THE EFFECT OF TRAINING ON A COMPLIANT SURFACE ON MUSCLE ACTIVATION AND CO-CONTRACTION AFTER ANTERIOR CRUCIATE LIGAMENT INJURY

Abstract

Background

Performing physical activities on compliant surfaces alters joints kinematics by decreasing joint motions. However, the effect of administering a training program on a compliant surface on muscle activities after anterior cruciate ligament (ACL) injury is unknown.

Hypothesis/Purpose

To compare the effects of training on a compliant surface and manual perturbation training on individual muscle activation and muscle co-contraction indexes after an ACL injury. It was hypothesized that patients who received training on the compliant surface would demonstrate higher individual and combined muscle activities compared to the manual group.

Method

Sixteen patients (participated in level I/II sports) who sustained an ACL injury and had not undergone reconstructive surgery participated in this preliminary study. Eight patients received training on a compliant surface (Compliant group) and data of eight patients matched by age and sex from a previous study who received manual perturbation training were used as a control group (Manual group). Patients in both groups completed standard three-dimensional gait motion analysis with surface electromyography (EMG) of several lower extremity muscles during gait. Muscle co-contraction index and individual muscle activations were computed during weight acceptance (WA) and mid-stance (MS) intervals. A 2x2 analysis of variance (ANOVA) was used with an alpha level of p<0.10 to account for the high EMG variability.

Results

The compliant group significantly increased muscle co-contraction of vastus lateralis-lateral hamstring (VL-LH), vastus medialis-gastrocnemius medialis (VM-MG), and vastus lateralis (VL) muscle activity during WA (p ≤ 0.035) and manual group significantly decreased VM-MG muscle co-contraction during WA (p=0.099) after training.

Conclusion

Administering training on a compliant surface provides different effects on muscle activation compared to manual perturbation training after an ACL injury. Training on a compliant surface caused increased muscle co-contraction indexes and individual muscle activation, while manual perturbation training decreased the VM-MG muscle co-contraction index.

Level of evidence

2b

INTRODUCTION

An intact anterior cruciate ligament ACL works as a passive knee stabilizer, limiting tibiofemoral motions in the anterior-posterior and rotational directions.^1,2 Rupturing the ACL is a common knee injury in sports activities among young athletes.³ The hallmark symptom in patients with an ACL injury, who have not undergone reconstructive surgery, is dynamic knee instability (inability to perform high-level activities without episodes of giving way while maintaining normal movement patterns)⁴ that may limit the individuals’ ability to participate in activities of daily living and sport.^5,6 Repetitive dynamic knee instability may negatively impact knee structures and lead to the development of knee osteoarthritis.^7–10 The occurrence of dynamic knee instability might be attributed to the failure of the neuromuscular system to respond to destabilizing stimuli when attempting to maintain knee stability. Therefore, implementing a rehabilitation program that can improve the neuromuscular system's ability to stabilize the knee by developing quick and sufficient muscular forces might be beneficial for patients with an ACL injury who opting non-operative managment.

Previous studies have revealed that athletes who perform hopping and running activities on compliant surfaces alter joint kinematics by decreases in joint motion.^11–14 Decreases in joint motion were attributed to the need to offset the increased compliance of the surface and maintain similar locomotion of the body's center of mass while on changing surfaces.^12,14 A recent study also found that patients with an ACL injury who received training on a compliant surface walked with reduced joint motion.¹⁵ Patients in the aforementioned study received training on an automated mechanical device with an embedded plate that moved vertically, simulating a compliant surface.¹⁶ Ultimately, the training involved the performance of high-demand physical activities, including jumping, landing, pivoting, and hopping on the compliant surface. The protocol was designed to train patients with an ACL injury to trigger the neuromuscular system during a training program that included destabilizing stimuli and to actively control knee stability. However, the effects of administering the training program on a compliant surface on muscle activity have not yet been studied.

