Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Monitoring recovery of gait balance control following concussion using an accelerometer David Howell 1, Louis Osternig, Li-Shan Chou n Department of Human Physiology, University of Oregon, Eugene, OR, USA
art ic l e i nf o
a b s t r a c t
Article history: Accepted 15 June 2015
Despite medical best-practice recommendations, no consistent standard exists to systematically monitor recovery from concussion. Studies utilizing camera-based systems have reported center-of-mass (COM) motion control deficits persisting in individuals with concussion up to two months post-injury. The use of an accelerometer may provide an efficient and sensitive method to monitor COM alterations following concussion that can be employed in clinical settings. This study examined: (1) frontal/sagittal plane acceleration characteristics during dual-task walking for individuals with concussion and healthy controls; and (2) the effectiveness of utilizing acceleration characteristics to classify concussed and healthy individuals via receiver operating characteristic (ROC) curve analyses. Individuals with concussion completed testing within 72 h as well as 1 week, 2 weeks, 1 month, and 2 months post-injury. Control subjects completed the same protocol in similar time increments. Participants walked and simultaneously completed a cognitive task while wearing an accelerometer attached to L5. Participants with concussion walked with significantly less peak medial–lateral acceleration during 55–75% gait cycle (p¼ 0.04) throughout the testing period compared with controls. Moderate levels of sensitivity and specificity were found at the 72 h and 1 week testing times (sensitivity ¼0.70, specificity ¼0.71). ROC analysis revealed significant AUC values at the 72 h (AUC ¼0.889) and two week (AUC ¼ 0.810) time points. Accelerometer-derived measurements may assist in detecting frontal plane control deficits during dual-task walking post-concussion, consistent with camera-based studies. These initial findings demonstrate potential for using accelerometry as a tool for clinicians to monitor gait balance control following concussion. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Acceleration Center-of-mass Balance Locomotion Mild traumatic brain injury
1. Introduction Gait balance disturbances have been identified in individuals following concussion. Individuals with concussion exhibit greater displacement and velocity of the whole body center-of-mass (COM) in the frontal plane and smaller peak COM velocity in the sagittal plane during dual-task walking than individually matched and uninjured controls (Howell et al., 2015a, 2015b, 2014, 2013; Parker et al., 2006). Such COM position or velocity differences could be due to poor momentum control resulting from difficulties in regulating COM acceleration. As the examination of COM acceleration during a sit-to-stand task has been shown to n Correspondence to: Department of Human Physiology, 122 Esslinger Hall, 1240 University of Oregon, Eugene, OR 97403. Tel.: þ1 541 346 3391; fax: þ1 541 346 2841. E-mail address: [email protected] (L.-S. Chou). 1 Current Position: The Micheli Center for Sports Injury Prevention, Waltham, MA, USA; Division of Sports Medicine, Department of Orthopedics, Brain Injury Center, Boston Children's Hospital, Boston, MA, USA.
differentiate individuals with functional limitations from healthy individuals (Fujimoto and Chou, 2014, 2012), the measurement of COM acceleration may help to identify how individuals control balance during daily activities. Recent studies have utilized accelerometers placed at L5 (King et al., 2014), the anterior midline of the pelvis (Furman et al., 2013), and the forehead (Brown et al., 2014) to increase the objectivity in measuring static balance postconcussion. However, no studies have examined acceleration during gait following concussion. Best-practice statements have recommended that balance testing is a critical component in the clinical examination of concussion (Harmon et al., 2013; McCrory et al., 2013). Laboratorybased studies have identified gait balance control deficits following concussion, particularly in whole body COM displacement and velocity (Howell et al., 2014, 2013). However, analysis of COM displacement and velocity may require substantial motion capture equipment not often available in clinical settings where postconcussion assessments are conducted. Hence, a need exists to develop protocols utilizing inexpensive, easily implemented
http://dx.doi.org/10.1016/j.jbiomech.2015.06.014 0021-9290/& 2015 Elsevier Ltd. All rights reserved.
Please cite this article as: Howell, D., et al., Monitoring recovery of gait balance control following concussion using an accelerometer. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.06.014i
D. Howell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
alternatives that provide accuracy and sensitivity comparable to camera-based systems for measuring gait balance control in a wide spectrum of clinical sites. As COM displacement and velocity are directly connected to COM acceleration, the use of accelerometers may offer a cost-effective, readily available alternative for measuring longitudinal alterations in the magnitude or pattern of COM motion during gait following concussion. Through the use of dual-task paradigms, previous investigations have observed that the completion of simultaneous motor and cognitive tasks are affected by concussion (Fait et al., 2013; Howell et al., 2013; Martini et al., 2011). As such, dual-task testing has been identified as a useful component in the management of concussion (Register-Mihalik et al., 2013b) and as method which may detect subtle neurological impairments following injury (Lee et al., 2013). Dual-task testing in concussion assessments may provide information about the ability to multitask in everyday activities (Howell et al., 2013) and allow for a greater ability to detect impairments following concussion than single-tasks focusing solely on motor or cognitive task completion. The objective of this study was to examine peak sagittal and frontal plane accelerations, measured with an accelerometer during dual-task walking, in a cohort of participants with concussion within 72 h of injury and regularly over the subsequent 2-month post-injury period and similarly in uninjured controls. The effectiveness of utilizing acceleration characteristics to classify subjects was also assessed. It was hypothesized that differences between groups (concussion and control) in forward and side-to-side movements would be detectable with a single accelerometer placed at L5, and that protocol measurements would demonstrate moderate to high sensitivity and specificity in discriminating between participants with and without concussion.
