Muscle activation patterns during gait: A hierarchical clustering analysis

• Published in: Biomedical signal processing and control Vol. 31; pp. 463 - 469
• Main Authors: Rosati, Samanta, Agostini, Valentina, Knaflitz, Marco, Balestra, Gabriella
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2017
• Subjects: Dendrogram; Gait analysis; Hierarchical clustering; Muscle activation patterns; Surface EMG; Hierarchical clustering; Dendrogram; Gait analysis; Muscle activation patterns; Surface EMG
• DOI: 10.1016/j.bspc.2016.09.017
• ISSN: 1746-8094

[Display omitted] •Muscle onset-offset timings during gait show a high stride-to-stride variability.•Different strides may show different timings even in the same activation modality.•A dendrogram algorithm was presented to cluster strides with similar timings.•The algorithm was validated on 5 lower limb muscles.•The concept of principal and secondary activations was introduced. Human gait is characterized by a large stride-to-stride variability of the muscle activation patterns (onset-offset timings). For this reason prolonged walking sessions lasting several minutes are analyzed. To interpret correctly the electromyographic (EMG) data collected during gait, it is important to group strides sharing similar EMG activation patterns. The aim of this work is to present and validate a method, based on hierarchical clustering, able to group strides showing homogeneous onset-offset activation intervals. Results show that the variability of the onset-offset timing is significantly reduced after clustering, for all of the five lower limb muscles considered to test this method. A by-product of the clustering procedure is the possibility to define and extract the principal activations of a muscle during gait. We define principal activations those activations that are necessary for the specific muscle contribution to the biomechanical function of walking. This concept may be useful whenever the dynamic performance of the muscle has to be compared in subsequent times, such as in patient’s follow-up or when the performance of a specific subject is to be compared to that of a group of selected subjects. The contribution presented in this work could be beneficial in implementing a personalized medicine approach to rehabilitation. Clinical gait analysis, enriched by hierarchical clustering of EMG patterns as well as by the quantitative assessment of muscles principal activations, could greatly contribute to the design of therapeutic treatments tailored on the patient’s needs.