Treatment involving manual perturbation training augmented with progressive quadriceps strengthening is often chosen for non-operative management of patients with an ACL injury.^2,17 Manual perturbation training is a type of neuromuscular exercise that includes applying destabilizing translation and rotation of the support surfaces (using tilting and roller boards) underneath the patient's feet while the patient attempts to respond to the translation.^18–21 Manual perturbation training has been shown to be an effective intervention for improving dynamic knee stability and restoring knee motion during gait.^18–21However, manual perturbation training is limited to a standing position and requires one-on-one treatment by the therapist during the training.²² The effects of administering a mechanical perturbation training program with a compliant surface, on the neuromuscular activity after an ACL injury, remain unknown. Therefore, the purpose of this study was to compare the effects of training on a compliant surface and manual perturbation training on individual muscle activation and muscle co-contraction indexes in patients after an ACL injury. It was hypothesized that patients who received training on the compliant surface would demonstrate higher individual and combined muscle activities compared to the manual group.

METHODS

Eight patients with an isolated unilateral ACL rupture ( < seven months from initial injury) who had not undergone reconstruction surgery and were between the ages of 14-55 were enrolled in this preliminary prospective study. Patients underwent 10 training sessions on a compliant surface (Compliant group) using the Reactive Agility System (Simbex LLC, Lebanon, NH). Data of another eight patients, matched by age and sex, from a prior randomized controlled study who had received manual perturbation training (Manual group) were used as a control group.^23,24 Patients in both groups were regular participants in level I or II sports activities that included jumping, cutting, pivoting, and lateral maneuvers (e.g. soccer, basketball, football, baseball, and field hockey) for at least 50 hours per year prior to their ACL injury.²⁵ For both groups, patients with serious lower limb injuries (e.g. fracture), concurrent ligamentous injury, repairable meniscus injury, osteochondral defect more than 1 cm², or less than 3 mm side-to-side difference in passive anterior knee laxity measured with a KT-2000 arthrometer²⁶ were excluded from this study. Complete ACL injury and concomitant knee injuries were confirmed using both physical examination and magnetic resonance imaging. The protocol of this study was approved by the Institutional Human Subjects Review Board at the University of Delaware. One patient was under the age of 18 years and therefore both parental consent and patient assent was obtained. Patients were mainly recruited from the physical therapy clinic at the University of Delaware and local physician clinics. All patients provided written informed consent for participation in this study.

Training program

Patients were enrolled in this study once they demonstrated no knee pain, minimal knee joint effusion as measured with the modified stroke test,²⁷ full knee joint range of motion, and ≥ 70% quadriceps index. The patients in the compliant group completed a training protocol on the Reactive Agility System device, which has an embedded plate. The embedded plate was controlled by software so that it was either locked (no motion) or moved in the vertical direction simulating a compliant surface. The protocol included ten sessions of dynamic stabilization training and plyometric training. The complete training protocol for the compliant surface was previously described in detail by Nawasreh et al.¹⁶ The training program included ten training sessions of 30 minutes that were administered three to five times per week. In short, the training program was divided into three phases, with the first phase (sessions 1-3) included bilateral limb training and emphasized dynamic stabilization. The second phase (sessions 4-7) included unilateral limb training and emphasized dynamic stabilization and plyometric exercises. The third phase (sessions 8-10) was similar to the second phase, except sports-specific activities were added. Patients were given instruction and information during the first session on how to perform the training, and verbal cues were provided to inform the patient on when the plate would change its status by dropping down. The training was progressed to be more random, with unanticipated dropdowns of the plate starting in the third session. The training program was completed on the mechanical device, with the changes to the state of the plate (dropping down) administered in an unanticipated manner at different phases of the training activities (i.e. jumping up, landing, turning, squatting, and standing up from squat). The magnitude and velocity of dropdowns were fixed throughout the program's phases. The difficulty of the program was systematically progressed throughout the training, including increases in the difficulty of the activity, progressing from double limbs to a single limb, increasing the duration for each set and number of repetitions, and by integrating sport-specific skills. Patients were progressed throughout the program using a criterion similar to that used for the manual perturbation training.¹⁸ Rest time was provided between sets to avoid fatigue, with soreness rules (Appendix A) implemented during sessions. The therapist monitored the patients’ responses to training and provided feedback to ensure that the activities were performed appropriately.

