Prodromal Diagnosis of Lewy Body Diseases Based on Actigraphy

Warning

This publication doesn't include Faculty of Economics and Administration. It includes Central European Institute of Technology. Official publication website can be found on muni.cz.
Authors

MIKULEC Marek GALAZ Zoltan MEKYSKA Jiri MUCHA Jan BRABENEC Luboš MORÁVKOVÁ Ivona REKTOROVÁ Irena

Year of publication 2022
Type Article in Proceedings
Conference 2022 45th International Conference on Telecommunications and Signal Processing (TSP)
MU Faculty or unit

Central European Institute of Technology

Citation
Web https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9851316
Doi http://dx.doi.org/10.1109/TSP55681.2022.9851316
Keywords actigraphy; machine learning; neurodegenerative diseases; Lewy body diseases; RBD; SHAP values; sleep diary; XGBoost
Description This paper is devoted to the computerized automated diagnosis of the prodromal state of Lewy body diseases (LBD) based on actigraphy. LBD is a group of neurodegenerative diseases that require early treatment to alleviate the course of the disease and improve the quality of the lives of patients. This work proposes a method of prodromal diagnosis of LBD based on quantitative analysis of actigraphic sleep data. A new method of sleep and wake detection based on the XGBoost classifier and the angle of the z-axis is introduced, which achieves 83 % accuracy and surpasses the results of state-of-the-art methods. Furthermore, a method that can distinguish subjects with pro-dromal LBD (50 subjects with Parkinson's disease, dementia with Lewy bodies or mild cognitive impairment) and healthy controls (63 subjects) with 94 % accuracy was introduced. The sensitivity of the method of 100 % and specificity of 91% was considered sufficient for clinical practice and the proposed methods can help develop decision-making tools that maximize the potential for an early and objective diagnosis of LBD.
Related projects:

You are running an old browser version. We recommend updating your browser to its latest version.