Brain predictors of fatigue in rheumatoid arthritis: A machine learning study

María Goñi; Neil Basu; Alison D Murray; Gordon D Waiter

doi:10.1371/journal.pone.0269952

Brain predictors of fatigue in rheumatoid arthritis: A machine learning study

María Goñi^* (Corresponding Author), Neil Basu, Alison D Murray, Gordon D Waiter

^*Corresponding author for this work

University of Glasgow

Research output: Contribution to journal › Article › peer-review

4 Citations (Scopus)

3 Downloads (Pure)

Abstract

BACKGROUND: Fatigue is a common and burdensome symptom in Rheumatoid Arthritis (RA), yet is poorly understood. Currently, clinicians rely solely on fatigue questionnaires, which are inherently subjective measures. For the effective development of future therapies and stratification, it is of vital importance to identify biomarkers of fatigue. In this study, we identify brain differences between RA patients who improved and did not improve their levels of fatigue based on Chalder Fatigue Scale variation (ΔCFS≥ 2), and we compared the performance of different classifiers to distinguish between these samples at baseline.

METHODS: Fifty-four fatigued RA patients underwent a magnetic resonance (MR) scan at baseline and 6 months later. At 6 months we identified those whose fatigue levels improved and those for whom it did not. More than 900 brain features across three data sets were assessed as potential predictors of fatigue improvement. These data sets included clinical, structural MRI (sMRI) and diffusion tensor imaging (DTI) data. A genetic algorithm was used for feature selection. Three classifiers were employed in the discrimination of improvers and non-improvers of fatigue: a Least Square Linear Discriminant (LSLD), a linear Support Vector Machine (SVM) and a SVM with Radial Basis Function kernel.

RESULTS: The highest accuracy (67.9%) was achieved with the sMRI set, followed by the DTI set (63.8%), whereas classification performance using clinical features was at the chance level. The mean curvature of the left superior temporal sulcus was most strongly selected during the feature selection step, followed by the surface are of the right frontal pole and the surface area of the left banks of the superior temporal sulcus.

CONCLUSIONS: The results presented evidence a superiority of brain metrics over clinical metrics in predicting fatigue changes. Further exploration of these methods may support clinicians to triage patients towards the most appropriate fatigue alleviating therapies.

Original language	English
Article number	e0269952
Number of pages	15
Journal	PloS ONE
Volume	17
Issue number	6
Early online date	27 Jun 2022
DOIs	https://doi.org/10.1371/journal.pone.0269952
Publication status	Published - 27 Jun 2022

Bibliographical note

Open Access via the PLOS OA agreement
Funding: NB received funding from Pfizer (https://www.pfizer.com/) for data collection and from the Sir Jules Thorn Charitable Trust (https://julesthorntrust.org.uk/) to fund this research as part of MG's PhD. GDW received funding from Roland Sutton Academic Trust (https://www.abdn.ac.uk/ims/research/abic/roland-sutton-academic-trust-1427.php, grant number 0093-R-21) to cover publication fees. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data Availability Statement

Data Availability: The data that supports the findings of this study are available in the University of Aberdeen Pure system (doi: 10.20392/a0aa9c28-729c-478b-aa69-efbd6028623c).

Access to Document

10.1371/journal.pone.0269952Licence: CC BY

Goni_et_al_Brain_predictors_Plos_one_vor
© 2022 Goñi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. https://creativecommons.org/licenses/by/4.0/
Final published version, 1.68 MBLicence: CC BY

Aberdeen Biomedical Imaging Centre
Gordon Waiter (Manager) & Teresa Morris (Facilities Co-ordinator)
Aberdeen Biomedical Imaging Centre
Research Facilities: Facility

Cite this

@article{ffa430117832478780a42e35ffc2be25,

title = "Brain predictors of fatigue in rheumatoid arthritis: A machine learning study",

abstract = "BACKGROUND: Fatigue is a common and burdensome symptom in Rheumatoid Arthritis (RA), yet is poorly understood. Currently, clinicians rely solely on fatigue questionnaires, which are inherently subjective measures. For the effective development of future therapies and stratification, it is of vital importance to identify biomarkers of fatigue. In this study, we identify brain differences between RA patients who improved and did not improve their levels of fatigue based on Chalder Fatigue Scale variation (ΔCFS≥ 2), and we compared the performance of different classifiers to distinguish between these samples at baseline.METHODS: Fifty-four fatigued RA patients underwent a magnetic resonance (MR) scan at baseline and 6 months later. At 6 months we identified those whose fatigue levels improved and those for whom it did not. More than 900 brain features across three data sets were assessed as potential predictors of fatigue improvement. These data sets included clinical, structural MRI (sMRI) and diffusion tensor imaging (DTI) data. A genetic algorithm was used for feature selection. Three classifiers were employed in the discrimination of improvers and non-improvers of fatigue: a Least Square Linear Discriminant (LSLD), a linear Support Vector Machine (SVM) and a SVM with Radial Basis Function kernel.RESULTS: The highest accuracy (67.9%) was achieved with the sMRI set, followed by the DTI set (63.8%), whereas classification performance using clinical features was at the chance level. The mean curvature of the left superior temporal sulcus was most strongly selected during the feature selection step, followed by the surface are of the right frontal pole and the surface area of the left banks of the superior temporal sulcus.CONCLUSIONS: The results presented evidence a superiority of brain metrics over clinical metrics in predicting fatigue changes. Further exploration of these methods may support clinicians to triage patients towards the most appropriate fatigue alleviating therapies.",

