Publication detail

Foreseeing wearing-off state in Parkinson’s disease patients, a multimodal approach with the usage of machine learning and wearables

SKIBIŃSKA, J. SHAH, A. CHANNA, A. SHAH SYED, M. SYED, Z. HOŠEK, J.

Original Title

Foreseeing wearing-off state in Parkinson’s disease patients, a multimodal approach with the usage of machine learning and wearables

Type

conference paper

Language

English

Original Abstract

Parkinson’s disease (PD) is one of the most common neurogenerative disorders in the world and therefore of great concern in the provision of healthcare to people. The motor and cognitive impairments manifesting in this disease have a significant impact on the deterioration of PD patients’ quality of life. The commonly administrated drug to minimise PD-caused impairments is levodopa. Nevertheless, PD patients could suffer because the applied dose is not enough or they will need the next medication, this phenomenon is so-called wearing-off. Thereby, the paper presents a method to forecast wearing-off from data of 12 patients from the 5th ABC challenge 15 minutes in advance. The proposed methodology uses features derived from the heart rate, stress level, sleep, and number of steps along with machine learning to predict wearing-off in patients. Experiments were conducted using five different machine-learning algorithms for all of the participants and targeting a personalised approach. The main contribution of the paper is the identification of the most suitable algorithm to foresee the wearing-off state. XGBoost classifier achieved 0.43 F1-score and 0.72 balanced accuracy. The advantage of this study is also depicting the most relevant features: mean condition, sleep-relevant, and time-relevant. Therefore, this solution has clinical value. Additionally, the benefit of this paper is the proposed patients-targeted solution to predict the wearing-off state. This methodology provides a promising result in terms of the possibility of automated sensor-based detection of medication wearing off in PD patients.

Keywords

Parkinson's disease, machine learning, wearable, wearing-off

Authors

SKIBIŃSKA, J.; SHAH, A.; CHANNA, A.; SHAH SYED, M.; SYED, Z.; HOŠEK, J.

Released

24. 1. 2025

Publisher

Taylor and Francis

ISBN

9781032648422

Book

Activity, Behavior, and Healthcare Computing

Edition

1

Pages from

1

Pages to

14

Pages count

14

URL

https://www.taylorfrancis.com/chapters/edit/10.1201/9781032648422-22/foreseeing-wearing-state-parkinson-disease-patients-multimodal-approach-usage-machine-learning-wearables-justyna-skibi%C5%84ska-muhammad-zaigham-abbas-shah-asma-channa-muhammad-shehram-shah-syed-zafi-sherhan-syed-jiri-hosek?context=ubx&refId=77feb6ba-d3f2-444a-858d-b5ef9d98d0a7

BibTex

@inproceedings{BUT184998,
  author="Justyna {Skibińska} and Abbas {Shah} and Asma {Channa} and Muhammad Shehram {Shah Syed} and Zafi Sherhan {Syed} and Jiří {Hošek}",
  title="Foreseeing wearing-off state in Parkinson’s disease patients, a multimodal approach with the usage of machine learning and wearables",
  booktitle="Activity, Behavior, and Healthcare Computing",
  year="2025",
  series="1",
  pages="14",
  publisher="Taylor and Francis",
  isbn="9781032648422",
  url="https://www.taylorfrancis.com/chapters/edit/10.1201/9781032648422-22/foreseeing-wearing-state-parkinson-disease-patients-multimodal-approach-usage-machine-learning-wearables-justyna-skibi%C5%84ska-muhammad-zaigham-abbas-shah-asma-channa-muhammad-shehram-shah-syed-zafi-sherhan-syed-jiri-hosek?context=ubx&refId=77feb6ba-d3f2-444a-858d-b5ef9d98d0a7"
}

Responsibility: Ing. Marek Strakoš