2. Methods 2.1. Participants Seventeen subjects were identified and recruited for testing. Ten subjects were diagnosed with a concussion by a healthcare professional (physician/athletic trainer) and seven uninjured subjects served as controls (Table 1). The definition of concussion was consistent with the latest best-practice statement (McCrory et al., 2013) as an injury caused by a direct blow to the head, face, neck, or elsewhere in the body resulting in impaired neurologic function and clinical symptoms. Prior to data collection, the institutional review board reviewed and approved the study protocol. All subjects and parents/guardians (if under the age of 18) provided written consent to participate in the study. Exclusion criteria for prospective subjects included: (1) lower extremity deficiency or injury, which may affect normal gait patterns; (2) history of cognitive deficiencies, such as permanent memory loss or concentration abnormalities; (3) history of three or more previous concussions; (4) loss of consciousness from the concussion lasting greater than one minute (for participants with concussion); (5) history of attention deficit hyperactivity disorder; or (6) a previously documented concussion within the past year. A prospective, repeated-measures design was employed where subjects with concussion reported to the laboratory for testing within 72 h post-injury as well as time points approximately 1 week, 2 weeks, 1 month, and 2 months post-injury. Table 1 Subject demographics: mean (SD) for each group of participants. Group
Concussion
Control
Sex Age (years) Height (cm) Mass (kg) Mechanism of injury (n)
7M/3F 19.0 (5.5) 170.9 (6.7) 67.1 (10.7) Head – head contact: 3 Head – opponent body contact: 2 Head – ground contact: 2 Bike accident: 2 Activities of daily living: 1
3M/4F 20.0 (4.5) 168.6 (8.4) 65.1 (10.5) N/A
Control subjects were initially assessed and then tested according to the same testing schedule as subjects with concussion. 2.2. Experimental protocol All study participants walked barefoot at a self-selected speed along a walkway while simultaneously completing an auditory Stroop test (dual-task). Gait balance control, under the Stroop test condition, has been shown previously to be affected in subjects with concussion for an extended post-injury period (Howell et al., 2014, 2013). The Stroop test consisted of the subject listening to four auditory stimuli: the recorded words “high” or “low” spoken in either a high or low pitch. Subjects were instructed to identify the pitch of the word, regardless of whether the pitch was congruent with the word. Each of the four stimuli was presented in random order at a specific time while walking. The first stimulus was presented once the subject had achieved steady state gait and was triggered by a photocell located several steps after gait initiation. Each of the three subsequent stimuli was presented one second following the previous response while the subject continued to walk. Prior to test administration, subjects were provided verbal instructions by a study investigator, along with examples of the auditory stimulus. Following confirmation that the subject understood and could adequately hear each auditory stimulus, a practice trial was completed with verbal accuracy feedback. After successful completion of the practice trial, experimental trials began. While walking, participants wore an accelerometer (Opal Sensor, APDM Inc., Portland, OR) attached with an elastic belt at L5, and linear acceleration was measured along three orthogonal axes so that the reference coordinate system xaxis was oriented vertically downward, the y-axis was oriented to the right, and the z-axis was oriented orthogonal to the x- and y-axes, toward the front. Accelerometer data were used to identify linear accelerations in the anterior–posterior and medial–lateral directions, obtained at a sampling frequency of 128 Hz and saved via data-logger for offline post-processing. In addition to the accelerometer, retro-reflective markers were placed on bony landmarks (Hahn and Chou, 2004) of the subject, and marker trajectory data were collected using a 10-camera motion analysis system (Motion Analysis Corp., Santa Rosa, CA) at a sampling rate of 60 Hz. Gait velocity was calculated as the mean forward velocity of the sacral marker during a gait cycle. 2.3. Data analysis For analysis of accelerometer signals, four trials were identified for each subject at each testing time point. Relevant raw data were filtered with a zero-lag, fourthorder Butterworth low pass filter with a cutoff frequency of 12 Hz. To remove the gravitational component from the accelerometer, participants also completed a 30second quiet standing trial. The RMS value in anterior and medial–lateral directions was subtracted from respective data. Heel strikes were used to identify the beginning and the end of the gait cycle, as previously demonstrated (Bugané et al., 2012; Zijlstra and Hof, 2003). Each trial was normalized to 100% gait cycle. The peak forward acceleration was identified at approximately 50% of the gait cycle (Fig. 1A). Three distinct peak medial–lateral accelerations were identified: (1) during 25–45% of the gait cycle, (2) during 45–55% of the gait cycle, and (3) during 55–75% of the gait cycle (Fig. 1B). Absolute values were used in further analysis to normalize between right and left directions. 2.4. Statistical analysis Peak accelerations were analyzed by two-way mixed effects analyses of covariance, with average gait velocity as a covariate, in order to determine the effect of group (concussion and control), time (72 h, 1 week, 2 weeks, 1 month, and 2 months post-injury) and the interaction between these two independent variables. For all omnibus tests, significance was set at p o 0.05. Follow-up pairwise comparisons were examined using the Bonferroni procedure to control family wise type I error. Effect size estimations for mean differences are reported as partial eta squared (ηp2) values. In order to evaluate the sensitivity and specificity of acceleration data to classify participants, and to evaluate the effectiveness of acceleration characteristics to discriminate between participants with concussion and healthy controls, further analyses were conducted on those variables which demonstrated a significant effect when tested using analysis of covariance. To determine sensitivity and specificity, predictive cutoff values which would indicate abnormal performance were determined using the lower limit of the 95% confidence interval (CI) of the control group mean at each testing time point (Register-Mihalik et al., 2013a). Sensitivity indicates the probability that a participant with concussion will be identified as abnormal using an established peak acceleration cutoff threshold. Specificity indicates the probability that a control participant would be correctly classified healthy using the same method (Register-Mihalik et al., 2013a). Furthermore, how well acceleration values accurately identify participants with a concussion was estimated using a receiver operating characteristic (ROC) curve analysis. The area under an ROC curve (AUC) was used as a measure of the ability of acceleration data to distinguish between the two participant groups, independent
Please cite this article as: Howell, D., et al., Monitoring recovery of gait balance control following concussion using an accelerometer. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.06.014i
D. Howell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ of any particular cutoff (Doi and Williams, 2013). The AUC was determined and tested against a null hypothesis of no classification ability (AUC ¼0.50). The maximum AUC value of 1.0 indicates a perfect identification of sensitivity and specificity, while an AUC of 0.50 indicates no discriminative value (Fan et al., 2006). All statistical analyses were performed with SPSS versions 20/21.
3. Results No significant demographic differences were observed between groups for age, height, or mass (Table 1). Concussion participants were assessed at 2 ( 70.8), 8 (7 1.3), 15 (71.6), 30 (7 2.4), and 55 (7 1.9) days post-injury. Control participants underwent an initial assessment and also at 7 ( 71.7), 14 (70.4), 27 (7 2.2), and 56
3
(73.1) days after the initial assessment. No study participant reported sustaining a concussion during the testing period. For self-selected average gait velocity during dual-task walking, subjects with concussion walked at an average (SD) of 1.01 m/s (0.14), 1.06 m/s (0.18), 1.12 m/s (0.12), 1.17 m/s (0.17), and 1.19 m/s (0.20) at the five chronological testing time points, respectively. Control subjects walked at an average (SD) of 1.25 m/s (0.13), 1.29 m/s (0.16), 1.27 m/s (0.13), 1.27 m/s (0.14), and 1.28 m/s (0.14) at the five chronological testing time points, respectively. A significant time group interaction (F¼8.28, p¼ 0.001, ηp2 ¼ 0.356) indicated that the concussion group walked with less average gait velocity than the control group at the 72 h (p¼ 0.003), 1 week (p ¼0.013), and 2 week (p ¼0.031) post-injury time points, but not at the 1 month or 2 month time points.