Data of eight patients from a previous study, matched by age and sex, were used as a control group.^23,24 These patients had received manual perturbation training that was previously described by Fitzgerald et al.¹⁸ These patients received purposeful manual manipulation of the support surfaces that was administered by therapists. The manual perturbation training included three conditions: rockerboard, rollerboard, and a combination of rollerboard and platform (Appendix B).¹⁹ The manual perturbation training and testing procedure were administered by well-trained and reliable therapists.

In addition to the perturbation training, patients in both groups with quadriceps strength index <80% received supervised, progressive quadriceps strength training of the involved lower extremity, augmented with neuromuscular electrical stimulation (NMES) in order to improve quadriceps strength. Patients with quadriceps strength index between 80 and 90% received supervised progressive strength training of the involved lower extremity without NMES. Patients with quadriceps strength index more than 90% were instructed to start a strengthening program to further improve their quadriceps strength of both lower extremities.²⁸

Testing

Patients completed standard three-dimensional (3-D) gait motion analysis testing within two weeks prior to (pre-test) and two weeks after the completion (post-test) of the training protocols. Motion analysis testing of walking was captured using 8-camera motion analysis system (VICON, Oxford Metrics Ltd, London, United Kingdom) at a sampling rate of 120Hz. Twenty static retro-reflective markers were placed on specific boney landmarks of the foot, ankle, shank, thighs, and pelvis to locate joints’ centers and segments’ positions. Five rigid shell clusters were also placed on the patient's pelvis and distal-lateral aspects of the shanks and thighs to track segments motions during walking trials. An embedded 6-component force plate (Bertec Corp, Worthington, Ohio) was simultaneously used to collect kinetic data at a sampling rate of 1080 Hz during walking. Eight walking trials were captured for each limb at a patient's self-selected walking speed with ± 5% variability. Data from five good trials were then post-processed using rigid body analyses and inverse dynamics in Visual3D software (C-Motion, Inc., Germantown, MD, USA) to build individual patient skeleton module. Joints kinematic and kinetic data were low pass filtered at 6 Hz and 40 Hz, respectively. Walking trials were normalized to 100% of the stance phase. Initial contact (IC) and toe-off (TO) of stance phase were determined using a 50 N force threshold of the vertical ground reaction force vector.

Electromyography (EMG) data were also collected during walking trials using the MA-300 EMG System (Motion Lab Systems, Baton Rouge, LA). Pre-amplified, stainless steel, bipolar surface electrodes with a double differential configuration (38x19x8mm preamp size, 2 X 12mm disk diameter with an 18mm inter-electrode spacing, and a 12x3 mm bar between the disks; input impedance > 100mΩ, CMRR > 100 dB at 65Hz). After cleaning the skin with isopropyl alcohol and non-sterile gauze pads, electrodes were placed at the mid-muscle belly and in parallel with the muscle fibers of the vastus lateralis, rectus femoris, and vastus medialis (VL, RF, and VM), lateral and medial hamstrings (LH, MH), lateral and medial gastrocnemius (LG, MG), and soleus (SOL) muscles.

EMG data were collected at 1080 Hz. Prior to walking trials, muscle activity was collected while performing maximum voluntary isometric contractions (MVIC) that lasted four seconds for each muscle. Maximum muscle activity was used for normalization in post-processing. The maximum muscles activities during MVICs were visually inspected to verify signal quality and gains were adjusted as necessary to avoid signal clipping. MVICs for gastrocnemii were tested while standing with the patient utilizing a countertop to provide a counter resistance while plantarflexing their feet and rising up on their toes. Quadriceps muscles were tested in a seated position, with knees at 60 ° flexion, and ankles anchored to a table through chains and padded cuffs. Tibialis anterior muscles were tested in a long sitting position with patients’ knees extended and the ankles secured to a table. Hamstring muscles were tested while subjects laid prone and the knees in 30 ° flexion, with ankles anchored to a table through chains and padded cuffs. Finally, SOL muscles were tested while the patients on hands and knees position on a testing table, with both feet off the table and secured to the table in full dorsiflexion position. A resting trial that lasted for two seconds was collected to determine no muscle activity. EMG data hardware filtered the data with a bandpass from 20-2000 Hz and an anti-aliasing filter of 1000Hz; data were output at ± 5 volts. Analog EMG data was converted to digital data via a 16-bit A-to-D board. EMG data were time normalized and synced with gait data.

