Revealing the unique features of each individual’s muscle activation signatures

Abstract

There is growing evidence that each individual has unique movement patterns, or signatures. The exact origin of these movement signatures, however, remains unknown. We developed an approach that can identify individual muscle activation signatures during two locomotor tasks (walking and pedalling). A linear support vector machine was used to classify 78 participants based on their electromyographic (EMG) patterns measured on eight lower limb muscles. To provide insight into decision-making by the machine learning classification model, a layer-wise relevance propagation (LRP) approach was implemented. This enabled the model predictions to be decomposed into relevance scores for each individual input value. In other words, it provided information regarding which features of the time-varying EMG profiles were unique to each individual. Through extensive testing, we have shown that the LRP results, and by extent the activation signatures, are highly consistent between conditions and across days. In addition, they are minimally influenced by the dataset used to train the model. Additionally, we proposed a method for visualizing each individual's muscle activation signature, which has several potential clinical and scientific applications. This is the first study to provide conclusive evidence of the existence of individual muscle activation signatures.

Autoren

Jeroen Aeles
Fabian Horst
Sebastian Lapuschkin
Lilian Lacourpaille
François Hug

DOI

10.1098/rsif.2020.0770

eISSN

1742-5662

ISSN

1742-5689

Ausgabe der Veröffentlichung

174

Zeitschrift

Journal of The Royal Society Interface

Sprache

en

Online publication date

2021

Paginierung

20200770 - 20200770

Datum der Veröffentlichung

2021

Status

Published

Herausgeber

The Royal Society

Herausgeber URL

http://dx.doi.org/10.1098/rsif.2020.0770

Datum der Datenerfassung

2021

Titel

Revealing the unique features of each individual's muscle activation signatures

Ausgabe der Zeitschrift

18

Datenquelle: Crossref

Abstract

There is growing evidence that each individual has unique movement patterns, or signatures. The exact origin of these movement signatures, however, remains unknown. We developed an approach that can identify individual muscle activation signatures during two locomotor tasks (walking and pedalling). A linear support vector machine was used to classify 78 participants based on their electromyographic (EMG) patterns measured on eight lower limb muscles. To provide insight into decision-making by the machine learning classification model, a layer-wise relevance propagation (LRP) approach was implemented. This enabled the model predictions to be decomposed into relevance scores for each individual input value. In other words, it provided information regarding which features of the time-varying EMG profiles were unique to each individual. Through extensive testing, we have shown that the LRP results, and by extent the activation signatures, are highly consistent between conditions and across days. In addition, they are minimally influenced by the dataset used to train the model. Additionally, we proposed a method for visualizing each individual's muscle activation signature, which has several potential clinical and scientific applications. This is the first study to provide conclusive evidence of the existence of individual muscle activation signatures.

Addresses

Laboratory 'Movement, Interactions, Performance' (EA 4334), University of Nantes, Nantes, France.

Autoren

Jeroen Aeles
Fabian Horst
Sebastian Lapuschkin
Lilian Lacourpaille
François Hug

DOI

10.1098/rsif.2020.0770

eISSN

1742-5662

Externe Identifier

PubMed Identifier: 33435843
PubMed Central ID: PMC7879771

Funding acknowledgements

German Ministry for Education and Research:
Institut Universitaire de France:
French national research agency: ANR-19-CE17-002-01

Open access

false

ISSN

1742-5689

Ausgabe der Veröffentlichung

174

Zeitschrift

Journal of the Royal Society, Interface

Schlüsselwörter

Muscles
Muscle, Skeletal
Humans
Electromyography
Walking
Movement
Machine Learning

Sprache

eng

Medium

Print-Electronic

Online publication date

2021

Paginierung

20200770

Datum der Veröffentlichung

2021

Status

Published

Datum der Datenerfassung

2021

Titel

Revealing the unique features of each individual's muscle activation signatures.