Fig. 1. Representative participant data: acceleration characteristic profile in the anterior–posterior (A) and medial–lateral (B) directions of movement during a single gait cycle.
Fig. 2. Mean 7 SE peak acceleration data for concussion and control groups across the two months of testing for peak forward (A) and peak medial–lateral (B/C/D) directions. Including average gait velocity as a covariate, results indicated a significant difference between groups across the two months of testing for peak acceleration 3 (D).
Please cite this article as: Howell, D., et al., Monitoring recovery of gait balance control following concussion using an accelerometer. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.06.014i
D. Howell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
Fig. 3. ROC curve analysis for the peak medial/lateral acceleration 3 variable at the 72 h and 2 week post-injury time points.
Analysis of covariance results indicated no significant interactions, or main effects of group or time for mean peak forward acceleration (Fig. 2A, p4 0.05). Peak medial–lateral acceleration 1 and peak acceleration 2 analyses also revealed no significant interaction effects, or main effects of group or time (Fig. 2B and C, p 40.05). Peak acceleration 3 results indicated that subjects with concussion displayed less peak frontal plane acceleration during 55–75% of the gait cycle during dual-task walking than control participants throughout the two months of testing [main effect of group; F(1, 14) ¼5.77, p ¼0.04, ηp2 ¼0.391; Fig. 2D]. As a main effect of group was detected for peak medial–lateral acceleration 3, sensitivity, specificity, and ROC curve analyses were conducted for this variable. Moderate levels of sensitivity and specificity were found at the 72 h (95% CI ¼2.72, 4.48 m/s2; sensitivity ¼0.70, specificity ¼0.71) and 1 week (95% CI ¼2.18, 4.71 m/s2; sensitivity¼0.70, specificity ¼0.71) testing times. The two week (95% CI ¼ 2.88, 4.75 m/s2; sensitivity ¼0.60, specificity ¼0.57), 1 month (95% CI ¼2.59, 4.74 m/s2; sensitivity ¼0.30, specificity¼ 0.71), and 2 month (95% CI ¼ 2.35, 4.66 m/s2; sensitivity¼0.40, specificity ¼0.57) post-injury time points demonstrated a weak ability to distinguish between groups. ROC analysis revealed significant AUC values at the 72 h (Fig. 3; AUC ¼0.889, p ¼0.01) and two week (Fig. 3; AUC¼0.810, p¼ 0.04) time points.
4. Discussion Data from this study indicate that during dual-task walking, participants with concussion displayed less peak medial–lateral acceleration than control participants during 55–75% gait cycle. This period represents the transition from double-support to single-support phases based on conventional gait analysis patterns (Perry and Burnfield, 1992). However, no effects of forward acceleration were observed. These data suggest that concussion may affect the ability to regulate whole-body side-to-side balance control during dual-task gait, possibly due to poor COM acceleration generation and regulation. This study provides evidence that acceleration alterations during dual-task walking may be present following concussion and the utilization of a single accelerometer may provide a method to detect these alterations without the requirement of
expensive camera-based motion analysis systems. These results appear consistent with previous camera-based studies reporting greater dual-task COM medial–lateral displacement and velocity for subjects with concussion as compared to healthy controls (Howell et al., 2014, 2013; Parker et al., 2006). Following concussion, the ability to effectively reduce whole body velocity toward the ipsilateral limb during the transition from the double to single support phase may be impaired (less peak acceleration 3, Fig. 2D). This also suggests an insufficiency to effectively redirect wholebody motion to the contralateral direction. Insufficient regulation of body momentum in the medial–lateral direction may have led to a greater and faster overall sway of the whole-body COM for the subjects tested in the current study. The study data indicated that participants with concussion were not significantly different from control participants in their ability to control forward momentum. Previous studies using camera-based systems have identified that concussion may initially affect the ability to maintain forward velocity during dual-task walking, but that this variable tended to normalize by two weeks post-injury (Howell et al., 2014, 2013; Parker et al., 2005). No differences between groups were detected for forward acceleration in the current study, suggesting that gait velocity may have influenced findings related to peak forward acceleration (Fig. 2A). However, differences were still detected between groups for peak medial–lateral acceleration 3 (Fig. 2D), suggesting that gait balance control differences in the frontal plane of movement may exist following concussion independent of gait velocity. Previous work reported obstacle navigation impairments while completing a Stroop task following concussion despite recovery of clinical symptoms (Fait et al., 2013). In a long-term retrospective analysis, altered gait patterns were reported to persist six years post-concussion (Martini et al., 2011). Using motion capture or instrumented walkways, researchers have observed effects following concussion on variables such as gait speed (Fait et al., 2013; Martini et al., 2011), gait termination strategies (Buckley et al., 2013), and COM displacement/velocity during gait (Howell et al., 2014, 2013). However, no studies have reported how whole-body acceleration during gait may be affected by concussion. One advantage of monitoring acceleration during movement is that it may provide further insights into how balance is controlled, and lead to a better understanding of the mechanisms underlying gait imbalance (Fujimoto and Chou, 2012). The finding that the peak medial–lateral acceleration 3 variable detected significant between-group differences throughout the two months of post-injury testing (Fig. 2D), and could effectively discriminate concussion participants from controls within 72 h and at two weeks post-injury using a ROC analysis, suggests that measuring acceleration may be a promising way to objectively track balance recovery following concussion and warrants further investigation into its clinical feasibility. The importance of gait and balance testing in the clinical examination of concussion has been recently documented (Harmon et al., 2013; McCrory et al., 2013). As such, accelerometry provides a readily available technology that can be straightforwardly employed in a clinical setting to facilitate the examination of motor and cognitive functions following concussion to potentially assist in clinical management. The current study was limited by uneven subject numbers and non-individually matched subject groups. However, the demographics of the two groups were not significantly different (Table 1), potentially mitigating the effects of this limitation. As this study enrolled a relatively small control sample, clinical interpretation of cutoff thresholds indicating abnormal performance should be done with caution. The sensitivity of the acceleration data to correctly classify participants with concussion decreased across the testing timeline, thus future studies should seek to establish normative reference ranges in cohorts of healthy
Please cite this article as: Howell, D., et al., Monitoring recovery of gait balance control following concussion using an accelerometer. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.06.014i
D. Howell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
individuals to provide a robust comparison to patients with concussion.
Conflict of interest statement The authors have no conflicts of interest which may have influenced the work reported in this study.
