Personalized Template-Based Step Detection From Inertial Measurement Units Signals in Multiple Sclerosis

Aliénor Vienne-Jumeau, Laurent Oudre, Albane Moreau, Flavien Quijoux, Sébastien Edmond, Mélanie Dandrieux, Eva Legendre, Pierre Paul Vidal, Damien Ricard, Aliénor Vienne-Jumeau, Laurent Oudre, Albane Moreau, Flavien Quijoux, Sébastien Edmond, Mélanie Dandrieux, Eva Legendre, Pierre Paul Vidal, Damien Ricard

Abstract

Background: Objective gait assessment is key for the follow-up of patients with progressive multiple sclerosis (pMS). Inertial measurement units (IMUs) provide reliable and yet easy quantitative gait assessment in routine clinical settings. However, to the best of our knowledge, no automated step-detection algorithm performs well in detecting severely altered pMS gait. Method: This article elaborates on a step-detection method based on personalized templates tested against a gold standard. Twenty-two individuals with pMS and 10 young healthy subjects (HSs) were instructed to walk on an electronic walkway wearing synchronized IMUs. Templates were derived from the IMU signals by using Initial and Final Contact times given by the walkway. These were used to detect steps from other gait trials of the same individual (intra-individual template-based detection, IITD) or another participant from the same group (pMS or HS) (intra-group template-based detection, IGTD). All participants were seen twice with a 6-month interval, with two measurements performed at each visit. Performance and accuracy metrics were computed, along with a similarity index (SId), which was computed as the mean distance between detected steps and their respective closest template. Results: For HS participants, both the IITD and the IGTD algorithms had precision and recall of 1.00 for detecting steps. For pMS participants, precision and recall ranged from 0.94 to 1.00 for IITD and 0.85 to 0.95 for IGTD depending on the level of disability. The SId was correlated with performance and the accuracy of the result. An SId threshold of 0.957 (IITD) and 0.963 (IGTD) could rule out decreased performance (F-measure ≤ 0.95), with negative predictive values of 0.99 and 0.96 with the IITD and IGTD algorithms. Also, the SId computed with the IITD and IGTD algorithms could distinguish individuals showing changes at 6-month follow-up. Conclusion: This personalized step-detection method has high performance for detecting steps in pMS individuals with severely altered gait. The algorithm can be self-evaluating with the SI, which gives a measure of the confidence the clinician can have in the detection. What is more, the SId can be used as a biomarker of change in disease severity occurring between the two measurement times.

Keywords: accelerometer; gait detection; gait disorders; gait quantification; inertial measurement unit; multiple sclerosis; wearable inertial sensors.

Copyright © 2020 Vienne-Jumeau, Oudre, Moreau, Quijoux, Edmond, Dandrieux, Legendre, Vidal and Ricard.

Figures

Figure 1
Figure 1
Definition of the different pairs of reference/detection trials analyzed by the intra-individual template-based detection (IITD) algorithm (A) and the intra-group template-based detection (IGTD) algorithm (B).
Figure 2
Figure 2
Example traces of the results of the step detection method for a pMS individual. The lines represent medio-lateral axis angular velocity (upper panel) and the magnitude of the norm of the acceleration (lower panel) recorded from the right (blue) and left (red) feet. The vertical lines display the Initial and Final Contacts as defined by the GR, and the triangles and circles display the ICs (triangles) and FCs (circles) as detected by our method. The shaded zone delimits the U-turn and is excluded from the analysis.
Figure 3
Figure 3
Receiver operating characteristic (ROC) curves for (A) the training cohort involving 15 people with pMS and (B) the test cohort involving seven people with pMS for the IITD detection method. Cutoffs were determined with the training cohort, and their predictive values were computed within the test cohort. Dashed curves are ROC curves for each configuration of the Monte Carlo cross-validation (and the nested four-fold cross-validation for the training set). Plain curves are the means of all dashed curves.
Figure 4
Figure 4
Receiver operating characteristic (ROC) curves for (A) the training cohort involving 15 people with pMS and (B) the test cohort involving seven people with pMS for the IGTD detection method. Cutoffs were determined with the training cohort, and their predictive values were computed within the test cohort. Dashed curves are ROC curves for each configuration of the Monte Carlo cross-validation (and the nested four-fold cross-validation for the training set). Plain curves are the means of all dashed curves.
Figure 5
Figure 5
SId from the IITD and IGTD detection methods depending on the change in disease status. The EDSS was not available for one patient.

