut as quickly as possible, but we prioritized controlling for approach velocity and did not measure the time to completion of the task. Therefore, future authors should evaluate whether effective deceleration in the penultimate step is related to performance-related outcomes during side-step cutting.

The primary limitation of our study was that only biomechanical indicators of deceleration were assessed. Recent research from Byrne et al provided evidence that planning time influenced an athlete’s body positioning during the penultimate step of reactive cutting.¹⁷  As changes in deceleration are influenced by whole-body mechanics and the individual’s ability to position the body to generate braking forces and change direction, future researchers should investigate the relationship between the kinematic movement strategies described previously and the biomechanical indicators of deceleration we explored during the penultimate and final steps of reactive cutting. Furthermore, the light stimulus used to direct the reactive cutting direction may be more demanding than the external stimuli the athlete would encounter in a typical sport setting.³⁴  Specifically, the typical sport environment provides environmental cues, such as the body position of an opposing player or the trajectory of a ball, that a well-trained athlete could pick up on to potentially increase planning time.³⁴  Another limitation of our study was the possibility that the use of the force plates for both steps created a spatial constraint for stride length and width. To address this, we staggered the position of the force plates to accommodate both right- and left-leg–dominant individuals and wider step lengths (Figure 2). However, participants may still have modified their stride to successfully contact the force plates during the cutting maneuver. Finally, although we found that an ineffective penultimate step braking strategy during reactive cutting resulted in greater braking demand during the final step, it is unclear whether increased braking impulse during the final step of reactive cutting is linked to a heightened injury risk. Future authors should consider the use of inertial measurement units to compare the deceleration profiles in a laboratory-based setting versus a typical sport environment to address these limitations and determine the ecological validity of our results.

Our study was the first to comprehensively assess the biomechanical indicators of deceleration during the penultimate and final steps of both preplanned and reactive side-step cutting maneuvers. The results supply strong evidence that a lack of planning time results in a deceleration profile in which lesser braking is achieved during the penultimate step and greater braking and medially directed impulse are required during the final step of reactive cutting. As a consequence, participants achieved a lesser cutting angle and exhibited longer stance times during the final step of reactive cutting. Future investigators should examine using augmented feedback strategies to improve penultimate step deceleration during reactive cutting maneuvers. Doing so may also improve performance while decreasing the braking demand during the final step of reactive cutting.

We thank Kyle DeRosia, Cole McCallister, and Kyra Knox for their help with participant recruitment and data collection.