Acknowledgments This work was supported by Veterans Administration subcontract awards [A4842C8 & A4843C], Translational Research Award from the University of Oregon and Peace Health Oregon Region, the Department of Defense-TATRC Award (W81XWH-111-0717), and the Eugene and Clarissa Evonuk Memorial Graduate Fellowship in Environmental, Cardiovascular, or Stress Physiology. The authors would like to acknowledge Mr. Quinn Peterson and Ms. Madison Murray for their assistance with data collection.
References Brown, H.J., Siegmund, G.P., Guskiewicz, K.M., Van Den Doel, K., Cretu, E., Blouin, J.S., 2014. Development and validation of an objective balance error scoring system. Med. Sci. Sports Exerc. 46, 1610–1616. Buckley, T.A., Munkasy, B.A., Tapia-Lovler, T.G., Wikstrom, E.A., 2013. Altered gait termination strategies following a concussion. Gait Posture 38, 549–551. Bugané, F., Benedetti, M.G., Casadio, G., Attala, S., Biagi, F., Manca, M., Leardini, A., 2012. Estimation of spatial-temporal gait parameters in level walking based on a single accelerometer: validation on normal subjects by standard gait analysis. Comput. Methods Programs Biomed. 108, 129–137. Doi, S.A.R., Williams, G.M. (Eds.), 2013. Methods of Clinical Epidemiology, Springer Series on Epidemiology and Public Health. Springer, Berlin, Heidelberg, p. 69. Fait, P., Swaine, B., Cantin, J.-F., Leblond, J., McFadyen, B.J., 2013. Altered integrated locomotor and cognitive function in elite athletes 30 days postconcussion: a preliminary study. J. Head Trauma Rehabil. 28, 293–301. Fan, J., Upadhye, S., Worster, A., 2006. Understanding receiver operating characteristic (ROC) curves. CJEM 8, 19–20. Fujimoto, M., Chou, L.-S., 2012. Dynamic balance control during sit-to-stand movement: an examination with the center of mass acceleration. J. Biomech. 45, 543–548. Fujimoto, M., Chou, L.-S., 2014. Region of stability derived by center of mass acceleration better identifies individuals with difficulty in sit-to-stand movement. Ann. Biomed. Eng. 42, 733–741. Furman, G.R., Lin, C.-C., Bellanca, J.L., Marchetti, G.F., Collins, M.W., Whitney, S.L., 2013. Comparison of the balance accelerometer measure and balance error
5
scoring system in adolescent concussions in sports. Am. J. Sports Med. 41, 1404–1410. Hahn, M.E., Chou, L.-S., 2004. Age-related reduction in sagittal plane center of mass motion during obstacle crossing. J. Biomech. 37, 837–844. Harmon, K.G., Drezner, J.A., Gammons, M., Guskiewicz, K.M., Halstead, M., Herring, S.A., Kutcher, J.S., Pana, A., Putukian, M., Roberts, W.O., 2013. American Medical Society for Sports Medicine position statement: concussion in sport. Br. J. Sports Med. 47, 15–26. Howell, D.R., Osternig, L.R., Chou, L.-S., 2013. Dual-task effect on gait balance control in adolescents with concussion. Arch. Phys. Med. Rehabil. 94, 1513–1520. Howell, D.R., Osternig, L.R., Chou, L.-S., 2015a. Adolescents demonstrate greater gait balance control deficits after concussion than young adults. Am. J. Sports Med. 43, 625–632. Howell, D.R., Osternig, L.R., Chou, L.-S., 2015b. Return to activity after concussion affects dual-task gait balance control recovery. Med. Sci. Sports Exerc. 47, 673–680. Howell, D.R., Osternig, L.R., Koester, M.C., Chou, L.-S., 2014. The effect of cognitive task complexity on gait stability in adolescents following concussion. Exp. Brain Res. 232, 1773–1782. King, L.A., Horak, F.B., Mancini, M., Pierce, D., Priest, K.C., Chesnutt, J., Sullivan, P., Chapman, J.C., 2014. Instrumenting the balance error scoring system for use with patients reporting persistent balance problems after mild traumatic brain injury. Arch. Phys. Med. Rehabil. 95, 353–359. Lee, H., Sullivan, S.J., Schneiders, A.G., 2013. The use of the dual-task paradigm in detecting gait performance deficits following a sports-related concussion: a systematic review and meta-analysis. J. Sci. Med. Sport 16, 2–7. Martini, D.N., Sabin, M.J., DePesa, S.A., Leal, E.W., Negrete, T.N., Sosnoff, J.J., Broglio, S.P., 2011. The chronic effects of concussion on gait. Arch. Phys. Med. Rehabil. 92, 585–589. McCrory, P., Meeuwisse, W.H., Aubry, M., Cantu, R.C., Dvořák, J., Echemendia, R.J., Engebretsen, L., Johnston, K., Kutcher, J.S., Raftery, M., Sills, A., Benson, B.W., Davis, G.A., Ellenbogen, R., Guskiewicz, K.M., Herring, S.A., Iverson, G.L., Jordan, B.D., Kissick, J., McCrea, M., McIntosh, A.S., Maddocks, D., Makdissi, M., Purcell, L., Putukian, M., Schneider, K., Tator, C.H., Turner, M., 2013. Consensus statement on concussion in sport: the 4th international conference on concussion in sport, Zurich, November 2012. J. Athl. Train. 48, 554–575. Parker, T.M., Osternig, L.R., Lee, H.-J., Donkelaar, P., van, Chou, L.-S., 2005. The effect of divided attention on gait stability following concussion. Clin. Biomech. 20, 389–395. Parker, T.M., Osternig, L.R., Van Donkelaar, P., Chou, L.-S., 2006. Gait stability following concussion. Med. Sci. Sports Exerc. 38, 1032–1040. Perry, J., Burnfield, J.M., 1992. Gait Analysis: Normal and Pathological Function. Slack Inc., Thorofare, NJ, pp. 10–16. Register-Mihalik, J.K., Guskiewicz, K.M., Mihalik, J.P., Schmidt, J.D., Kerr, Z.Y., McCrea, M.A., 2013a. Reliable change, sensitivity, and specificity of a multidimensional concussion assessment battery. J. Head Trauma Rehabil. 28, 274–283. Register-Mihalik, J.K., Littleton, A.C., Guskiewicz, K.M., 2013b. Are divided attention tasks useful in the assessment and management of sport-related concussion? Neuropsychol. Rev. 23, 300–313. Zijlstra, W., Hof, A.L., 2003. Assessment of spatio-temporal gait parameters from trunk accelerations during human walking. Gait Posture 18, 1–10.