References

    1. Heesen C, Bohm J, Reich C, Kasper J, Goebel M, Gold S. Patient perception of bodily functions in multiple sclerosis: gait and visual function are the most valuable. Mult Scler J. (2008) 14:988–91. 10.1177/1352458508088916
    1. Fischer JS, Rudick RA, Cutter GR, Reingold SC. The multiple sclerosis functional composite measure (MSFC): an integrated approach to MS clinical outcome assessment. Mult Scler J. (1999) 5:244–50. 10.1191/135245899678846168
    1. Goodman AD, Brown TR, Cohen JA, Krupp LB, Schapiro R, Schwid SR, et al. . Dose comparison trial of sustained-release fampridine in multiple sclerosis. Neurology. (2008) 71:1134–41. 10.1212/01.wnl.0000326213.89576.0e
    1. Kragt JJ, van der Linden FaH, Nielsen JM, Uitdehaag BMJ, Polman CH. Clinical impact of 20% worsening on timed 25-foot walk and 9-hole PEG Test in multiple sclerosis. Mult Scler. (2006) 12:594–8. 10.1177/1352458506070768
    1. Greene BR, Doheny EP, Walsh C, Cunningham C, Crosby L, Kenny RA. Evaluation of falls risk in community-dwelling older adults using body-worn sensors. Gerontology. (2012) 58:472–80. 10.1159/000337259
    1. Marschollek M, Rehwald A, Wolf KH, Gietzelt M, Nemitz G, zu Schwabedissen HM, et al. . Sensors vs. experts - a performance comparison of sensor-based fall risk assessment vs. conventional assessment in a sample of geriatric patients. BMC Med Inform Decis Mak. (2011) 11:48. 10.1186/1472-6947-11-48
    1. Vaney C, Gattlen B, Lugon-Moulin V, Meichtry A, Hausammann R, Foinant D, et al. Robotic-assisted step training (Lokomat) not superior to equal intensity of over-ground rehabilitation in patients with multiple sclerosis. Neurorehabil Neural Repair. (2012) 26:212–21. 10.1177/1545968311425923
    1. Vienne A, Barrois RP, Buffat S, Ricard D, Vidal PP. Inertial sensors to assess gait quality in patients with neurological disorders: a systematic review of technical and analytical challenges. Front Psychol. (2017) 8:817. 10.3389/fpsyg.2017.00817
    1. Senden R, Savelberg HHCM, Grimm B, Heyligers IC, Meijer K. Accelerometry-based gait analysis, an additional objective approach to screen subjects at risk for falling. Gait Post. (2012) 36:296–300. 10.1016/j.gaitpost.2012.03.015
    1. Barrois R, Gregory T, Oudre L, Moreau T, Truong C, Aram Pulini A, et al. . An automated recording method in clinical consultation to rate the limp in lower limb osteoarthritis. PLoS ONE. (2016) 11:e0164975. 10.1371/journal.pone.0164975
    1. Truong C, Barrois-Müller R, Moreau T, Provost C, Vienne-Jumeau A, Moreau A, et al. A data set for the study of human locomotion with inertial measurements units. Image Process On Line. (2019) 9:381–90. 10.5201/ipol.2019.265
    1. Panebianco GP, Bisi MC, Stagni R, Fantozzi S. Analysis of the performance of 17 algorithms from a systematic review: influence of sensor position, analysed variable and computational approach in gait timing estimation from IMU measurements. Gait Post. (2018) 66:76–82. 10.1016/j.gaitpost.2018.08.025
    1. Storm FA, Nair KPS, Clarke AJ, Van der Meulen JM, Mazza C. Free-living and laboratory gait characteristics in patients with multiple sclerosis. PLoS ONE. (2018) 13:e0196463. 10.1371/journal.pone.0196463
    1. Comber L, Galvin R, Coote S. Gait deficits in people with multiple sclerosis: a systematic review and meta-analysis. Gait Post. (2017) 51:25–35. 10.1016/j.gaitpost.2016.09.026
    1. Bertoli M, Cereatti A, Trojaniello D, Avanzino L, Pelosin E, Del Din S, et al. . Estimation of spatio-temporal parameters of gait from magneto-inertial measurement units: multicenter validation among Parkinson, mildly cognitively impaired and healthy older adults. Biomed Eng Online. (2018) 17:58. 10.1186/s12938-018-0488-2