Sub types

Research Support, Non-U.S. Gov't
research-article
Journal Article

Ausgabe der Zeitschrift

18

Files

https://royalsocietypublishing.org/doi/pdf/10.1098/rsif.2020.0770 https://europepmc.org/articles/PMC7879771?pdf=render

Datenquelle: Europe PubMed Central

Abstract

There is growing evidence that each individual has unique movement patterns, or signatures. The exact origin of these movement signatures, however, remains unknown. We developed an approach that can identify individual muscle activation signatures during two locomotor tasks (walking and pedalling). A linear support vector machine was used to classify 78 participants based on their electromyographic (EMG) patterns measured on eight lower limb muscles. To provide insight into decision-making by the machine learning classification model, a layer-wise relevance propagation (LRP) approach was implemented. This enabled the model predictions to be decomposed into relevance scores for each individual input value. In other words, it provided information regarding which features of the time-varying EMG profiles were unique to each individual. Through extensive testing, we have shown that the LRP results, and by extent the activation signatures, are highly consistent between conditions and across days. In addition, they are minimally influenced by the dataset used to train the model. Additionally, we proposed a method for visualizing each individual's muscle activation signature, which has several potential clinical and scientific applications. This is the first study to provide conclusive evidence of the existence of individual muscle activation signatures.

Autoren

Jeroen Aeles
Fabian Horst
Sebastian Lapuschkin
Lilian Lacourpaille
François Hug

Autoren-URL

https://www.ncbi.nlm.nih.gov/pubmed/33435843

DOI

10.1098/rsif.2020.0770

eISSN

1742-5662

Externe Identifier

PubMed Central ID: PMC7879771

Ausgabe der Veröffentlichung

174

Zeitschrift

J R Soc Interface

Schlüsselwörter

electromyography
machine learning
motor control
movement
neural control
Electromyography
Humans
Machine Learning
Movement
Muscle, Skeletal
Muscles
Walking

Sprache

eng

Country

England

Paginierung

20200770

Datum der Veröffentlichung

2021

Status

Published

Datum, an dem der Datensatz öffentlich gemacht wurde

2021

Titel

Revealing the unique features of each individual's muscle activation signatures.

Sub types

Journal Article
Research Support, Non-U.S. Gov't

Ausgabe der Zeitschrift

18

Datenquelle: PubMed

Abstract

There is growing evidence that each individual has unique movement patterns, or signatures. The exact origin of these movement signatures, however, remains unknown. We developed an approach that can identify individual muscle activation signatures during two locomotor tasks (walking and pedalling). A linear support vector machine was used to classify 78 participants based on their electromyographic (EMG) patterns measured on eight lower limb muscles. To provide insight into decision-making by the machine learning classification model, a layer-wise relevance propagation (LRP) approach was implemented. This enabled the model predictions to be decomposed into relevance scores for each individual input value. In other words, it provided information regarding which features of the time-varying EMG profiles were unique to each individual. Through extensive testing, we have shown that the LRP results, and by extent the activation signatures, are highly consistent between conditions and across days. In addition, they are minimally influenced by the dataset used to train the model. Additionally, we proposed a method for visualizing each individual’s muscle activation signature, which has several potential clinical and scientific applications. This is the first study to provide conclusive evidence of the existence of individual muscle activation signatures.

Autoren

Jeroen Aeles
Fabian Horst
Sebastian Lapuschkin
Lilian Lacourpaille
François Hug

DOI

10.1098/rsif.2020.0770

Zeitschrift

Journal of the Royal Society, Interface

Notes

file: http://www.ncbi.nlm.nih.gov/pubmed/33435843 file: http://www.ncbi.nlm.nih.gov/pubmed/33435843

Artikelnummer

174

Paginierung

20200770 - 20200770

Datum der Veröffentlichung

2021

Datum der Datenerfassung

2021

Titel

Revealing the unique features of each individual’s muscle activation signatures

Sub types

article

Ausgabe der Zeitschrift

18

Datenquelle: Manual

Revealing the unique features of each individual’s muscle activation signatures

Files

Werkzeuge