Please cite this article as: Howell, D., et al., Monitoring recovery of gait balance control following concussion using an accelerometer. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.06.014i
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Monitoring recovery of gait balance control following concussion using an accelerometer David Howell 1, Louis Osternig, Li-Shan Chou n Department of Human Physiology, University of Oregon, Eugene, OR, USA
art ic l e i nf o
a b s t r a c t
Article history: Accepted 15 June 2015
Despite medical best-practice recommendations, no consistent standard exists to systematically monitor recovery from concussion. Studies utilizing camera-based systems have reported center-of-mass (COM) motion control deficits persisting in individuals with concussion up to two months post-injury. The use of an accelerometer may provide an efficient and sensitive method to monitor COM alterations following concussion that can be employed in clinical settings. This study examined: (1) frontal/sagittal plane acceleration characteristics during dual-task walking for individuals with concussion and healthy controls; and (2) the effectiveness of utilizing acceleration characteristics to classify concussed and healthy individuals via receiver operating characteristic (ROC) curve analyses. Individuals with concussion completed testing within 72 h as well as 1 week, 2 weeks, 1 month, and 2 months post-injury. Control subjects completed the same protocol in similar time increments. Participants walked and simultaneously completed a cognitive task while wearing an accelerometer attached to L5. Participants with concussion walked with significantly less peak medial–lateral acceleration during 55–75% gait cycle (p¼ 0.04) throughout the testing period compared with controls. Moderate levels of sensitivity and specificity were found at the 72 h and 1 week testing times (sensitivity ¼0.70, specificity ¼0.71). ROC analysis revealed significant AUC values at the 72 h (AUC ¼0.889) and two week (AUC ¼ 0.810) time points. Accelerometer-derived measurements may assist in detecting frontal plane control deficits during dual-task walking post-concussion, consistent with camera-based studies. These initial findings demonstrate potential for using accelerometry as a tool for clinicians to monitor gait balance control following concussion. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Acceleration Center-of-mass Balance Locomotion Mild traumatic brain injury
1. Introduction Gait balance disturbances have been identified in individuals following concussion. Individuals with concussion exhibit greater displacement and velocity of the whole body center-of-mass (COM) in the frontal plane and smaller peak COM velocity in the sagittal plane during dual-task walking than individually matched and uninjured controls (Howell et al., 2015a, 2015b, 2014, 2013; Parker et al., 2006). Such COM position or velocity differences could be due to poor momentum control resulting from difficulties in regulating COM acceleration. As the examination of COM acceleration during a sit-to-stand task has been shown to n Correspondence to: Department of Human Physiology, 122 Esslinger Hall, 1240 University of Oregon, Eugene, OR 97403. Tel.: þ1 541 346 3391; fax: þ1 541 346 2841. E-mail address: [email protected] (L.-S. Chou). 1 Current Position: The Micheli Center for Sports Injury Prevention, Waltham, MA, USA; Division of Sports Medicine, Department of Orthopedics, Brain Injury Center, Boston Children's Hospital, Boston, MA, USA.
differentiate individuals with functional limitations from healthy individuals (Fujimoto and Chou, 2014, 2012), the measurement of COM acceleration may help to identify how individuals control balance during daily activities. Recent studies have utilized accelerometers placed at L5 (King et al., 2014), the anterior midline of the pelvis (Furman et al., 2013), and the forehead (Brown et al., 2014) to increase the objectivity in measuring static balance postconcussion. However, no studies have examined acceleration during gait following concussion. Best-practice statements have recommended that balance testing is a critical component in the clinical examination of concussion (Harmon et al., 2013; McCrory et al., 2013). Laboratorybased studies have identified gait balance control deficits following concussion, particularly in whole body COM displacement and velocity (Howell et al., 2014, 2013). However, analysis of COM displacement and velocity may require substantial motion capture equipment not often available in clinical settings where postconcussion assessments are conducted. Hence, a need exists to develop protocols utilizing inexpensive, easily implemented
http://dx.doi.org/10.1016/j.jbiomech.2015.06.014 0021-9290/& 2015 Elsevier Ltd. All rights reserved.
Please cite this article as: Howell, D., et al., Monitoring recovery of gait balance control following concussion using an accelerometer. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.06.014i
D. Howell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
alternatives that provide accuracy and sensitivity comparable to camera-based systems for measuring gait balance control in a wide spectrum of clinical sites. As COM displacement and velocity are directly connected to COM acceleration, the use of accelerometers may offer a cost-effective, readily available alternative for measuring longitudinal alterations in the magnitude or pattern of COM motion during gait following concussion. Through the use of dual-task paradigms, previous investigations have observed that the completion of simultaneous motor and cognitive tasks are affected by concussion (Fait et al., 2013; Howell et al., 2013; Martini et al., 2011). As such, dual-task testing has been identified as a useful component in the management of concussion (Register-Mihalik et al., 2013b) and as method which may detect subtle neurological impairments following injury (Lee et al., 2013). Dual-task testing in concussion assessments may provide information about the ability to multitask in everyday activities (Howell et al., 2013) and allow for a greater ability to detect impairments following concussion than single-tasks focusing solely on motor or cognitive task completion. The objective of this study was to examine peak sagittal and frontal plane accelerations, measured with an accelerometer during dual-task walking, in a cohort of participants with concussion within 72 h of injury and regularly over the subsequent 2-month post-injury period and similarly in uninjured controls. The effectiveness of utilizing acceleration characteristics to classify subjects was also assessed. It was hypothesized that differences between groups (concussion and control) in forward and side-to-side movements would be detectable with a single accelerometer placed at L5, and that protocol measurements would demonstrate moderate to high sensitivity and specificity in discriminating between participants with and without concussion.