Three-dimensional gait and EMG data were post-processed using Visual 3D (C-Motion Inc., Germantown, Maryland). The EMG data were DC corrected, band-pass filtered at 30-350 Hz, and full-wave rectified. A linear envelope was created with a 6 Hz low-pass, second-order, phase-corrected, Butterworth filter. The linear envelope for each muscle was then normalized to the muscle's maximum activity, identified as the maximum for each muscle from either the MVIC testing or the walking trails.

Muscle co-contraction indexes, defined as the simultaneous activation of the agonist and anti-agonistic muscles (VL-LH, VM-MH, VL-LG, and VM-MG), were computed from the corresponding muscles’ linear envelope using the following formula:

$$\begin{array}{l}co-contractionindex=\left(\frac{\text{less}\text{active}\text{muscle}}{\text{more}\text{active}\text{muscle}}\right)\times \\ \left(\text{less}\text{active}\text{muscle}+\text{more}\text{active}\text{muscle}\right)\end{array}$$

²⁹

The average muscle activity was also computed for each muscle (VM, RF, VL, MH, LH, MG, LG, and SOL) from the corresponding muscles’ linear envelope. Individual muscle activity and muscle co-contraction indexes for muscle pairs were averaged during two intervals of the stance phase of gait: weight acceptance (WA) and mid-stance (MS) intervals. The WA interval was defined from 100ms prior to IC (to account for the electromechanical delay)³⁰ to peak knee flexion (PKF). The MS interval was defined from PKF to peak knee extension (PKE).

Statistics

Independent t-tests were used to determine significant differences between groups (Compliant and Manual) for patients’ demographics, muscle co-contraction index, and individual muscle activation measures at the pre-test session. A 2x2 analysis of variance (ANOVA) was used to evaluate differences between groups (Compliant and Manual), and over time (pre-test and post-test) for muscle co-contraction and individual muscle activation measures. The alpha level for comparisons of muscle activity and muscle co-contraction variables was conservatively set at p < 0.10 to account for the high variability in the EMG data.³¹ The alpha level for comparison between subjects’ demographics was set at p < 0.05. All statistical analyses were performed using the SPSS 22.0 (IBM Company, Chicago, Illinois, USA). Bonferroni corrections were used to adjust for multiple comparisons when necessary, and Eta squared were calculated as an indicator of effect size (ES). Univariate ANOVAs were used for measures that were significantly different between groups at pre-testing with the corresponding values of the significantly different measures entered as covariates.

RESULTS

Twelve patients were recruited to receive mechanical training on a compliant surface (compliant group), however, only eight of them completed the study. Reasons for the four patients who did not complete the study include one patient stopped participation in the study for inconvenience of commuting, one patient stopped participation in the study for academic purposes, one patient dropped and did not return to the follow-up, and one patient reported pain during pre-training testing and which excluded the patient from being able to continue participation in the study.

There were no differences between groups for the patients’ demographic and walking speed measures at pre-testing (p>0.122) (Table 1). Significant differences between groups, at pre-testing, were found for muscle co-contraction indexes of VM-MH during WA (Manual: 0.20 ± 0.13, Compliant: 0.10 ± 0.04, p=0.07) and for individual muscle activity of VM during MS (Manual: 0.07 ± 0.03, Compliant: 0.10 ± 0.03, p=0.065).

Table 1.

Patient demographic measures for manual and compliant groups at pre-training testing reported as Mean (SD).

