Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Predicting Parkinson's Disease Progression Using Statistical and Neural Mixed Effects Models: Comparative Study on Longitudinal Biomarkers

Ran Tong, Lanruo Wang, Tong Wang, Wei Yan

TL;DR

This work addresses predicting Parkinson's disease progression using longitudinal voice biomarkers to forecast total UPDRS scores, confronting within-subject correlations and nonlinear progression. It benchmarks traditional Linear Mixed Models against Generalized Neural Network Mixed Models (GNMM) and Neural Mixed Effects (NME) on the Oxford Parkinson's Telemonitoring Voice dataset. The results show that Generalized Additive Mixed Models (GAMM) provide the best predictive performance and model fit, while GNMM and NME offer limited gains under the studied conditions, underscoring the value of nonlinear temporal effects and random effects with interpretability. The findings inform telemonitoring workflows by highlighting when simpler, flexible mixed-effects models may outperform more complex neural architectures on realistic, limited-sample clinical data.

Abstract

Predicting Parkinson's Disease (PD) progression is crucial, and voice biomarkers offer a non-invasive method for tracking symptom severity (UPDRS scores) through telemonitoring. Analyzing this longitudinal data is challenging due to within-subject correlations and complex, nonlinear patient-specific progression patterns. This study benchmarks LMMs against two advanced hybrid approaches: the Generalized Neural Network Mixed Model (GNMM) (Mandel 2021), which embeds a neural network within a GLMM structure, and the Neural Mixed Effects (NME) model (Wortwein 2023), allowing nonlinear subject-specific parameters throughout the network. Using the Oxford Parkinson's telemonitoring voice dataset, we evaluate these models' performance in predicting Total UPDRS to offer practical guidance for PD research and clinical applications.

Predicting Parkinson's Disease Progression Using Statistical and Neural Mixed Effects Models: Comparative Study on Longitudinal Biomarkers

TL;DR

This work addresses predicting Parkinson's disease progression using longitudinal voice biomarkers to forecast total UPDRS scores, confronting within-subject correlations and nonlinear progression. It benchmarks traditional Linear Mixed Models against Generalized Neural Network Mixed Models (GNMM) and Neural Mixed Effects (NME) on the Oxford Parkinson's Telemonitoring Voice dataset. The results show that Generalized Additive Mixed Models (GAMM) provide the best predictive performance and model fit, while GNMM and NME offer limited gains under the studied conditions, underscoring the value of nonlinear temporal effects and random effects with interpretability. The findings inform telemonitoring workflows by highlighting when simpler, flexible mixed-effects models may outperform more complex neural architectures on realistic, limited-sample clinical data.

Abstract

Predicting Parkinson's Disease (PD) progression is crucial, and voice biomarkers offer a non-invasive method for tracking symptom severity (UPDRS scores) through telemonitoring. Analyzing this longitudinal data is challenging due to within-subject correlations and complex, nonlinear patient-specific progression patterns. This study benchmarks LMMs against two advanced hybrid approaches: the Generalized Neural Network Mixed Model (GNMM) (Mandel 2021), which embeds a neural network within a GLMM structure, and the Neural Mixed Effects (NME) model (Wortwein 2023), allowing nonlinear subject-specific parameters throughout the network. Using the Oxford Parkinson's telemonitoring voice dataset, we evaluate these models' performance in predicting Total UPDRS to offer practical guidance for PD research and clinical applications.

Paper Structure

This paper contains 12 sections, 37 equations, 3 figures, 6 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Dataset
  5. Traditional Methods: Linear Mixed--Effects Model (LMM)
  6. Generalized Additive Mixed Model (GAMM)
  7. Generalized Neural Network Mixed Model (GNMM) for Non‑linear Longitudinal Modeling
  8. Evaluation.
  9. Neural Mixed-Effects (NME) Model for Longitudinal UPDRS Prediction
  10. NME Parameterization.
  11. Analysis of the Results
  12. Summary and Future Work

Figures (3)

  • Figure 1: Data analysis of the dataset
  • Figure 2: Top row: Residuals, fixed effect Q-Q, and random effect Q-Q plots from the original model using total_UPDRS. Bottom row: Diagnostics after log-transforming the response. Transformation improves variance stabilization and normality.
  • Figure 3: Estimated spline effect of test_time from GAMM, showing nonlinear progression over time.