2. Methods 2.1. Participants Seventeen subjects were identified and recruited for testing. Ten subjects were diagnosed with a concussion by a healthcare professional (physician/athletic trainer) and seven uninjured subjects served as controls (Table 1). The definition of concussion was consistent with the latest best-practice statement (McCrory et al., 2013) as an injury caused by a direct blow to the head, face, neck, or elsewhere in the body resulting in impaired neurologic function and clinical symptoms. Prior to data collection, the institutional review board reviewed and approved the study protocol. All subjects and parents/guardians (if under the age of 18) provided written consent to participate in the study. Exclusion criteria for prospective subjects included: (1) lower extremity deficiency or injury, which may affect normal gait patterns; (2) history of cognitive deficiencies, such as permanent memory loss or concentration abnormalities; (3) history of three or more previous concussions; (4) loss of consciousness from the concussion lasting greater than one minute (for participants with concussion); (5) history of attention deficit hyperactivity disorder; or (6) a previously documented concussion within the past year. A prospective, repeated-measures design was employed where subjects with concussion reported to the laboratory for testing within 72 h post-injury as well as time points approximately 1 week, 2 weeks, 1 month, and 2 months post-injury. Table 1 Subject demographics: mean (SD) for each group of participants. Group
Concussion
Control
Sex Age (years) Height (cm) Mass (kg) Mechanism of injury (n)
7M/3F 19.0 (5.5) 170.9 (6.7) 67.1 (10.7) Head – head contact: 3 Head – opponent body contact: 2 Head – ground contact: 2 Bike accident: 2 Activities of daily living: 1
3M/4F 20.0 (4.5) 168.6 (8.4) 65.1 (10.5) N/A
Control subjects were initially assessed and then tested according to the same testing schedule as subjects with concussion. 2.2. Experimental protocol All study participants walked barefoot at a self-selected speed along a walkway while simultaneously completing an auditory Stroop test (dual-task). Gait balance control, under the Stroop test condition, has been shown previously to be affected in subjects with concussion for an extended post-injury period (Howell et al., 2014, 2013). The Stroop test consisted of the subject listening to four auditory stimuli: the recorded words “high” or “low” spoken in either a high or low pitch. Subjects were instructed to identify the pitch of the word, regardless of whether the pitch was congruent with the word. Each of the four stimuli was presented in random order at a specific time while walking. The first stimulus was presented once the subject had achieved steady state gait and was triggered by a photocell located several steps after gait initiation. Each of the three subsequent stimuli was presented one second following the previous response while the subject continued to walk. Prior to test administration, subjects were provided verbal instructions by a study investigator, along with examples of the auditory stimulus. Following confirmation that the subject understood and could adequately hear each auditory stimulus, a practice trial was completed with verbal accuracy feedback. After successful completion of the practice trial, experimental trials began. While walking, participants wore an accelerometer (Opal Sensor, APDM Inc., Portland, OR) attached with an elastic belt at L5, and linear acceleration was measured along three orthogonal axes so that the reference coordinate system xaxis was oriented vertically downward, the y-axis was oriented to the right, and the z-axis was oriented orthogonal to the x- and y-axes, toward the front. Accelerometer data were used to identify linear accelerations in the anterior–posterior and medial–lateral directions, obtained at a sampling frequency of 128 Hz and saved via data-logger for offline post-processing. In addition to the accelerometer, retro-reflective markers were placed on bony landmarks (Hahn and Chou, 2004) of the subject, and marker trajectory data were collected using a 10-camera motion analysis system (Motion Analysis Corp., Santa Rosa, CA) at a sampling rate of 60 Hz. Gait velocity was calculated as the mean forward velocity of the sacral marker during a gait cycle. 2.3. Data analysis For analysis of accelerometer signals, four trials were identified for each subject at each testing time point. Relevant raw data were filtered with a zero-lag, fourthorder Butterworth low pass filter with a cutoff frequency of 12 Hz. To remove the gravitational component from the accelerometer, participants also completed a 30second quiet standing trial. The RMS value in anterior and medial–lateral directions was subtracted from respective data. Heel strikes were used to identify the beginning and the end of the gait cycle, as previously demonstrated (Bugané et al., 2012; Zijlstra and Hof, 2003). Each trial was normalized to 100% gait cycle. The peak forward acceleration was identified at approximately 50% of the gait cycle (Fig. 1A). Three distinct peak medial–lateral accelerations were identified: (1) during 25–45% of the gait cycle, (2) during 45–55% of the gait cycle, and (3) during 55–75% of the gait cycle (Fig. 1B). Absolute values were used in further analysis to normalize between right and left directions. 2.4. Statistical analysis Peak accelerations were analyzed by two-way mixed effects analyses of covariance, with average gait velocity as a covariate, in order to determine the effect of group (concussion and control), time (72 h, 1 week, 2 weeks, 1 month, and 2 months post-injury) and the interaction between these two independent variables. For all omnibus tests, significance was set at p o 0.05. Follow-up pairwise comparisons were examined using the Bonferroni procedure to control family wise type I error. Effect size estimations for mean differences are reported as partial eta squared (ηp2) values. In order to evaluate the sensitivity and specificity of acceleration data to classify participants, and to evaluate the effectiveness of acceleration characteristics to discriminate between participants with concussion and healthy controls, further analyses were conducted on those variables which demonstrated a significant effect when tested using analysis of covariance. To determine sensitivity and specificity, predictive cutoff values which would indicate abnormal performance were determined using the lower limit of the 95% confidence interval (CI) of the control group mean at each testing time point (Register-Mihalik et al., 2013a). Sensitivity indicates the probability that a participant with concussion will be identified as abnormal using an established peak acceleration cutoff threshold. Specificity indicates the probability that a control participant would be correctly classified healthy using the same method (Register-Mihalik et al., 2013a). Furthermore, how well acceleration values accurately identify participants with a concussion was estimated using a receiver operating characteristic (ROC) curve analysis. The area under an ROC curve (AUC) was used as a measure of the ability of acceleration data to distinguish between the two participant groups, independent
Please cite this article as: Howell, D., et al., Monitoring recovery of gait balance control following concussion using an accelerometer. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.06.014i
D. Howell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ of any particular cutoff (Doi and Williams, 2013). The AUC was determined and tested against a null hypothesis of no classification ability (AUC ¼0.50). The maximum AUC value of 1.0 indicates a perfect identification of sensitivity and specificity, while an AUC of 0.50 indicates no discriminative value (Fan et al., 2006). All statistical analyses were performed with SPSS versions 20/21.
3. Results No significant demographic differences were observed between groups for age, height, or mass (Table 1). Concussion participants were assessed at 2 ( 70.8), 8 (7 1.3), 15 (71.6), 30 (7 2.4), and 55 (7 1.9) days post-injury. Control participants underwent an initial assessment and also at 7 ( 71.7), 14 (70.4), 27 (7 2.2), and 56
3
(73.1) days after the initial assessment. No study participant reported sustaining a concussion during the testing period. For self-selected average gait velocity during dual-task walking, subjects with concussion walked at an average (SD) of 1.01 m/s (0.14), 1.06 m/s (0.18), 1.12 m/s (0.12), 1.17 m/s (0.17), and 1.19 m/s (0.20) at the five chronological testing time points, respectively. Control subjects walked at an average (SD) of 1.25 m/s (0.13), 1.29 m/s (0.16), 1.27 m/s (0.13), 1.27 m/s (0.14), and 1.28 m/s (0.14) at the five chronological testing time points, respectively. A significant time group interaction (F¼8.28, p¼ 0.001, ηp2 ¼ 0.356) indicated that the concussion group walked with less average gait velocity than the control group at the 72 h (p¼ 0.003), 1 week (p ¼0.013), and 2 week (p ¼0.031) post-injury time points, but not at the 1 month or 2 month time points.