Variables	Manual group (n=8)	Compliant group (n=8)	p-value
Subjects #(women/men)	8 (4/4)	8 (4/4)
Age (years)	32.48 (12.43)	33.50 (13.43)	0.877
BMI (N/M2)	29.13 (5.03)	25.70 (3.03)	0.122
Time from injury to pre-testing session (weeks)	9.55 (10.45)	8.61 (6.19)	0.831
Walking speed (m/s)	1.54 (0.16)	1.52 (0.14)	0.583

BMI- mody mass index, N/M² = Newtons/meters squared, m/s = meters/second

Individual muscle activity

A significant group-by-time interaction was found for the VL muscle activation during WA (p=0.037, ES: 0.28) (Table 2). Post hoc testing indicated that the compliant group significantly increased the VL muscle activity during WA (p=0.027). Significant main effect of group was found for LH muscle activation during WA (p=0.075, ES: 0.209, Manual: 0.158 ± 0.10, Compliant: 0.229 ± 0.10) and during MS (p=0.083, ES: 0.199, Manual: 0.077 ± 0.13, Compliant: 0.161 ± 0.13).

Table 2.

Muscle co-contraction indexes and individual muscle activity of manual and compliant groups at pre and post-testing, presented as mean (SD), statistical comparisions and effect sizes.

	Pre-testing		Post-testing		Group x Time (p-value)	Effect Size
	Manual (n=8)	Compliant (n=8)	Manual (n=8)	Compliant (n=8)
VL-LH during WA	0.21(0.10)	0.13 (0.09)	0.17 (0.07)	0.26 (0.22)	0.01	0.38
VL-LH during MS	0.05 (0.03)	0.07 (0.07)	0.05 (0.05)	0.15 (0.24)	0.38	0.06
VM-MH during WA	0.20 (0.13)	0.10 (0.04)	0.13 (0.7)	0.25 (0.44)	0.20	0.12
VM-MH during MS	0.06 (0.06)	0.04 (0.04)	0.047 (0.049)	0.21 (0.47)	0.33	0.07
VL-LG during WA	0.097 (0.04)	0.08 (0.04)	0.13 (0.07)	0.36 (0.39)	0.12	0.16
VL-LG during MS	0.12 (0.15)	0.8 (0.07)	0.13 (0.23)	0.18 (0.24)	0.43	0.05
VM-MG during WA	0.094 (0.06)	0.08 (0.03)	0.07 (0.04)	0.35 (0.46)	0.099	0.18
VM-MG during MS	0.06 (0.04)	0.07 (0.05)	0.07 (0.08)	0.22 (0.49)	0.48	0.04
MH activity during WA	0.12 (0.08)	0.13 (0.8)	0.13 (0.62)	0.21 (0.22)	0.39	0.05
MH activity during MS	0.06 (0.05)	0.04 (0.03)	0.07 (0.06)	0.14 (0.25)	0.33	0.07
LH activity during WA	0.15 (0.06)	0.21 (0.08)	0.17 (0.07)	0.25 (0.14)	0.62	0.02
LH activity during MS	0.077 (0.05)	0.12 (0.11)	0.077 (0.06)	0.20 (0.199)	0.33	0.07
LG activity during WA	0.08 (0.03)	0.07 (0.05)	0.08 (0.03)	0.22 (0.23)	0.103	0.18
LG activity during MS	0.195 (0.08)	0.14 (0.06)	0.188 (0.05)	0.18 (0.24)	0.62	0.02
MG activity during WA	0.066 (0.05)	0.068 (0.04)	0.053 (0.026)	0.17 (0.24)	0.20	0.12
MG activity during MS	0.21 (0.09)	0.25 (0.08)	0.22 (0.14)	0.31 (0.20)	0.59	0.02
SOL activity during WA	0.12 (0.06)	0.11 (0.08)	0.10 (0.05)	0.18 (0.26)	0.31	0.08
SOL activity during MS	0.28 (0.08)	0.28 (0.087)	0.27 (0.068)	0.30 (0.23)	0.70	0.01
VL activity during WA	0.16 (0.07)	0.16 (0.15)	0.16 (0.06)	0.26 (0.19)	0.037	0.28
VL activity during MS	0.047 (0.02)	0.077 (0.06)	0.037 (0.026)	0.12 (0.137)	0.13	0.16
VM activity during WA	0.12 (0.04)	0.14 (0.012)	0.12 (0.05)	0.18 (0.24)	0.42	0.05
VM activity during MS	0.07 (0.04)	0.10 (0.02)	0.049 (0.036)	0.14 (0.26)	0.82	0.08

