Академический Документы
Профессиональный Документы
Культура Документы
com/jmt
ISSN: 0309-1902 (print), 1464-522X (electronic)
J Med Eng Technol, 2013; 37(8): 498510
! 2013 Informa UK Ltd. DOI: 10.3109/03091902.2013.837529
INNOVATION
Faculty of Science, Universiti Brunei Darussalam, Jalan Tungku Link, Gadong BE1410, Brunei Darussalam and 2Sports Medicine & Research Centre,
Hassan Bolkiah National Stadium, Berakas, Brunei Darussalam
Abstract
Keywords
A hardware/software co-design for assessing post-Anterior Cruciate Ligament (ACL) reconstruction ambulation is presented. The knee kinematics and neuromuscular data during
walking (26 km h1) have been acquired using wireless wearable motion and electromyography (EMG) sensors, respectively. These signals were integrated by superimposition and mixed
signals processing techniques in order to provide visual analyses of bio-signals and
identification of the recovery progress of subjects. Monitoring overlapped signals simultaneously helps in detecting variability and correlation of knee joint dynamics and muscles
activities for an individual subject as well as for a group. The recovery stages of subjects have
been identified based on combined features (knee flexion/extension and EMG signals) using an
adaptive neuro-fuzzy inference system (ANFIS). The proposed system has been validated for
28 test subjects (healthy and ACL-reconstructed). Results of ANFIS showed that the ambulation
data can be used to distinguish subjects at different levels of recuperation after ACL
reconstruction.
1. Introduction
Anterior Cruciate Ligament (ACL) rupture leads to various
short- and long-term impacts on the injured subjects including
dynamic knee joint instability, neuromuscular impairments,
loss of proprioception, cartilage degeneration and early onset
of osteoarthritis [14]. In order to restore the knee kinematics,
kinetics and neurophysiologic dysfunction, ACL reconstruction is generally recommended. However, recent clinical
studies indicate the occurrence of cartilage degeneration
and early progression of osteoarthritis in subjects having
ACL reconstruction, which suggests incomplete restoration
of normal kinematics and neuromuscular function after ACL
surgery and a lack of proper monitoring of rehabilitation
progress [510].
The alterations in biomechanical features of the knee
during walking following ACL reconstruction have been
evaluated in different studies. Reduced knee flexion angles/
moments at mid-stance, extension deficit during swing phase,
gait variability and muscles weakness have been noticed
in ACL reconstructed (ACL-R) subjects even after 12 years
of surgery [1015]. Thus, in order to re-establish knee joint
and neuromuscular control and avoiding or minimizing postoperative complications, a careful monitoring of recovery
process after ACL surgery becomes crucial.
History
Received 8 February 2013
Revised 5 August 2013
Accepted 20 August 2013
20
13
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
S. M. N. Arosha Senanayake*1, Owais Ahmed Malik1, Pg. Mohammad Iskandar1, and Dansih Zaheer2
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
DOI: 10.3109/03091902.2013.837529
499
2. Methodology
2.1. System architecture
The hardware/software co-design of the proposed system is
shown in Figure 1. The recovery assessment system consisted
of two major hardware components which are (1) KinetiSense
(ClevMed. Inc, Cleveland, OH) with four wireless microelectro-mechanical motion sensor units and (2) a wireless
BioCapture (ClevMed. Inc) surface electromyography monitoring unit with electrodes. The body-mounted motion sensors
were used to measure the subjects lower extremity motion
during walking in terms of angular rate and linear acceleration. These angular rates and linear acceleration were then
used to compute the knee joint movements in the sagittal
plane. The electromyography sensors provided the muscle
Figure 1. Hardware/software co-design with information flow from wireless wearable sensors to software components for providing recovery
classification and visual biofeedback analysis for post-ACL reconstruction recovery.
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
500
S. M. N. A. Senanayake et al.
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
DOI: 10.3109/03091902.2013.837529
501
Figure 2. Placement of motion and EMG sensors on both lower limbs of a human body to collect data for knee joint and neuromuscular movements.
(a) Front view, (b) back view.
