Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

PPINtonus: Early Detection of Parkinson's Disease Using Deep-Learning Tonal Analysis

Varun Reddy

TL;DR

PPINtonus tackles early PD detection by analyzing voice through deep learning augmented with synthetic data. It combines cGAN-based data generation with PRAAT-derived vocal features in a dense neural network, designed to be robust to household noise and real-world recording conditions. The approach demonstrates high performance, achieving 92.5% accuracy and 92.7% precision with recall of 1.0, and identifies sustained vowel phonation as a particularly informative BVM. This work highlights a scalable, non-invasive screening tool that could significantly improve PD diagnosis timeliness in resource-constrained regions, provided data diversity and edge-device deployment are further optimized.

Abstract

PPINtonus is a system for the early detection of Parkinson's Disease (PD) utilizing deep-learning tonal analysis, providing a cost-effective and accessible alternative to traditional neurological examinations. Partnering with the Parkinson's Voice Project (PVP), PPINtonus employs a semi-supervised conditional generative adversarial network to generate synthetic data points, enhancing the training dataset for a multi-layered deep neural network. Combined with PRAAT phonetics software, this network accurately assesses biomedical voice measurement values from a simple 120-second vocal test performed with a standard microphone in typical household noise conditions. The model's performance was validated using a confusion matrix, achieving an impressive 92.5 \% accuracy with a low false negative rate. PPINtonus demonstrated a precision of 92.7 \%, making it a reliable tool for early PD detection. The non-intrusive and efficient methodology of PPINtonus can significantly benefit developing countries by enabling early diagnosis and improving the quality of life for millions of PD patients through timely intervention and management.

PPINtonus: Early Detection of Parkinson's Disease Using Deep-Learning Tonal Analysis

TL;DR

PPINtonus tackles early PD detection by analyzing voice through deep learning augmented with synthetic data. It combines cGAN-based data generation with PRAAT-derived vocal features in a dense neural network, designed to be robust to household noise and real-world recording conditions. The approach demonstrates high performance, achieving 92.5% accuracy and 92.7% precision with recall of 1.0, and identifies sustained vowel phonation as a particularly informative BVM. This work highlights a scalable, non-invasive screening tool that could significantly improve PD diagnosis timeliness in resource-constrained regions, provided data diversity and edge-device deployment are further optimized.

Abstract

PPINtonus is a system for the early detection of Parkinson's Disease (PD) utilizing deep-learning tonal analysis, providing a cost-effective and accessible alternative to traditional neurological examinations. Partnering with the Parkinson's Voice Project (PVP), PPINtonus employs a semi-supervised conditional generative adversarial network to generate synthetic data points, enhancing the training dataset for a multi-layered deep neural network. Combined with PRAAT phonetics software, this network accurately assesses biomedical voice measurement values from a simple 120-second vocal test performed with a standard microphone in typical household noise conditions. The model's performance was validated using a confusion matrix, achieving an impressive 92.5 \% accuracy with a low false negative rate. PPINtonus demonstrated a precision of 92.7 \%, making it a reliable tool for early PD detection. The non-intrusive and efficient methodology of PPINtonus can significantly benefit developing countries by enabling early diagnosis and improving the quality of life for millions of PD patients through timely intervention and management.
Paper Structure (9 sections, 5 figures, 3 tables)

This paper contains 9 sections, 5 figures, 3 tables.

Table of Contents

  1. Background
  2. Methodology
  3. Data Collection
  4. Deep Learning Methodology
  5. Real Time PD Detection
  6. Proposed Model Architecture
  7. Results
  8. Discussion
  9. Conclusion

Figures (5)

  • Figure 1: Illustration of a Generative Adversarial Network (Gharakhanian).
  • Figure 2: Figure 2: Differences in specific BVMs between Parkinson's patients and healthy individuals (Rusz)
  • Figure 3: cGAN model able to model pre-existing training data over 10,000 epochs
  • Figure 4: (a) Training and validation accuracy over 100 epochs. (b) Training and validation loss over 100 epochs.
  • Figure 5: Accuracy of various vocal tests in extracting reliable Biomedical Voice Measurements (BVMs) for Parkinson's Disease detection.