Fig. 1. Representative participant data: acceleration characteristic profile in the anterior–posterior (A) and medial–lateral (B) directions of movement during a single gait cycle.
Fig. 2. Mean 7 SE peak acceleration data for concussion and control groups across the two months of testing for peak forward (A) and peak medial–lateral (B/C/D) directions. Including average gait velocity as a covariate, results indicated a significant difference between groups across the two months of testing for peak acceleration 3 (D).
Please cite this article as: Howell, D., et al., Monitoring recovery of gait balance control following concussion using an accelerometer. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.06.014i
D. Howell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
Fig. 3. ROC curve analysis for the peak medial/lateral acceleration 3 variable at the 72 h and 2 week post-injury time points.
Analysis of covariance results indicated no significant interactions, or main effects of group or time for mean peak forward acceleration (Fig. 2A, p4 0.05). Peak medial–lateral acceleration 1 and peak acceleration 2 analyses also revealed no significant interaction effects, or main effects of group or time (Fig. 2B and C, p 40.05). Peak acceleration 3 results indicated that subjects with concussion displayed less peak frontal plane acceleration during 55–75% of the gait cycle during dual-task walking than control participants throughout the two months of testing [main effect of group; F(1, 14) ¼5.77, p ¼0.04, ηp2 ¼0.391; Fig. 2D]. As a main effect of group was detected for peak medial–lateral acceleration 3, sensitivity, specificity, and ROC curve analyses were conducted for this variable. Moderate levels of sensitivity and specificity were found at the 72 h (95% CI ¼2.72, 4.48 m/s2; sensitivity ¼0.70, specificity ¼0.71) and 1 week (95% CI ¼2.18, 4.71 m/s2; sensitivity¼0.70, specificity ¼0.71) testing times. The two week (95% CI ¼ 2.88, 4.75 m/s2; sensitivity ¼0.60, specificity ¼0.57), 1 month (95% CI ¼2.59, 4.74 m/s2; sensitivity ¼0.30, specificity¼ 0.71), and 2 month (95% CI ¼ 2.35, 4.66 m/s2; sensitivity¼0.40, specificity ¼0.57) post-injury time points demonstrated a weak ability to distinguish between groups. ROC analysis revealed significant AUC values at the 72 h (Fig. 3; AUC ¼0.889, p ¼0.01) and two week (Fig. 3; AUC¼0.810, p¼ 0.04) time points.
4. Discussion Data from this study indicate that during dual-task walking, participants with concussion displayed less peak medial–lateral acceleration than control participants during 55–75% gait cycle. This period represents the transition from double-support to single-support phases based on conventional gait analysis patterns (Perry and Burnfield, 1992). However, no effects of forward acceleration were observed. These data suggest that concussion may affect the ability to regulate whole-body side-to-side balance control during dual-task gait, possibly due to poor COM acceleration generation and regulation. This study provides evidence that acceleration alterations during dual-task walking may be present following concussion and the utilization of a single accelerometer may provide a method to detect these alterations without the requirement of
expensive camera-based motion analysis systems. These results appear consistent with previous camera-based studies reporting greater dual-task COM medial–lateral displacement and velocity for subjects with concussion as compared to healthy controls (Howell et al., 2014, 2013; Parker et al., 2006). Following concussion, the ability to effectively reduce whole body velocity toward the ipsilateral limb during the transition from the double to single support phase may be impaired (less peak acceleration 3, Fig. 2D). This also suggests an insufficiency to effectively redirect wholebody motion to the contralateral direction. Insufficient regulation of body momentum in the medial–lateral direction may have led to a greater and faster overall sway of the whole-body COM for the subjects tested in the current study. The study data indicated that participants with concussion were not significantly different from control participants in their ability to control forward momentum. Previous studies using camera-based systems have identified that concussion may initially affect the ability to maintain forward velocity during dual-task walking, but that this variable tended to normalize by two weeks post-injury (Howell et al., 2014, 2013; Parker et al., 2005). No differences between groups were detected for forward acceleration in the current study, suggesting that gait velocity may have influenced findings related to peak forward acceleration (Fig. 2A). However, differences were still detected between groups for peak medial–lateral acceleration 3 (Fig. 2D), suggesting that gait balance control differences in the frontal plane of movement may exist following concussion independent of gait velocity. Previous work reported obstacle navigation impairments while completing a Stroop task following concussion despite recovery of clinical symptoms (Fait et al., 2013). In a long-term retrospective analysis, altered gait patterns were reported to persist six years post-concussion (Martini et al., 2011). Using motion capture or instrumented walkways, researchers have observed effects following concussion on variables such as gait speed (Fait et al., 2013; Martini et al., 2011), gait termination strategies (Buckley et al., 2013), and COM displacement/velocity during gait (Howell et al., 2014, 2013). However, no studies have reported how whole-body acceleration during gait may be affected by concussion. One advantage of monitoring acceleration during movement is that it may provide further insights into how balance is controlled, and lead to a better understanding of the mechanisms underlying gait imbalance (Fujimoto and Chou, 2012). The finding that the peak medial–lateral acceleration 3 variable detected significant between-group differences throughout the two months of post-injury testing (Fig. 2D), and could effectively discriminate concussion participants from controls within 72 h and at two weeks post-injury using a ROC analysis, suggests that measuring acceleration may be a promising way to objectively track balance recovery following concussion and warrants further investigation into its clinical feasibility. The importance of gait and balance testing in the clinical examination of concussion has been recently documented (Harmon et al., 2013; McCrory et al., 2013). As such, accelerometry provides a readily available technology that can be straightforwardly employed in a clinical setting to facilitate the examination of motor and cognitive functions following concussion to potentially assist in clinical management. The current study was limited by uneven subject numbers and non-individually matched subject groups. However, the demographics of the two groups were not significantly different (Table 1), potentially mitigating the effects of this limitation. As this study enrolled a relatively small control sample, clinical interpretation of cutoff thresholds indicating abnormal performance should be done with caution. The sensitivity of the acceleration data to correctly classify participants with concussion decreased across the testing timeline, thus future studies should seek to establish normative reference ranges in cohorts of healthy
Please cite this article as: Howell, D., et al., Monitoring recovery of gait balance control following concussion using an accelerometer. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.06.014i
D. Howell et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
individuals to provide a robust comparison to patients with concussion.
Conflict of interest statement The authors have no conflicts of interest which may have influenced the work reported in this study.
