EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Real-World Classification of Pain States in Multiple Sclerosis Patients using Wearable Sensor Data and Ensemble Machine Learning

Real-World-Klassifikation von Schmerzzuständen bei Patient:innen mit Multipler Sklerose mithilfe von Wearable-Sensordaten und Ensemble-Machine-Learning

Serena Moscato, Silvia Orlandi, Stefania Pozzi, Giada Lullini, Loredana Sabattini, Fabio La Porta, Lorenzo Chiari

Digital Biomarkers · 01.10.2026 · 10 (1) · 202-215

DOI: 10.1159/dib/acpag001 ↗

Auf einen Blick

Bei 18 MS-Patient:innen wurden Wearable-Daten mit Selbstangaben zu Schmerzen verglichen. Die besten Modelle erreichten F1-Werte bis 0,78.

Fragestellung

Können Wearable-Physiologie-Signale Schmerzzustände bei Menschen mit Multipler Sklerose klassifizieren?

Methode

18 MS-Patient:innen wurden 48 h mit dem Empatica E4 überwacht. Aus 5-, 10- und 15-Minuten-Fenstern vor stündlichen Schmerzangaben wurden Merkmale extrahiert und mit 5 ML-Algorithmen per Leave-one-subject-out-CV getestet.

Ergebnisse

In der univariaten Analyse waren nach Korrektur 2 von 43 Merkmalen signifikant. Das beste F1-Ergebnis (0,78) erzielte ein SVM-Modell mit 2 Merkmalen aus 5-Minuten-Fenstern; XGBoost zeigte die beste Sensitivität/Spezifität (0,69/0,68).

Einschränkungen

Die Ergebnisse sollten in größeren, unabhängigen Kohorten validiert werden.

Schlussfolgerung der Autor:innen

Wearable-physiologische Signale können für die Klassifikation von Schmerzzuständen bei Menschen mit MS relevant sein; die Ergebnisse sind vorläufig.

Kernaussagen

  • 2 von 43 Merkmalen unterschieden sich signifikant nach Korrektur
  • Bestes F1: 0,78 mit SVM aus 5-Minuten-Fenstern
  • Vorläufige Daten; Validierung in größeren Kohorten nötig

Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.

Kurz gefasst

Fachartikel in der Zeitschrift Digital Biomarkers, online erschienen am 01.10.2026. Verfasst von Serena Moscato und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Electrical, Electronic, and Information Engineering "Guglielmo Marconi" - DEI, University of Bologna, Bologna, Italy. Die Arbeit nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

Auf einen BlickAbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Pain assessment remains challenging due to its subjective nature and reliance on self-reports. These challenges are particularly relevant for multiple sclerosis (MS) patients, for whom pain is among the most common and fluctuating symptoms, occurring in multiple forms. Wearable sensors combined with machine learning (ML) offer the potential for continuous and more objective pain monitoring by leveraging physiological signals in real-world contexts.

Methods

We monitored 18 MS patients over 48 h using the Empatica E4 wristband to collect photoplethysmography, electrodermal activity, skin temperature, and accelerometer data. Physiological features were extracted from 5-, 10-, and 15-min time windows preceding hourly self-reported pain assessment during waking hours. Univariate associations with pain states were evaluated using linear mixed effects models, correcting for multiple comparisons. For classification purposes, feature selection was performed via recursive feature elimination, and 5 ML algorithms were trained and tested with leave-one-subject-out cross-validation. Model performance was evaluated using the F1 score and the sensitivity-specificity balance.

Results

For univariate analysis, after correction, only 2 out of 43 features showed significant differences between pain and no-pain states. For the classification task, models based on 5-min windows generally performed the best. The highest F1 score (0.78) was achieved by a Support Vector Machine using 2 features from 5-min windows, while XGBoost achieved the best sensitivity-specificity balance (0.69/0.68) with 2 features from 15-min windows.

Conclusion

Wearable physiological signals may contain information relevant to pain-state classification in people with MS. However, these results should be considered preliminary, requiring validation in larger, independent cohorts to establish the robustness and generalizability of the proposed approach.

Schlagwörter

Stichwörter der Autor:innen

EnsembleMachine learningPain assessmentReal-world dataWearable sensors

Autor:innen und Institutionen

  • Department of Electrical, Electronic, and Information Engineering "Guglielmo Marconi" - DEI, University of Bologna, Bologna, Italy.
  • Health Sciences and Technologies, Interdepartmental Center for Industrial Research (CIRI-SDV), University of Bologna, Bologna, Italy.
  • Rehabilitation Bioengineering Laboratory, IRCCS Istituto delle Scienze Neurologiche di Bologna, Bologna, Italy.
  • DATER Riabilitazione Ospedaliera, UA Riabilitazione, Ospedale Bellaria, Azienda Unità Sanitaria Locale di Bologna, Bologna, Italy.
  • IRCCS Istituto delle Scienze Neurologiche di Bologna, Bologna, Italy.
  • UOSI Riabilitazione Sclerosi Multipla, IRCCS Istituto delle Scienze Neurologiche di Bologna, Bologna, Italy.

Zitieren

APA

Moscato, S., Orlandi, S., Pozzi, S., Lullini, G., Sabattini, L., La Porta, F., & Chiari, L. (2026). Real-World Classification of Pain States in Multiple Sclerosis Patients using Wearable Sensor Data and Ensemble Machine Learning. Digital Biomarkers, 10(1), 202-215. https://doi.org/10.1159/dib/acpag001

Vancouver

Moscato S, Orlandi S, Pozzi S, Lullini G, Sabattini L, La Porta F, et al. Real-World Classification of Pain States in Multiple Sclerosis Patients using Wearable Sensor Data and Ensemble Machine Learning. Digital Biomarkers. 2026;10(1):202-215. doi:10.1159/dib/acpag001

BibTeX
@article{Moscato2026,
  author = {Moscato, Serena and Orlandi, Silvia and Pozzi, Stefania and Lullini, Giada and Sabattini, Loredana and La Porta, Fabio and Chiari, Lorenzo},
  title = {Real-World Classification of Pain States in Multiple Sclerosis Patients using Wearable Sensor Data and Ensemble Machine Learning},
  journal = {Digital Biomarkers},
  year = {2026},
  volume = {10},
  number = {1},
  pages = {202--215},
  doi = {10.1159/dib/acpag001}
}
Kostenlos starten →