502
S. M. N. A. Senanayake et al.
70
60
50
40
30
20
10
0
10
20
3.4
3.5
3.6
3.7
3.8
3.9
4.1
4.2
Figure 3. Identification of heel strike during ambulation with knee angle changes (---) using shank angular rate () from motion sensors and anteriorposterior acceleration ( ___ ) from BioCapture system to synchronize kinematics and EMG data for signals integration and overlapping.
80
70
50
40
30
0.02
20
0.04
60
Knee Flexion/Extension (degrees)
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
4.3
x 104
Sampling Rate
10
10
15
16
17
18
19
20
21
22
23
24
0
25
Time (sec)
Figure 4. Superimposition of knee flexion/extension ( ___ ) and vastus lateralis EMG envelop ( ) for a healthy subject during walking at a speed
of 3 km h1 to represent the correlation between both signals.
existing literature [45]. MSt was further divided into two halves
to identify the joints and muscles activity in more detail. For
kinematics data, the average knee flexion/extension during
each selected phase was calculated. For EMG data classification, in previous studies, various kinds of features have been
used in time and frequency domains such as mean absolute
value, spectral analysis, zero-crossing and auto-regressive
coefficients, etc. In this study, a combination of time, frequency
and time-frequency EMG features has been used. The root
mean square (RMS) value for filtered EMG data was used as a
time-domain feature for each selected gait phase (1).
s
Z T2
1
EMGRMS
EMGt2 dt
1
T2 T1 T1
For time-frequency analysis, CWT has been chosen as it
has been proved effective in classification of non-stationary
DOI: 10.3109/03091902.2013.837529
(a)
503
70
60
50
40
30
20
10
0
10
10
20
30
40
50
60
70
80
90
100
60
70
80
90
100
(b)
70
60
50
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
40
30
20
10
0
10
10
20
30
40
50
Gait Cycle (%)
Figure 5. Knee flexion/extension variations in the (a) ACL-R and (b) ACL-I leg of a subject at 3 km h1 speed after 2 months of surgery ( ___ )
average, (---) lower limit, (- - -) upper limit representing restricted extension during 2050% of the gait cycle.
3. Results
3.1. Biofeedback: Visual analyses
3.1.1. Intra- and inter-subjects variations in knee
flexion/extension
Intra- and inter-subjects variations in knee flexion/extension
of both legs for a selected number of gait cycles at different
504
S. M. N. A. Senanayake et al.
(a)
70
60
50
40
30
20
10
0
10
10
20
30
40
50
60
70
80
90
100
60
70
80
90
100
(b)
70
60
50
40
30
20
10
0
10
10
20
30
40
50
Gait Cycle (%)
Figure 6. Knee flexion/extension variations in the (a) ACL-R and (b) ACL-I leg of a subject at 4 km h1 speed after 2 months of surgery ( ___ )
average, (---) lower limit, (- - -) upper limit representing restricted extension during 2050% of the gait cycle.
70
60
50
Knee Angle (degrees)
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
40
30
20
10
0
10
10
20
30
40
50
60
70
80
90
100
Figure 7. Knee flexion/extension variations in the subjects at 4 km h1 speed of a healthy subject ( ___ ) average, after 1 year of surgery (- - -),
after 2 months of surgery (---), representing the visible differences during 2050% of gait cycle and maximum peak flexion values/timings.
DOI: 10.3109/03091902.2013.837529
505
800
700
% of Mean EMG
600
500
400
300
200
100
0
3.5
4.5
5.5
6.5
7.5
Figure 8. Comparison of muscle percentage of mean strength and activation for vastus lateralis ( ) and vastus medialis () in the ACL-R leg
of a subject after 1 year of surgery at 4 km h1 speed representing weak vastus medialis.
300
% of Mean EMG
250
200
150
100
50
3.5
4.5
5.5
x 104
Sampling Rate
Figure 9. Comparison of percentage of mean strength and activation for vastus medialis muscles of ACL-I leg ( ) and ACL-R leg ( ___ ) of a
subject after 1 year of surgery at 4 km h1 speed, representing lower mean strength for vastus medialis in ACL-R leg.