    1. Dijkstra B, Kamsma YP, Zijlstra W. Detection of gait and postures using a miniaturized triaxial accelerometer-based system: accuracy in patients with mild to moderate Parkinson's disease. Arch Phys Med Rehabil. (2010) 91:1272–7. 10.1016/j.apmr.2010.05.004
    1. Shin SH, Park CG. Adaptive step length estimation algorithm using optimal parameters and movement status awareness. Med Eng Phys. (2011) 33:1064–71. 10.1016/j.medengphy.2011.04.009
    1. Novak D, Rebersek P, De Rossi SMM, Donati M, Podobnik J, Beravs T, et al. . Automated detection of gait initiation and termination using wearable sensors. Med Eng Phys. (2013) 35:1713–20. 10.1016/j.medengphy.2013.07.003
    1. Marschollek M, Goevercin M, Wolf KH, Song B, Gietzelt M, Haux R, et al. . A performance comparison of accelerometry-based step detection algorithms on a large, non-laboratory sample of healthy and mobility-impaired persons. In: Conference proceedings: Annual International Conference of the IEEE Engineering in Medicine and Biology Society IEEE Engineering in Medicine and Biology Society Annual Conference (Vancouver, BC: ) (2008). p. 1319–22. 10.1109/IEMBS.2008.4649407
    1. Trojaniello D, Ravaschio A, Hausdorff JM, Cereatti A. Comparative assessment of different methods for the estimation of gait temporal parameters using a single inertial sensor: application to elderly, post-stroke, Parkinson's disease and Huntington's disease subjects. Gait Post. (2015) 42:310–6. 10.1016/j.gaitpost.2015.06.008
    1. Prakash C, Kumar R, Mittal N. Recent developments in human gait research: parameters, approaches, applications, machine learning techniques, datasets and challenges. Artif Intell Rev. (2018) 49:1–40. 10.1007/s10462-016-9514-6
    1. Mantilla J, Oudre L, Barrois R, Vienne A, Ricard D. Template-DTW based on inertial signals: Preliminary results for step characterization. In: 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) (Jeju Island: ) (2017). p. 2267–70. 10.1109/EMBC.2017.8037307
    1. Barth J, Oberndorfer C, Pasluosta C, Schülein S, Gassner H, Reinfelder S, et al. . Stride segmentation during free walk movements using multi-dimensional subsequence dynamic time warping on inertial sensor data. Sensors. (2015) 15:6419–40. 10.3390/s150306419
    1. Barth J, Oberndorfer C, Kugler P, Schuldhaus D, Winkler J, Klucken J, et al. . Subsequence dynamic time warping as a method for robust step segmentation using gyroscope signals of daily life activities. In: 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) (Osaka: IEEE; ) (2013). p. 6744–7. 10.1109/EMBC.2013.6611104
    1. Oudre L, Barrois-Muller R, Moreau T, Truong C, Vienne-Jumeau A, Ricard D, et al. . Template-based step detection with inertial measurement units. Sensors. (2018) 18:4033. 10.3390/s18114033
    1. Soaz C, Diepold K. Step detection and parameterization for gait assessment using a single waist-worn accelerometer. IEEE Trans Biomed Eng. (2016) 63:933–42. 10.1109/TBME.2015.2480296
    1. Trojaniello D, Cereatti A, Pelosin E, Avanzino L, Mirelman A, Hausdorff JM, et al. . Estimation of step-by-step spatio-temporal parameters of normal and impaired gait using shank-mounted magneto-inertial sensors: application to elderly, hemiparetic, Parkinsonian and choreic gait. J Neuroeng Rehabil. (2014) 11:152. 10.1186/1743-0003-11-152
    1. Vienne-Jumeau A, Oudre L, Moreau A, Quijoux F, Vidal PP, Ricard D. Comparing gait trials with greedy template matching. Sensors. (2019) 19:3089. 10.3390/s19143089
    1. Polman CH, Reingold SC, Banwell B, Clanet M, Cohen JA, Filippi M, et al. . Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. Ann Neurol. (2011) 69:292–302. 10.1002/ana.22366
    1. Zecevic AA, Salmoni AW, Speechley M, Vandervoort AA. Defining a fall and reasons for falling: comparisons among the views of seniors, health care providers, and the research literature. Gerontologist. (2006) 46:367–76. 10.1093/geront/46.3.367
