Wavelet analysis reveals differential lower limb muscle activity patterns long after anterior cruciate ligament reconstruction

https://doi.org/10.1016/j.jbiomech.2022.110957Get rights and content

Abstract

The purpose of this study was to test whether differences in muscle activity patterns between anterior cruciate ligament-reconstructed patients (ACLR) and healthy controls could be detected 10 to 15 years post-surgery using a machine learning classification approach. Eleven ACLR subjects and 12 healthy controls were recruited from an ongoing prospective randomized clinical trial. Surface electromyography (EMG) signals were recorded from gastrocnemius medialis and lateralis, tibialis anterior, vastus medialis, rectus femoris, biceps femoris, and semitendinosus muscles. Muscle activity was analyzed using wavelet analysis and examined within four sub-phases of the hop test, as well as an average of the task as a whole. K-nearest neighbor machine learning combined with a leave-one-out validation was used to classify the muscle activity patterns as either ACLR or Control. When muscle activity was averaged across the whole hop task, activity patterns for all muscles except the tibialis anterior were identified as being different between the study cohorts. ACLR patients demonstrated continuous muscle activities that spanned take-off, airborne, and landing hop phases versus healthy controls who displayed timed and regulated islets of muscle activities specific to each hop phase. The most striking features were 25–50% greater relative quadriceps intensity and approximately 66% diminished biceps femoris intensity in ACLR patients. The current findings are in contrast to previous work using conventional co-contraction and muscle activation onset EMG measures of the same dataset, underscoring the sensitivity and potential of the wavelet approach coupled with machine learning to reveal meaningful adaptation strategies in this at-risk population.

Introduction

Anterior cruciate ligament (ACL) reconstruction remains the clinical gold standard for restoring joint stability, re-establishing knee function, and enabling a return to sport following ACL injury (Chalmers et al., 2014, Siegel et al., 2012). However, compensatory muscle activity patterns (Hurd and Snyder-Mackler, 2007), altered neuromuscular coordination (Moraiti, et al., 2010) and persistent abnormal biomechanics (Delahunt et al., 2013, Deneweth et al., 2010, Devita et al., 1998, Georgoulis et al., 2010, Salem et al., 2003, Xergia et al., 2011) have been reported in ACL reconstructed (ACLR) patients, even after completing rehabilitation protocols and returning to pre-injury activities. It remains unclear whether these functional changes are permanent, to what degree they remain different from uninjured persons, and whether subtle differences in muscle activity could modulate abnormal joint motions and the long-term risk of post-traumatic osteoarthritis (PTOA).

To this end, surface electromyography (EMG) has been a valuable tool to quantify muscle activity across various healthy and injured cohort studies (Hanson et al., 2008, Ortiz et al., 2008, Shanbehzadeh et al., 2017, Wikstrom et al., 2008). Nevertheless, one of the challenges with EMG analyses is that extracting discrete outcomes (e.g., muscle activation onset, co-contraction indices, and normalized EMG signal amplitude) from continuous and complex electrical signals requires a priori selection of temporospatial variables. Consequently, much of the higher-order EMG signal information, such as the interrelationship between the magnitude, timing, and frequency, is discarded, creating the possibility that crucial physiologic muscle activity differences may remain undetected. This variable selection and/or omission problem is especially relevant to patient populations where subtle neuromuscular deficits may persist in ways we do not yet fully understand, such as long-term ACLR patients (>10 years post-reconstruction).

Conversely, wavelet analysis decomposes EMG signals into a series of overlapping non-linearly scaled wavelets that are superimposed to create a visual representation of the complex underlying signal attributes (von Tscharner, 2000). In this way, muscle EMG events with a given intensity at one frequency can be simultaneously compared to others at different frequencies over a continuous time frame. Although not yet widely adopted in orthopaedic biomechanics, the wavelet approach has been applied previously in studies of muscle activation patterns during running (Nüesch et al., 2012, Von Tscharner and Goepfert, 2006, von Tscharner et al., 2003), walking (Kuntze et al., 2015b, Mohr et al., 2019, von Tscharner and Valderrabano, 2010), stair climbing (Kuntze, et al., 2015a), and single-leg squats (Bishop, et al., 2020). In these works, supervised classification algorithms were applied to the wavelet patterns to characterize and compare continuous EMG wavelet properties.

Here, we applied wavelet analysis to the EMG signals recorded during a single leg hop-for-distance activity. We tested the hypothesis that differences in wavelet EMG patterns between ACLR and healthy controls could be detected at 10–15 years post-surgery and classified using a machine learning approach.