1000
900
800
% of Mean EMG
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
x 104
Sampling Rate
700
600
500
400
300
200
100
0
15
15.5
16
16.5
17
17.5
18
18.5
19
19.5
20
Time (sec)
Figure 10. Comparison of percentage of mean strength and activation for biceps femoris muscles of ACL-I leg ( ) and ACL-R leg ( ___ ) of a
subject after 1 year of surgery at 5 km h1 speed, representing more or less similar strength for muscles in both legs.
506
S. M. N. A. Senanayake et al.
(a)
60
300
40
250
30
200
20
150
10
% of Mean EMG
50
100
50
10
25
25.5
26
26.5
27
27.5
Time (sec)
40
300
200
20
% of Mean EMG
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
(b) 60
100
0
24.5
25
25.5
26
26.5
27
100
27.5
Time (sec)
Figure 11. Activation timings and percentage of mean strengths of vastus lateralis ( ) of (a) ACL-I and (b) ACL-R leg vs knee flexion/extension
( ___ ) of a subject after 2 months of surgery showing (a) visible and (b) low contraction of vastus lateralis during terminal swing, load response and
mid stance phases affecting the knee flexion/extension movements.
DOI: 10.3109/03091902.2013.837529
70
1000
60
50
40
500
30
20
% of Mean EMG
(a)
507
10
0
10
15
15.5
16
16.5
17
17.5
18
18.5
19
0
19.5
Time (sec)
60
400
50
40
30
20
200
% of Mean EMG
70
10
0
10
17.5
18
18.5
19
19.5
20
20.5
0
21
Time (sec)
(c) 70
60
400
50
40
30
20
200
% of Mean EMG
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
(b)
10
0
18
18.5
19
19.5
20
20.5
21
Time (sec)
Figure 12. Activation timings and strengths of (a) biceps femoris, (b) semitendinosus and (c) vastus lateralis ( ) vs knee flexion/extension (____) of
a healthy subject at a speed of 3 km h1.
4. Discussion
The use of wireless body mounted sensors and the integration
of knee kinematics and neuromuscular signals have been
found helpful in assessing the recovery progress of ACL
reconstructed subjects. The monitoring of bio-signals and
their superimposition provide the co-ordination of knee
kinematics and muscles movements during ambulation. The
effect of each muscle, causing the variability of gait patterns,
can be detected for individual subjects using the proposed
system. Moreover, this study also suggests that the recovery
stage of subjects can be accurately identified based on the
508
S. M. N. A. Senanayake et al.
4
x 107
3.9
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
3.8
3.7
3.6
3.5
3.4
3.3
3.2
3.1
3
10
20
30
40
50
60
70
80
90
100
Epoch
Walking speed
2 km h1
3 km h1
4 km h1
5 km h1
6 km h1
Class
Group
Group
Group
Group
Group
Group
Group
Group
Group
Group
Group
Group
Group
Group
Group
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
Specificity
(%)
Sensitivity
(%)
F-measure
(%)
96.15%
88.46%
92.59%
100.00%
95.65%
95.45%
94.12%
95.45%
95.00%
93.75%
100.00%
92.86%
100.00%
100.00%
92.31%
80.00%
80.00%
77.78%
80.00%
100.00%
87.50%
80.00%
80.00%
85.71%
100.00%
83.33%
90.00%
100.00%
88.89%
100.00%
84.21%
76.20%
77.78%
88.89%
93.34%
87.50%
84.21%
80.00%
85.71%
94.11%
90.90%
90.00%
100.00%
94.11%
94.11%
DOI: 10.3109/03091902.2013.837529
3.
4.
5.
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
6.
7.
8.
9.
10.
11.
12.
Acknowledgement
This work is supported by the University Research Council
(URC) grant scheme at the Universiti Brunei Darussalam
under the grant No: UBD/PNC2/2/RG/1(195). with the title
Integrated Motion Analysis System (IMAS). The authors
appreciate the sports medicine centre at ministry of sports and
performance optimization centre at ministry of defence,
Brunei Darussalam for providing Brunei national athletes as
test subjects who had undergone the rehabilitation process
due to ACL surgeries as well as healthy test subjects involved
for non-invasive rehabilitation experiments. Further, the
authors also acknowledge the support provided by Mr
Illepurma Ranasinghe (head physical strength and conditioning) and Ms Maria Leah (physiotherapist) from the performance optimization centre at the Ministry of Defence.