    1. Huxham F, Gong J, Baker R, Morris M, Iansek R. Defining spatial parameters for non-linear walking. Gait Post. (2006) 23:159–63. 10.1016/j.gaitpost.2005.01.001
    1. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology. (1983) 33:1444–52. 10.1212/WNL.33.11.1444
    1. Fisk JD, Pontefract A, Ritvo PG, Archibald CJ, Murray TJ. The impact of fatigue on patients with multiple sclerosis. Can J Neurol Sci. (1994) 21:9–14. 10.1017/S0317167100048691
    1. Fisk JD, Ritvo PG, Ross L, Haase DA, Marrie TJ, Schlech WF. Measuring the functional impact of fatigue: initial validation of the fatigue impact scale. Clin Infect Dis. (1994) 18(Suppl. 1):S79–83. 10.1093/clinids/18.Supplement_1.S79
    1. Learmonth YC, Dlugonski DD, Pilutti LA, Sandroff BM, Motl RW. The reliability, precision and clinically meaningful change of walking assessments in multiple sclerosis. Mult Scler J. (2013) 19:1784–91. 10.1177/1352458513483890
    1. Nordin A, Taft C, Lundgren-Nilsson A, Dencker A. Minimal important differences for fatigue patient reported outcome measures-a systematic review. BMC Med Res Methodol. (2016) 16:62. 10.1186/s12874-016-0167-6
    1. Hastie T, Tibshirani R, Friedman J. Model assessment and selection. In: Hastie T, Tibshirani R, Friedman J, editors. The Elements of Statistical Learning: Data Mining, Inference, and Prediction. New York, NY: Springer; (2009). p. 219–59. 10.1007/978-0-387-84858-7_7
    1. Youden WJ. Index for rating diagnostic tests. Cancer. (1950) 3:32–5. 10.1002/1097-0142(1950)3:1<2::AID-CNCR2820030106>;2-3
    1. Perkins NJ, Schisterman EF. The inconsistency of “optimal” cut-points using two ROC based criteria. Am J Epidemiol. (2006) 163:670–5. 10.1093/aje/kwj063
    1. Jasiewicz JM, Allum JHJ, Middleton JW, Barriskill A, Condie P, Purcell B, et al. . Gait event detection using linear accelerometers or angular velocity transducers in able-bodied and spinal-cord injured individuals. Gait Post. (2006) 24:502–9. 10.1016/j.gaitpost.2005.12.017
    1. Mannini A, Genovese V, Maria Sabatini A. Online decoding of hidden Markov models for gait event detection using foot-mounted gyroscopes. IEEE J Biomed Health Inform. (2014) 18:1122–30. 10.1109/JBHI.2013.2293887
    1. Lee H, Kim J, Myoung H, Lee J, Lee K. Repeatability of the accelerometric-based method to detect step events for hemiparetic stroke patients. In: 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society (Boston, MA: ) (2011). p. 5195–8.
    1. Chia Bejarano N, Ambrosini E, Pedrocchi A, Ferrigno G, Monticone M, Ferrante S. A novel adaptive, real-time algorithm to detect gait events from wearable sensors. IEEE Trans Neural Syst Rehabil Eng. (2015) 23:413–22. 10.1109/TNSRE.2014.2337914
    1. Gonzalez RC, Lopez AM, Rodriguez-Uria J, Alvarez D, Alvarez JC. Real-time gait event detection for normal subjects from lower trunk accelerations. Gait Post. (2010) 31:322–5. 10.1016/j.gaitpost.2009.11.014
    1. Teufl W, Lorenz M, Miezal M, Taetz B, Frohlich M, Bleser G. Towards inertial sensor based mobile gait analysis: event-detection and spatio-temporal parameters. Sensors. (2018) 19:38. 10.3390/s19010038
    1. Pau M, Caggiari S, Mura A, Corona F, Leban B, Coghe G, et al. . Clinical assessment of gait in individuals with multiple sclerosis using wearable inertial sensors: comparison with patient-based measure. Mult Scler Relat Disord. (2016) 10:187–91. 10.1016/j.msard.2016.10.007
    1. Leeflang MMG, Moons KGM, Reitsma JB, Zwinderman AH. Bias in sensitivity and specificity caused by data-driven selection of optimal cutoff values: mechanisms, magnitude, and solutions. Clin Chem. (2008) 54:729–37. 10.1373/clinchem.2007.096032

Source: PubMed

Sponsors and Collaborators

Medical Conditions

Drug Interventions

3
Subscribe