author = "Mar{\'i}a Go{\~n}i and Neil Basu and Murray, {Alison D} and Waiter, {Gordon D}",

note = "Open Access via the PLOS OA agreement Funding: NB received funding from Pfizer (https://www.pfizer.com/) for data collection and from the Sir Jules Thorn Charitable Trust (https://julesthorntrust.org.uk/) to fund this research as part of MG's PhD. GDW received funding from Roland Sutton Academic Trust (https://www.abdn.ac.uk/ims/research/abic/roland-sutton-academic-trust-1427.php, grant number 0093-R-21) to cover publication fees. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.",

year = "2022",

month = jun,

day = "27",

doi = "10.1371/journal.pone.0269952",

language = "English",

volume = "17",

journal = "PloS ONE",

issn = "1932-6203",

publisher = "PUBLIC LIBRARY SCIENCE",

number = "6",

}

TY - JOUR

T1 - Brain predictors of fatigue in rheumatoid arthritis

T2 - A machine learning study

AU - Goñi, María

AU - Basu, Neil

AU - Murray, Alison D

AU - Waiter, Gordon D

N1 - Open Access via the PLOS OA agreement Funding: NB received funding from Pfizer (https://www.pfizer.com/) for data collection and from the Sir Jules Thorn Charitable Trust (https://julesthorntrust.org.uk/) to fund this research as part of MG's PhD. GDW received funding from Roland Sutton Academic Trust (https://www.abdn.ac.uk/ims/research/abic/roland-sutton-academic-trust-1427.php, grant number 0093-R-21) to cover publication fees. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

PY - 2022/6/27

Y1 - 2022/6/27

N2 - BACKGROUND: Fatigue is a common and burdensome symptom in Rheumatoid Arthritis (RA), yet is poorly understood. Currently, clinicians rely solely on fatigue questionnaires, which are inherently subjective measures. For the effective development of future therapies and stratification, it is of vital importance to identify biomarkers of fatigue. In this study, we identify brain differences between RA patients who improved and did not improve their levels of fatigue based on Chalder Fatigue Scale variation (ΔCFS≥ 2), and we compared the performance of different classifiers to distinguish between these samples at baseline.METHODS: Fifty-four fatigued RA patients underwent a magnetic resonance (MR) scan at baseline and 6 months later. At 6 months we identified those whose fatigue levels improved and those for whom it did not. More than 900 brain features across three data sets were assessed as potential predictors of fatigue improvement. These data sets included clinical, structural MRI (sMRI) and diffusion tensor imaging (DTI) data. A genetic algorithm was used for feature selection. Three classifiers were employed in the discrimination of improvers and non-improvers of fatigue: a Least Square Linear Discriminant (LSLD), a linear Support Vector Machine (SVM) and a SVM with Radial Basis Function kernel.RESULTS: The highest accuracy (67.9%) was achieved with the sMRI set, followed by the DTI set (63.8%), whereas classification performance using clinical features was at the chance level. The mean curvature of the left superior temporal sulcus was most strongly selected during the feature selection step, followed by the surface are of the right frontal pole and the surface area of the left banks of the superior temporal sulcus.CONCLUSIONS: The results presented evidence a superiority of brain metrics over clinical metrics in predicting fatigue changes. Further exploration of these methods may support clinicians to triage patients towards the most appropriate fatigue alleviating therapies.

AB - BACKGROUND: Fatigue is a common and burdensome symptom in Rheumatoid Arthritis (RA), yet is poorly understood. Currently, clinicians rely solely on fatigue questionnaires, which are inherently subjective measures. For the effective development of future therapies and stratification, it is of vital importance to identify biomarkers of fatigue. In this study, we identify brain differences between RA patients who improved and did not improve their levels of fatigue based on Chalder Fatigue Scale variation (ΔCFS≥ 2), and we compared the performance of different classifiers to distinguish between these samples at baseline.METHODS: Fifty-four fatigued RA patients underwent a magnetic resonance (MR) scan at baseline and 6 months later. At 6 months we identified those whose fatigue levels improved and those for whom it did not. More than 900 brain features across three data sets were assessed as potential predictors of fatigue improvement. These data sets included clinical, structural MRI (sMRI) and diffusion tensor imaging (DTI) data. A genetic algorithm was used for feature selection. Three classifiers were employed in the discrimination of improvers and non-improvers of fatigue: a Least Square Linear Discriminant (LSLD), a linear Support Vector Machine (SVM) and a SVM with Radial Basis Function kernel.RESULTS: The highest accuracy (67.9%) was achieved with the sMRI set, followed by the DTI set (63.8%), whereas classification performance using clinical features was at the chance level. The mean curvature of the left superior temporal sulcus was most strongly selected during the feature selection step, followed by the surface are of the right frontal pole and the surface area of the left banks of the superior temporal sulcus.CONCLUSIONS: The results presented evidence a superiority of brain metrics over clinical metrics in predicting fatigue changes. Further exploration of these methods may support clinicians to triage patients towards the most appropriate fatigue alleviating therapies.

U2 - 10.1371/journal.pone.0269952

DO - 10.1371/journal.pone.0269952

M3 - Article

C2 - 35759489

SN - 1932-6203

VL - 17

JO - PloS ONE

JF - PloS ONE

IS - 6

M1 - e0269952

ER -

Brain predictors of fatigue in rheumatoid arthritis: A machine learning study

Abstract

Bibliographical note

Data Availability Statement

Access to Document

Fingerprint

Equipment

Aberdeen Biomedical Imaging Centre

Cite this