Fall Risk Prediction in Multiple Sclerosis Using Postural Sway Measures: A Machine Learning Approach

Ruopeng Sun, Katherine L. Hsieh, Jacob J. Sosnoff

Research output: Contribution to journalArticle

Abstract

Numerous postural sway metrics have been shown to be sensitive to balance impairment and fall risk in individuals with MS. Yet, there are no guidelines concerning the most appropriate postural sway metrics to monitor impairment. This investigation implemented a machine learning approach to assess the accuracy and feature importance of various postural sway metrics to differentiate individuals with MS from healthy controls as a function of physiological fall risk. 153 participants (50 controls and 103 individuals with MS) underwent a static posturography assessment and a physiological fall risk assessment. Participants were further classified into four subgroups based on fall risk: controls, low-risk MS (n = 34), moderate-risk MS (n = 27), high-risk MS (n = 42). Twenty common sway metrics were derived following standard procedures and subsequently used to train a machine learning algorithm (random forest – RF, with 10-fold cross validation) to predict individuals’ fall risk grouping. The sway-metric based RF classifier had high accuracy in discriminating controls from MS individuals (>86%). Sway sample entropy was identified as the strongest feature for classification of low-risk MS individuals from healthy controls. Whereas for all other comparisons, mediolateral sway amplitude was identified as the strongest predictor for fall risk groupings.

Original languageEnglish (US)
Article number16154
JournalScientific reports
Volume9
Issue number1
DOIs
StatePublished - Dec 1 2019

Fingerprint

Multiple Sclerosis
Machine Learning
Entropy
Guidelines

ASJC Scopus subject areas

  • General

Cite this

Fall Risk Prediction in Multiple Sclerosis Using Postural Sway Measures : A Machine Learning Approach. / Sun, Ruopeng; Hsieh, Katherine L.; Sosnoff, Jacob J.

In: Scientific reports, Vol. 9, No. 1, 16154, 01.12.2019.

Research output: Contribution to journalArticle

@article{e5b7e300ff1f44aab1a1385e7f4d7067,
title = "Fall Risk Prediction in Multiple Sclerosis Using Postural Sway Measures: A Machine Learning Approach",
abstract = "Numerous postural sway metrics have been shown to be sensitive to balance impairment and fall risk in individuals with MS. Yet, there are no guidelines concerning the most appropriate postural sway metrics to monitor impairment. This investigation implemented a machine learning approach to assess the accuracy and feature importance of various postural sway metrics to differentiate individuals with MS from healthy controls as a function of physiological fall risk. 153 participants (50 controls and 103 individuals with MS) underwent a static posturography assessment and a physiological fall risk assessment. Participants were further classified into four subgroups based on fall risk: controls, low-risk MS (n = 34), moderate-risk MS (n = 27), high-risk MS (n = 42). Twenty common sway metrics were derived following standard procedures and subsequently used to train a machine learning algorithm (random forest – RF, with 10-fold cross validation) to predict individuals’ fall risk grouping. The sway-metric based RF classifier had high accuracy in discriminating controls from MS individuals (>86{\%}). Sway sample entropy was identified as the strongest feature for classification of low-risk MS individuals from healthy controls. Whereas for all other comparisons, mediolateral sway amplitude was identified as the strongest predictor for fall risk groupings.",
author = "Ruopeng Sun and Hsieh, {Katherine L.} and Sosnoff, {Jacob J.}",
year = "2019",
month = "12",
day = "1",
doi = "10.1038/s41598-019-52697-2",
language = "English (US)",
volume = "9",
journal = "Scientific Reports",
issn = "2045-2322",
publisher = "Nature Publishing Group",
number = "1",

}

TY - JOUR

T1 - Fall Risk Prediction in Multiple Sclerosis Using Postural Sway Measures

T2 - A Machine Learning Approach

AU - Sun, Ruopeng

AU - Hsieh, Katherine L.

AU - Sosnoff, Jacob J.

PY - 2019/12/1

Y1 - 2019/12/1

N2 - Numerous postural sway metrics have been shown to be sensitive to balance impairment and fall risk in individuals with MS. Yet, there are no guidelines concerning the most appropriate postural sway metrics to monitor impairment. This investigation implemented a machine learning approach to assess the accuracy and feature importance of various postural sway metrics to differentiate individuals with MS from healthy controls as a function of physiological fall risk. 153 participants (50 controls and 103 individuals with MS) underwent a static posturography assessment and a physiological fall risk assessment. Participants were further classified into four subgroups based on fall risk: controls, low-risk MS (n = 34), moderate-risk MS (n = 27), high-risk MS (n = 42). Twenty common sway metrics were derived following standard procedures and subsequently used to train a machine learning algorithm (random forest – RF, with 10-fold cross validation) to predict individuals’ fall risk grouping. The sway-metric based RF classifier had high accuracy in discriminating controls from MS individuals (>86%). Sway sample entropy was identified as the strongest feature for classification of low-risk MS individuals from healthy controls. Whereas for all other comparisons, mediolateral sway amplitude was identified as the strongest predictor for fall risk groupings.

AB - Numerous postural sway metrics have been shown to be sensitive to balance impairment and fall risk in individuals with MS. Yet, there are no guidelines concerning the most appropriate postural sway metrics to monitor impairment. This investigation implemented a machine learning approach to assess the accuracy and feature importance of various postural sway metrics to differentiate individuals with MS from healthy controls as a function of physiological fall risk. 153 participants (50 controls and 103 individuals with MS) underwent a static posturography assessment and a physiological fall risk assessment. Participants were further classified into four subgroups based on fall risk: controls, low-risk MS (n = 34), moderate-risk MS (n = 27), high-risk MS (n = 42). Twenty common sway metrics were derived following standard procedures and subsequently used to train a machine learning algorithm (random forest – RF, with 10-fold cross validation) to predict individuals’ fall risk grouping. The sway-metric based RF classifier had high accuracy in discriminating controls from MS individuals (>86%). Sway sample entropy was identified as the strongest feature for classification of low-risk MS individuals from healthy controls. Whereas for all other comparisons, mediolateral sway amplitude was identified as the strongest predictor for fall risk groupings.

UR - http://www.scopus.com/inward/record.url?scp=85074622744&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85074622744&partnerID=8YFLogxK

U2 - 10.1038/s41598-019-52697-2

DO - 10.1038/s41598-019-52697-2

M3 - Article

C2 - 31695127

AN - SCOPUS:85074622744

VL - 9

JO - Scientific Reports

JF - Scientific Reports

SN - 2045-2322

IS - 1

M1 - 16154

ER -