WA: Weight acceptance interval of stance phase, MS: mid-stance interval of stance phase, VL: vastus lateralis, VM: vastus medialis, LH: lateral hamstrings, MH: medial hamstrings, LG: lateral gastrocnemius, MG: medial gastrocnemius, SOL: soleus muscle. Boldfaced p-values indicate statistically significant differences at p<0.10.

Muscle co-contraction

Significant group-by-time interactions were found for VL-LH (p=0.012, ES: 0.38) and VM-MG during WA (p=0.099, ES: 0.18) (Table 2). Post hoc testing indicated that the compliant group increased the muscle co-contraction indexes of VL-LH and VM-MG during WA significantly after training (p<0.035). Whereas, the manual group significantly decreased the muscle co-contraction index of VM-MG during WA after training (p=0.099). Significant main effect of time was observed for the VL-LG during WA (Pre-testing: 0.09 ± 0.04, Post-testing: 0.24 ± 0.28, p=0.06, ES: 0.23) (Table 2)

DISCUSSION

The results of this study revealed that both training modes have an effect on muscle activation in patients with an ACL injury. A noteworthy finding of this study is that the training on a compliant surface appears to have induced a different effect on muscle activation compared to manual perturbation training. Moreover, the training on a compliant surface increased some of the muscle co-contraction indexes and individual muscles’ activities in patients with an ACL injury in order to increase joint stiffness.

Both training modes altered the muscular activation patterns after an ACL injury, which might be due to the fact that perturbation training was designed to target the neuromuscular system. It worths to mention that the effect size of the changes to muscle activation was small. However, the two forms of perturbation training induced differential effects on muscle co-contraction indexes and individual muscle activation in patients with an ACL injury. This might be attributed to the differences in the training protocols between groups and the muscular demands that correspond with the respective training activities. The training program on a compliant surface may have led to increased muscle activity in order to stiffen the joints and compensate for the compliance of the surface. Further, the training program on a compliant surface consisted of physical activities that required high muscle activation to generate the sufficient muscle forces required for jumping up, turning, and hopping to and from the compliant surface. In comparison, manual perturbation training did not require high muscle activation as it was limited to a standing position.

The results of this study revealed that training on a compliant surface induced statistically significant changes in individual and gross muscle activation compared to the manual perturbation training. However, the effect size of this change is only considered small. It has been found previously that performing physical activities on compliant surfaces can lead to decreased joint motion and increased joint stiffness to offset the increase in compliance of the surfaces, and to maintain the locomotion of the body's center of mass during walking on surfaces with different level of compliance.^11–13,¹⁵ Therefore, the increased muscle co-contraction indexes may explain the decreased knee joint angles reported in previous studies.^11–13 In this study, the training program on a compliant surface included physical activities with emphasis on dynamic stabilization training and plyometric activities that typically require greater muscle force generation. While, in this study, the patients were tested using standard 3-D motion analysis of during gait, which did not include any compliant surfaces, the results suggest that the administering training on a compliant surface may have effects on the muscle activation that transfer to other activities. The transfer of joint protective muscle activation patterns effects may be helpful for improving knee stability while performing dynamic physical activities.

The results of this study also indicate that patients who received their training on a compliant surface increased the individual muscle activation of VL during WA. It has been previously reported that after ACL rehabilitation, the VL muscle contributes to constraining the internal rotation of the tibia as a protective mechanism.³² The increased VL muscle activation during WA may indicate a successful adaptation to control knee stability in the sagittal and frontal planes during the loading response. The compliant group demonstrated a higher LH muscle activation during WA and MS compared to the manual group at both pre- and post-testing sessions. It is unclear why the compliant group had higher LH muscle activity compared to the manual group. Higher LH muscle activity in the compliant group might have resulted from the physical activity they engaged in prior to and after the ACL injury. Additionally, it might have resulted as a compensatory strategy to an ACL injury, knowing that not all patients respond similarly to the ACL injury.^6,29 Numerous studies reported that patients after ACL injury demonstrate an increase in knee flexor muscles activation during the stance phase of gait cycle.^33–35 This, in part, might help stabilize the knee joint, as the knee flexors work as an agonist for the ACL.⁴ The increase in LH muscle activity may counteract anterior tibial translation induced by the effect of eccentric and concentric contractions of the quadriceps during WA and MS respectively.