Declaration of interest
13.
14.
15.
16.
17.
18.
References
1. Murrell, G.A., Maddali, S., Horovitz, L., Oakley, S.P., and Warren,
R.F., 2001, The effects of time course after anterior cruciate
ligament injury in correlation with meniscal and cartilage loss.
American Journal of Sports Medicine, 29, 914.
2. Reider, B., Arcand, M.A., Diehl, L.H., Mroczek, K., Abulencia, A.,
Stroud, C.C., Palm, M., Gilbertson, J., and Staszak, P., 2003,
19.
20.
509
J Med Eng Technol Downloaded from informahealthcare.com by University Brunei Darussalam on 10/30/13
For personal use only.
510
S. M. N. A. Senanayake et al.
21. Shaw, T., Chipchase, L.S., and Williams, M.T., 2004, A users guide
to outcome measurement following ACL reconstruction. Physical
Therapy in Sport, 5, 5767.
22. Akkaya, N., Ardic, F., Ozgen, M., Akkaya, S, Sahin, F., and Kilic,
A., 2012, Efficacy of electromyographic biofeedback and electrical
stimulation following arthroscopic partial meniscectomy: A
randomized controlled trial. Clinical Rehabilitation, 26, 224236.
23. Boucher, T., Wang, S., Trudelle-Jackson, E., and Olson, S., 2009,
Effectiveness of surface electromyographic biofeedback-triggered
neuromuscular electrical stimulation on knee rehabilitation.
North American Journal of Sports Physical Therapy, 4, 100109.
24. Feller, J., Hoser, C., and Webster, K., 2000, EMG biofeedback
assisted KT-1000 evaluation of anterior tibial displacement. Knee
Surgery, Sports Traumatology Arthroscopy, 8, 132136.
25. Scanlan, S.F., Chaudhari, A.M., Dyrby, C.O., and Andriacchi, T.P.,
2010, Differences in tibial rotation during walking in ACL
reconstructed and healthy contralateral knees. Journal of
Biomechanics, 43, 18171822.
26. Mayagoitia, R.E., Nene, A.V., and Veltink, P.H., 2002,
Accelerometer and rate gyroscope measurement of kinematics:
An inexpensive alternative to optical motion analysis systems.
Journal of Biomechanics, 35, 537542.
27. Gouwanda, D., and Senanayake, S.M.N.A., 2008, Emerging trends
of body-mounted sensors in sports and human gait analysis. In:
Abu Osman NA, Ibrahim F, Wan Abas WAB, Abdul Rahman HS,
Ting H-N, Magjarevic R, editors. IFMBE Proceedings (Berlin:
Springer), pp. 715718.
28. Atallah, L., Jones, G.G., Ali, R., Leong, J.J.H., Lo, B., and Yang,
G.-Z., 2011, Observing recovery from knee-replacement surgery by
using wearable sensors. Proceedings of the 2011 International
Conference on Body Sensor Networks: IEEE Computer Society,
2934.
29. Ayoade, M., Morton, L., and Baillie, L., 2011, Investigating
the feasibility of a wireless motion capture system to aid in the
rehabilitation of total knee replacement patients. Paper appeared in
5th International Conference on
Pervasive Computing
Technologies for Healthcare (PervasiveHealth), 2011 (Dublin,
Ireland), pp. 404407.
30. Saito, H., Watanabe, T., and Arifin, A., 2009, Ankle and knee joint
angle measurements during gait with wearable sensor system for
rehabilitation. IFMBE Proceedings, 25, 506509.
31. Hewett, T.E., Di Stasi, S.L., and Myer, G.D., 2013, Current
concepts for injury prevention in athletes after anterior cruciate
ligament reconstruction. American Journal of Sports Medicine, 41,
216224.
