Shotgun approaches to gait analysis: insights & limitations

R.G. Kaptein, D. Wezenberg, T. Ijmker, J.H.P. Houdijk, P.J. Beek, C.J.C. Lamoth, A. Daffertshofer

    Research output: Contribution to JournalArticleAcademicpeer-review

    Abstract

    Background: Identifying features for gait classification is a formidable problem. The number of candidate measures is legion. This calls for proper, objective criteria when ranking their relevance. Methods: Following a shotgun approach we determined a plenitude of kinematic and physiological gait measures and ranked their relevance using conventional analysis of variance (ANOVA) supplemented by logistic and partial least squares (PLS) regressions. We illustrated this approach using data from two studies involving stroke patients, amputees, and healthy controls. Results: Only a handful of measures turned out significant in the ANOVAs. The logistic regressions, by contrast, revealed various measures that clearly discriminated between experimental groups and conditions. The PLS regression also identified several discriminating measures, but they did not always agree with those of the logistic regression. Discussion & conclusion: Extracting a measure's classification capacity cannot solely rely on its statistical validity but typically requires proper post-hoc analysis. However, choosing the latter inevitably introduces some arbitrariness, which may affect outcome in general. We hence advocate the use of generic expert systems, possibly based on machine-learning.
    LanguageEnglish
    Article number120
    JournalJournal of NeuroEngineering and Rehabilitation
    Volume11
    DOIs
    Publication statusPublished - 2014

    Fingerprint

    Firearms
    Least-Squares Analysis
    Gait
    Analysis of Variance
    Logistic Models
    Expert Systems
    Amputees
    Biomechanical Phenomena
    Stroke
    Machine Learning

    Cite this

    @article{6dfdbb9c9d554654b7ee25c3402edfab,
    title = "Shotgun approaches to gait analysis: insights & limitations",
    abstract = "Background: Identifying features for gait classification is a formidable problem. The number of candidate measures is legion. This calls for proper, objective criteria when ranking their relevance. Methods: Following a shotgun approach we determined a plenitude of kinematic and physiological gait measures and ranked their relevance using conventional analysis of variance (ANOVA) supplemented by logistic and partial least squares (PLS) regressions. We illustrated this approach using data from two studies involving stroke patients, amputees, and healthy controls. Results: Only a handful of measures turned out significant in the ANOVAs. The logistic regressions, by contrast, revealed various measures that clearly discriminated between experimental groups and conditions. The PLS regression also identified several discriminating measures, but they did not always agree with those of the logistic regression. Discussion & conclusion: Extracting a measure's classification capacity cannot solely rely on its statistical validity but typically requires proper post-hoc analysis. However, choosing the latter inevitably introduces some arbitrariness, which may affect outcome in general. We hence advocate the use of generic expert systems, possibly based on machine-learning.",
    author = "R.G. Kaptein and D. Wezenberg and T. Ijmker and J.H.P. Houdijk and P.J. Beek and C.J.C. Lamoth and A. Daffertshofer",
    year = "2014",
    doi = "10.1186/1743-0003-11-120",
    language = "English",
    volume = "11",
    journal = "Journal of NeuroEngineering and Rehabilitation",
    issn = "1743-0003",
    publisher = "BioMed Central",

    }

    Shotgun approaches to gait analysis: insights & limitations. / Kaptein, R.G.; Wezenberg, D.; Ijmker, T.; Houdijk, J.H.P.; Beek, P.J.; Lamoth, C.J.C.; Daffertshofer, A.

    In: Journal of NeuroEngineering and Rehabilitation, Vol. 11, 120, 2014.

    Research output: Contribution to JournalArticleAcademicpeer-review

    TY - JOUR

    T1 - Shotgun approaches to gait analysis: insights & limitations

    AU - Kaptein, R.G.

    AU - Wezenberg, D.

    AU - Ijmker, T.

    AU - Houdijk, J.H.P.

    AU - Beek, P.J.

    AU - Lamoth, C.J.C.

    AU - Daffertshofer, A.

    PY - 2014

    Y1 - 2014

    N2 - Background: Identifying features for gait classification is a formidable problem. The number of candidate measures is legion. This calls for proper, objective criteria when ranking their relevance. Methods: Following a shotgun approach we determined a plenitude of kinematic and physiological gait measures and ranked their relevance using conventional analysis of variance (ANOVA) supplemented by logistic and partial least squares (PLS) regressions. We illustrated this approach using data from two studies involving stroke patients, amputees, and healthy controls. Results: Only a handful of measures turned out significant in the ANOVAs. The logistic regressions, by contrast, revealed various measures that clearly discriminated between experimental groups and conditions. The PLS regression also identified several discriminating measures, but they did not always agree with those of the logistic regression. Discussion & conclusion: Extracting a measure's classification capacity cannot solely rely on its statistical validity but typically requires proper post-hoc analysis. However, choosing the latter inevitably introduces some arbitrariness, which may affect outcome in general. We hence advocate the use of generic expert systems, possibly based on machine-learning.

    AB - Background: Identifying features for gait classification is a formidable problem. The number of candidate measures is legion. This calls for proper, objective criteria when ranking their relevance. Methods: Following a shotgun approach we determined a plenitude of kinematic and physiological gait measures and ranked their relevance using conventional analysis of variance (ANOVA) supplemented by logistic and partial least squares (PLS) regressions. We illustrated this approach using data from two studies involving stroke patients, amputees, and healthy controls. Results: Only a handful of measures turned out significant in the ANOVAs. The logistic regressions, by contrast, revealed various measures that clearly discriminated between experimental groups and conditions. The PLS regression also identified several discriminating measures, but they did not always agree with those of the logistic regression. Discussion & conclusion: Extracting a measure's classification capacity cannot solely rely on its statistical validity but typically requires proper post-hoc analysis. However, choosing the latter inevitably introduces some arbitrariness, which may affect outcome in general. We hence advocate the use of generic expert systems, possibly based on machine-learning.

    U2 - 10.1186/1743-0003-11-120

    DO - 10.1186/1743-0003-11-120

    M3 - Article

    VL - 11

    JO - Journal of NeuroEngineering and Rehabilitation

    T2 - Journal of NeuroEngineering and Rehabilitation

    JF - Journal of NeuroEngineering and Rehabilitation

    SN - 1743-0003

    M1 - 120

    ER -