Journal DOI 10.68956/dyh5-kwte

Creative Commons source reading · Neurodegeneration biomarkers and longitudinal studies

Machine learning-based prediction of longitudinal cognitive decline in early Parkinson's disease using multimodal features.

Hannes Almgren; Milton Camacho; Alexandru Hanganu; Mekale Kibreab; Richard Camicioli; Zahinoor Ismail; Nils D Forkert; Oury Monchi

Originally published in Scientific reports · 2023-08-14

From the article

Selected source text. JournalOS removes markup and shortens excerpts where indicated. The full article provides the complete context.

Abstract

Patients with Parkinson's Disease (PD) often suffer from cognitive decline. Accurate prediction of cognitive decline is essential for early treatment of at-risk patients. The aim of this study was to develop and evaluate a multimodal machine learning model for the prediction of continuous cognitive decline in patients with early PD. We included 213 PD patients from the Parkinson's Progression Markers Initiative (PPMI) database. Machine learning was used to predict change in Montreal Cognitive Assessment (MoCA) score using the difference between baseline and 4-years follow-up data as outcome. Input features were categorized into four sets: clinical test scores, cerebrospinal fluid (CSF) biomarkers, brain volumes, and genetic variants. All combinations of input feature sets were added to a basic model, which consisted of demographics and baseline cognition. An iterative scheme using RReliefF-based feature ranking and support vector regression in combination with tenfold cross validation was used to determine the optimal number of predictive features and to evaluate model performance for each combination of input feature sets.

Discussion

In this study we developed, compared, and evaluated unimodal and multimodal machine learning models to predict cognitive decline in Parkinson’s disease (PD) over four years using a wide range of features from different modalities, including CSF biomarkers, genetics, regional brain volumes, and clinical test scores. Our best performing model included CSF markers and clinical test scores, in addition to simple demographics and baseline cognition.

Conclusion

Our best machine learning model for cognitive decline in de novo Parkinson’s disease (PD) consisted of CSF biomarkers and clinical test scores, in addition to basic demographics and baseline cognition. Pearson correlation between predicted and observed MoCA change was 0.44. The predictive features in our model included features that have previously been linked to Alzheimer’s disease (AD), such as CSF amyloid-beta and tau concentration. Thus, our study adds to the body of evidence showing an important overlap between features related to cognitive decline in PD and AD.

Editorial reading guide

  • Was the evidence cross-sectional, longitudinal or experimental?
  • How did the authors account for age, disease stage and differences among participants?