Section snippets

Subjects and criteria

Twenty-two subjects were recruited from an ongoing prospective randomized controlled trial that has followed ACLR and matched healthy control subjects over the past 15 years (NCT00434837) (Akelman et al., 2016, Fleming et al., 2021, Fleming et al., 2013): 11 ACLR subjects (5 males, 6 females; age = 34.7 ± 9.9 years; BMI = 27.7 ± 4.0; 11.9 ± 1.3 years post-follow-up) and 11 healthy control subjects (7 males, 4 females; age = 38.8 ± 6.5 years; BMI = 25 ± 3.2; 11.9 ± 1.3 years post initial

Results

The average wavelet muscle activities for the lower limb muscles are shown in Fig. 3. The intensity patterns in GM, GL, VM, RF, BF, and ST muscle groups illustrate that the Control group exhibited more regulated and phase-specific islets of muscle activities (e.g., white dashed rectangle in ACLR GM and VM activity versus the same regions in Control patterns in Fig. 3). In contrast, ACLR subjects demonstrated greater overall muscle activity that overlapped multiple hop phases (e.g., green dashed

Discussion

Analyzing EMG signals using wavelet analysis coupled with a machine learning approach made it possible to classify patterns of muscle activity unique to ACLR subjects, confirming our hypothesis. The finding that muscle activity patterns differed between the ACLR subjects and uninjured Controls contrasts our previous work where we failed to detect significant differences using conventional discrete EMG analyses in these same subjects (Behnke, et al., 2021). Based solely on muscle co-contraction

Declaration of Competing Interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Dr. Fleming receives royalties from Springer Publishing and a stipend from Sage Publishing as an associate editor for a medical journal and is a Miach Orthopaedics LLC founder. All other authors do not have any conflicts of interest that may have influenced this studys results. Results of the present study are presented honestly and without fabrication,

Acknowledgment

This study was supported by the National Institutes of Health [NIAMS K99/R00-AR069004, R01-AR047910, R01-AR074973, and NIGMS P30-GM122732 (Bioengineering Core of the COBRE Centre for Skeletal Health and Repair)], the Lucy Lippitt Endowment, and the RIH Orthopaedic Foundation. The authors would like to thank Cynthia Chrostek and Orianna Duncan for their assistance with subject recruitment and Erika Tavares for her guidance with data collection at the Keck Foundation XROMM Laboratory. The authors

References (49)

  • G.J. Salem et al.

    Bilateral kinematic and kinetic analysis of the squat exercise after anterior cruciate ligament reconstruction

    Arch. Phys. Med. Rehabil.

    (2003)
  • K.A. Taylor et al.

    Measurement of in vivo anterior cruciate ligament strain during dynamic jump landing

    J. Biomech.

    (2011)
  • V. von Tscharner

    Intensity analysis in time-frequency space of surface myoelectric signals by wavelets of specified resolution

    J. Electromyogr. Kinesiol.

    (2000)
  • V. Von Tscharner et al.

    Estimation of the interplay between groups of fast and slow muscle fibers of the tibialis anterior and gastrocnemius muscle while running

    J. Electromyogr. Kinesiol.

    (2006)
  • V. von Tscharner et al.

    Classification of multi muscle activation patterns of osteoarthritis patients during level walking

    J. Electromyogr. Kinesiol.

    (2010)
  • V. von Tscharner et al.

    Changes in EMG signals for the muscle tibialis anterior while running barefoot or with shoes resolved by non-linearly scaled wavelets

    J. Biomech.

    (2003)
  • M.R. Akelman et al.

    Effect of Matching or Overconstraining Knee Laxity During Anterior Cruciate Ligament Reconstruction on Knee Osteoarthritis and Clinical Outcomes: A Randomized Controlled Trial With 84-Month Follow-up

    Am. J. Sports Med.

    (2016)
  • S.D. Barber et al.

    Quantitative assessment of functional limitations in normal and anterior cruciate ligament-deficient knees

    Clin. Orthop. Relat. Res.

    (1990)
  • L.A. Bolgla et al.

    Reliability of lower extremity functional performance tests

    J. Orthop. Sports Phys. Ther.

    (1997)
  • M.T. Cavanaugh et al.

    Intrasession and Intersession Reliability of Quadriceps' and Hamstrings' Electromyography During a Standardized Hurdle Jump Test With Single Leg Landing

    J Strength Cond Res

    (2017)
  • P.N. Chalmers et al.

    Does ACL reconstruction alter natural history?: A systematic literature review of long-term outcomes

    J. Bone Joint Surg. Am.

    (2014)
  • J.D. Chappell et al.

    Kinematics and electromyography of landing preparation in vertical stop-jump: risks for noncontact anterior cruciate ligament injury

    Am. J. Sports Med.

    (2007)
  • M.S. Coats-Thomas et al.

    Effects of ACL reconstruction surgery on muscle activity of the lower limb during a jump-cut maneuver in males and females

    J. Orthop. Res.

    (2013)
  • E. Delahunt et al.

    Lower limb kinematics and dynamic postural stability in anterior cruciate ligament-reconstructed female athletes

    J Athl Train

    (2013)
  • Cited by (0)

    1

    Co-first authors.

    View full text