Bayesian network modeling of risk and prodromal markers of Parkinson’s disease

PLOS ONE RESEARCH ARTICLE Bayesian network modeling of risk and prodromal markers of Parkinson’s disease Meemansa Sood ID1,2, Ulrike Suenkel3, Anna-Katharina von Thaler4, Helena U. Zacharias5,6, Kathrin Brockmann4,7, Gerhard W. Eschweiler3,8, Walter Maetzler9, Daniela Berg4,9, Holger Fröhlich ID1,2☯*, Sebastian Heinzel ID9,10☯* a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 1 Department of Bioinformatics, Fraunhofer Institute for Algorithms and Scientific Computing (SCAI), Sankt Augustin, Germany, 2 Bonn-Aachen International Center for IT, University of Bonn, Bonn, Germany, 3 Department of Psychiatry and Psychotherapy, Tübingen University Hospital, Tübingen, Germany, 4 Department of Neurodegeneration, Hertie Institute for Clinical Brain Research, University of Tübingen, Tübingen, Germany, 5 Department of Internal Medicine I, University Medical Center Schleswig-Holstein, Campus Kiel, Kiel, Germany, 6 Institute of Clinical Molecular Biology, Kiel University and University Medical Center Schleswig-Holstein, Campus Kiel, Kiel, Germany, 7 German Center for Neurodegenerative Diseases, University of Tübingen, Tübingen, Germany, 8 Geriatric Center, Tübingen University Hospital, Tübingen, Germany, 9 Department of Neurology, Christian-Albrechts University, Kiel, Germany, 10 Institute of Medical Informatics and Statistics, Kiel University, University Hospital Schleswig-Holstein, Kiel, Germany ☯ These authors contributed equally to this work. * (SH); (HF) OPEN ACCESS Citation: Sood M, Suenkel U, von Thaler A-K, Zacharias HU, Brockmann K, Eschweiler GW, et al. (2023) Bayesian network modeling of risk and prodromal markers of Parkinson’s disease. PLoS ONE 18(2): e0280609. https://doi.org/10.1371/ journal.pone.0280609 Editor: Tommaso Martino, Policlinico Riuniti of Foggia: Neuroscience Department, S.C. Ospedaliera of Neurology-Stroke Unit, ITALY Received: October 10, 2022 Accepted: January 3, 2023 Published: February 24, 2023 Copyright: © 2023 Sood 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. Data Availability Statement: Given that this study uses a compehensive body of prospective TREND cohort data and to prevent misuse, the longitudinal data of >1200 participants cannot be shared publicly. Data will be shared (upon request) with researchers and for research purposes only. For data requests, please contact the Organizational team and Steering committee of the TREND study via: info(at)trend-studie.de. Abstract Parkinson’s disease (PD) is characterized by a long prodromal phase with a multitude of markers indicating an increased PD risk prior to clinical diagnosis based on motor symptoms. Current PD prediction models do not consider interdependencies of single predictors, lack differentiation by subtypes of prodromal PD, and may be limited and potentially biased by confounding factors, unspecific assessment methods and restricted access to comprehensive marker data of prospective cohorts. We used prospective data of 18 established risk and prodromal markers of PD in 1178 healthy, PD-free individuals and 24 incident PD cases collected longitudinally in the Tübingen evaluation of Risk factors for Early detection of NeuroDegeneration (TREND) study at 4 visits over up to 10 years. We employed artificial intelligence (AI) to learn and quantify PD marker interdependencies via a Bayesian network (BN) with probabilistic confidence estimation using bootstrapping. The BN was employed to generate a synthetic cohort and individual marker profiles. Robust interdependencies were observed for BN edges from age to subthreshold parkinsonism and urinary dysfunction, sex to substantia nigra hyperechogenicity, depression, non-smoking and to constipation; depression to symptomatic hypotension and excessive daytime somnolence; solvent exposure to cognitive deficits and to physical inactivity; and non-smoking to physical inactivity. Conversion to PD was interdependent with prior subthreshold parkinsonism, sex and substantia nigra hyperechogenicity. Several additional interdependencies with lower probabilistic confidence were identified. Synthetic subjects generated via the BN based representation of the TREND study were realistic as assessed through multiple comparison approaches of real and synthetic data. Altogether our work demonstrates the potential of modern AI approaches (specifically BNs) both for modelling and understanding interdependencies between PD risk and prodromal markers, PLOS ONE | https://doi.org/10.1371/journal.pone.0280609 February 24, 2023 1 / 15 PLOS ONE Funding: The TREND study is being conducted at the University Hospital Tübingen and has been supported by the Hertie Institute for Clinical Brain Research, the DZNE, the Geriatric Center of Tübingen, the Center for Integrative Neuroscience, Teva Pharmaceutical Industries, Union Chimique Belge, Janssen Pharmaceuticals, the International Parkinson Foundation and the German Research Society (DFG). This work was (partially) funded by DIGIPD (01KU2110), a project supported by the Federal Ministry of Science and Education (BMBF), under the frame of ERA PerMed. HUZ is supported by the Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung; BMBF) within the framework of the e:Med research and funding concept (grant 01ZX1912A). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. There was no additional external funding received for this study. There was no additional external funding received for this study. Competing interests: The authors have declared that no competing interests exist. Network of prodromal Parkinson markers which are so far not accounted for in PD prediction models, as well as for generating realistic synthetic data. Introduction Parkinson’s disease (PD) is characterized by progressive neurodegeneration that has usually advanced for many years before it is clinically diagnosed [1]. In addition to old age, a multitude of risk markers, such as genetic factors, lifestyle, environmental factors, (comorbid) diseases (e.g., diabetes) as well as biomarkers (e.g., low plasma urate levels, hyperechogenicity of the substantia nigra) have been shown to indicate an increased risk of PD in prospective studies [2, 3]. Moreover, prodromal markers like depression, autonomous dysfunction, REM-sleep behavior disorder (RBD), subtle motor signs and pathological dopaminergic imaging [3–5] may already indicate early neurodegenerative processes that can ultimately lead to the clinical diagnosis of PD. The International Parkinson and Movement Disorder Society (MDS) research criteria for prodromal PD [3, 6] have been designed to review and continually update the predictive values of risk and prodromal markers (...truncated)