The patients who received training on compliant surface demonstrated an increased in muscle co-contraction indexes of the lateral (VL-LH and VL-LG) and medial (VM-MG) muscles that cross the knee joint during WA interval of the gait cycle. The overall increased of gross muscle co-contraction may help in stabilizing the knee joint agonist anterior tibial translation and internal knee rotation and prevent shear forces between the joint surfaces during dynamic activities. Therefore, the increase in the muscle co-contraction indexes may help protect the knee joint integrity and prevent the development of knee osteoarthritis.^36–38 The increased of VM-MG muscle co-contraction index may also control the knee valgus motion during the loading response of the gait cycle. As there is a lack of control data regarding levels of “normal” muscle co-contraction during walking in healthy athletes, it is hard to determine whether the increased in the muscle co-contraction indexes of the compliant group is a positive compensatory strategy or not. However, previous work by Palmieri-Smith and colleagues revealed that healthy men and women both demonstrate an imbalance between the medial and lateral quadriceps-hamstring muscle co-contraction, with women demonstrating higher medial-to-lateral differences.³⁹ Letafatkar and colleagues reported increased co-contraction in healthy women who had neuromuscular deficits after perturbation training.^40,41 As a response, the increased co-contraction indexes of this study may be a positive adaptation that contributes to knee joint stability.⁴

The increased gross muscle co-contraction activity, from a clinical standpoint, may be concerning as it may cause knee stiffness that compromises normal knee mechanics during functional activities (i.e. reduced knee motion and moments).^20,35,42,43Additionally, it has been advocated that long-lasting increases in muscle co-contraction along with decreases knee motion may have negative impacts on patient's performance during high demand activities.^44,45 Furthermore, it may have a negative impact on knee joint integrity as it might increase the net compressive forces between the articular surfaces and cause an initiation or progression of knee osteoarthritis.^44,46 Therefore, further work is needed to investigate the impact of increased muscle co-contraction on knee joint health. In comparison, the results of this study also indicate that manual perturbation training decreased the muscle co-contraction for VM-MG during WA. This may support utilizing the manual perturbation training to improve the neuromuscular response without compromising knee joint motion. Therefore, it cannot be concluded, based on the findings of this study, that administering a training program on the compliant surface is superior to the manual perturbation training. Instead, both types of training might be integrated into the rehabilitation program after the ACL injury.

The limitations of this study include the small sample size; therefore, the results should be taken as preliminary findings. Further work with a larger sample may be needed. The relational component of this study relied on data from a previous study for the manual group, however, the study was conducted with the same protocols, by well-trained investigators with similar requirements for reliability for the manual perturbation training, as well as during clinical and motion analysis testing. Another limitation may be related to the differences in the two training protocols, in terms of the demand of tasks and the types and directions of the perturbation stimuli provided, on a compliant surface from that of manual perturbation training. However, training on the compliant surface was used an alternative training while the manual perturbation training has been used as a standard treatment for non-operative ACL rehabilitation programs. Another limitation of this study is the usage of a large alpha value, chosen to account for the variability in the EMG data during walking trials. With a typical alpha value of p < 0.05 small differences in muscle activation would not have been detected.

CONCLUSIONS

The results of this study indicate that administering training on a compliant surface induced effects different from that of the effects seen due to manual perturbation training. Additionally, training on a compliant surface increased the gross muscle co-contraction indexes and individual muscle activation during gait compared to the manual group. The increases seen in the muscle co-contraction and individual muscle activation in the compliant group may reflect the adaptations that take place in the neuromuscular system to provide dynamic knee stability.