32. Leporace, G., Batista, L.A., Muniz, A.M., Zeitoune, G., Luciano,
T., and Metsavaht, L., Nadal, J., 2012, Classification of gait
kinematics of anterior cruciate ligament reconstructed subjects
using principal component analysis and regressions modelling.
Conference Proceedings IEEE Engineering in Medicine & Biology
Society, 2012, 65146517.
33. Toffola, L.D., Patel, S., Ozsecen, M.Y., Ramachandran, R., and
Bonato, P., 2012, A wearable system for long-term monitoring of
knee kinematics. International Conference on Biomedical and
Health Informatics (BHI), IEEE-EMBS (Shenzhen, China), 57
January 2012, pp. 188191.
34. Christanell, F., Hoser, C., Huber, R., Fink, C., and Luomajoki, H.,
2012, The influence of electromyographic biofeedback therapy on
knee extension following anterior cruciate ligament reconstruction:
A randomized controlled trial. Sports Medicine, Arthroscopy,
Rehabilitation, Therapy & Technology, 4, 41.
35. Senanayake, S.M.N.A., Malik, O.A., Iskandar, P.M., and Zaheer,
D., 2012, A hybrid intelligent system for recovery and performance
evaluation after anterior cruciate ligament injury. 12th International
Conference on Intelligent Systems Design and Applications (ISDA),
2012, IEEE. pp. 986991.
36. Englehart, K., Hudgins, B., Parker, P.A., and Stevenson, M., 1999,
Classification of the myoelectric signal using time-frequency based
representations. Medical Engineering & Physics, 21, 431438.
37. Kilby, J., Mawston, G., Gholam Hosseini, H., 2007, Continuous
wavelet transform analysis for the classification of surface electromyography signals. In: Magjarevic R, Nagel JHD, editors. World
Congress on Medical Physics and Biomedical Engineering 2006
SE 246. 14th edn. IFMBE Proceedings: Springer Berlin
Heidelberg DA 2007/01/01; 2007. p 10341036 LA English.
38. Khezri, M., and Jahed, M., 2011, A neuro-fuzzy inference
system for sEMG-based identification of hand motion commands.
IEEE Transactions on Industrial Electronics, 58, 19521960.
39. Kocer, S., 2010, Classification of EMG signals using neuro-fuzzy
system and diagnosis of neuromuscular diseases. Journal of
Medical Systems, 34, 321329.
40. Hermens, H.J., Freriks, B., Disselhorst-Klug, C., and Rau, G., 2000,
Development of recommendations for SEMG sensors and sensor
placement procedures. Journal of Electromyography Kinesiology,
10, 361374.
41. Madgwick, S., 2010, An efficient orientation filter for inertial and
inertial/magnetic sensor arrays. Bristol, UK: Department of
Mechanical Engineering, University of Bristol.
42. Merletti, R., 1999, Standards for reporting EMG data. Journal of
Electromyography and Kinesiology, 9, IIIIV.
43. Jasiewicz, J.M., Allum, J.H.J., Middleton, J.W., Barriskill, A.,
Condie, P., Purcell, B., and Li, R.C.T., 2006, Gait event detection
using linear accelerometers or angular velocity transducers in ablebodied and spinal-cord injured individuals. Gait & Posture, 24,
502509.
44. Evans, A.L., Duncan, G., and Gilchrist, W., 1991, Recording
accelerations in body movements. Medical & Biological
Engineering & Computing, 29, 102104.
45. Perry, J., 1992, Gait analysis: Normal and pathological function
(NJ, USA: SLACK Inc.).
46. Ismail, A.R., and Asfour, S.S., 1998, Continuous wavelet transform
application to EMG signals during human gait. Conference Record
of the Thirty-Second Asilomar Conference on Signals, Systems &
Computers, 1, 325329.
47. Kilby, J., and Hosseini, H.G., 2006, Extracting effective features
of SEMG using continuous wavelet transform. Conference
Proceedings IEEE Engineering in Medicine & Biology Society, 1,
17041707.
48. Jang, J.-S.R., 1993, ANFIS: Adaptive-network-based fuzzy inference system. IEEE Transactions on Systems, Man, and Cybernetics,
23, 665685.