Acknowledgments This work was supported by Veterans Administration subcontract awards [A4842C8 & A4843C], Translational Research Award from the University of Oregon and Peace Health Oregon Region, the Department of Defense-TATRC Award (W81XWH-111-0717), and the Eugene and Clarissa Evonuk Memorial Graduate Fellowship in Environmental, Cardiovascular, or Stress Physiology. The authors would like to acknowledge Mr. Quinn Peterson and Ms. Madison Murray for their assistance with data collection.
References Brown, H.J., Siegmund, G.P., Guskiewicz, K.M., Van Den Doel, K., Cretu, E., Blouin, J.S., 2014. Development and validation of an objective balance error scoring system. Med. Sci. Sports Exerc. 46, 1610–1616. Buckley, T.A., Munkasy, B.A., Tapia-Lovler, T.G., Wikstrom, E.A., 2013. Altered gait termination strategies following a concussion. Gait Posture 38, 549–551. Bugané, F., Benedetti, M.G., Casadio, G., Attala, S., Biagi, F., Manca, M., Leardini, A., 2012. Estimation of spatial-temporal gait parameters in level walking based on a single accelerometer: validation on normal subjects by standard gait analysis. Comput. Methods Programs Biomed. 108, 129–137. Doi, S.A.R., Williams, G.M. (Eds.), 2013. Methods of Clinical Epidemiology, Springer Series on Epidemiology and Public Health. Springer, Berlin, Heidelberg, p. 69. Fait, P., Swaine, B., Cantin, J.-F., Leblond, J., McFadyen, B.J., 2013. Altered integrated locomotor and cognitive function in elite athletes 30 days postconcussion: a preliminary study. J. Head Trauma Rehabil. 28, 293–301. Fan, J., Upadhye, S., Worster, A., 2006. Understanding receiver operating characteristic (ROC) curves. CJEM 8, 19–20. Fujimoto, M., Chou, L.-S., 2012. Dynamic balance control during sit-to-stand movement: an examination with the center of mass acceleration. J. Biomech. 45, 543–548. Fujimoto, M., Chou, L.-S., 2014. Region of stability derived by center of mass acceleration better identifies individuals with difficulty in sit-to-stand movement. Ann. Biomed. Eng. 42, 733–741. Furman, G.R., Lin, C.-C., Bellanca, J.L., Marchetti, G.F., Collins, M.W., Whitney, S.L., 2013. Comparison of the balance accelerometer measure and balance error
5
scoring system in adolescent concussions in sports. Am. J. Sports Med. 41, 1404–1410. Hahn, M.E., Chou, L.-S., 2004. Age-related reduction in sagittal plane center of mass motion during obstacle crossing. J. Biomech. 37, 837–844. Harmon, K.G., Drezner, J.A., Gammons, M., Guskiewicz, K.M., Halstead, M., Herring, S.A., Kutcher, J.S., Pana, A., Putukian, M., Roberts, W.O., 2013. American Medical Society for Sports Medicine position statement: concussion in sport. Br. J. Sports Med. 47, 15–26. Howell, D.R., Osternig, L.R., Chou, L.-S., 2013. Dual-task effect on gait balance control in adolescents with concussion. Arch. Phys. Med. Rehabil. 94, 1513–1520. Howell, D.R., Osternig, L.R., Chou, L.-S., 2015a. Adolescents demonstrate greater gait balance control deficits after concussion than young adults. Am. J. Sports Med. 43, 625–632. Howell, D.R., Osternig, L.R., Chou, L.-S., 2015b. Return to activity after concussion affects dual-task gait balance control recovery. Med. Sci. Sports Exerc. 47, 673–680. Howell, D.R., Osternig, L.R., Koester, M.C., Chou, L.-S., 2014. The effect of cognitive task complexity on gait stability in adolescents following concussion. Exp. Brain Res. 232, 1773–1782. King, L.A., Horak, F.B., Mancini, M., Pierce, D., Priest, K.C., Chesnutt, J., Sullivan, P., Chapman, J.C., 2014. Instrumenting the balance error scoring system for use with patients reporting persistent balance problems after mild traumatic brain injury. Arch. Phys. Med. Rehabil. 95, 353–359. Lee, H., Sullivan, S.J., Schneiders, A.G., 2013. The use of the dual-task paradigm in detecting gait performance deficits following a sports-related concussion: a systematic review and meta-analysis. J. Sci. Med. Sport 16, 2–7. Martini, D.N., Sabin, M.J., DePesa, S.A., Leal, E.W., Negrete, T.N., Sosnoff, J.J., Broglio, S.P., 2011. The chronic effects of concussion on gait. Arch. Phys. Med. Rehabil. 92, 585–589. McCrory, P., Meeuwisse, W.H., Aubry, M., Cantu, R.C., Dvořák, J., Echemendia, R.J., Engebretsen, L., Johnston, K., Kutcher, J.S., Raftery, M., Sills, A., Benson, B.W., Davis, G.A., Ellenbogen, R., Guskiewicz, K.M., Herring, S.A., Iverson, G.L., Jordan, B.D., Kissick, J., McCrea, M., McIntosh, A.S., Maddocks, D., Makdissi, M., Purcell, L., Putukian, M., Schneider, K., Tator, C.H., Turner, M., 2013. Consensus statement on concussion in sport: the 4th international conference on concussion in sport, Zurich, November 2012. J. Athl. Train. 48, 554–575. Parker, T.M., Osternig, L.R., Lee, H.-J., Donkelaar, P., van, Chou, L.-S., 2005. The effect of divided attention on gait stability following concussion. Clin. Biomech. 20, 389–395. Parker, T.M., Osternig, L.R., Van Donkelaar, P., Chou, L.-S., 2006. Gait stability following concussion. Med. Sci. Sports Exerc. 38, 1032–1040. Perry, J., Burnfield, J.M., 1992. Gait Analysis: Normal and Pathological Function. Slack Inc., Thorofare, NJ, pp. 10–16. Register-Mihalik, J.K., Guskiewicz, K.M., Mihalik, J.P., Schmidt, J.D., Kerr, Z.Y., McCrea, M.A., 2013a. Reliable change, sensitivity, and specificity of a multidimensional concussion assessment battery. J. Head Trauma Rehabil. 28, 274–283. Register-Mihalik, J.K., Littleton, A.C., Guskiewicz, K.M., 2013b. Are divided attention tasks useful in the assessment and management of sport-related concussion? Neuropsychol. Rev. 23, 300–313. Zijlstra, W., Hof, A.L., 2003. Assessment of spatio-temporal gait parameters from trunk accelerations during human walking. Gait Posture 18, 1–10.
Please cite this article as: Howell, D., et al., Monitoring recovery of gait balance control following concussion using an accelerometer. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.06.014i
