Pre-trained Convolutional Neural Networks Identify Parkinson’s Disease from Spectrogram Images of Voice Samples
2024

Using AI to Detect Parkinson's Disease from Voice Samples

publication Evidence: moderate

Author Information

Author(s): Rahmatallah Yasir, Kemp Aaron, Iyer Anu, Pillai Lakshmi, Larson-Prior Linda, Virmani Tuhin, Prior Fred

Hypothesis

Can convolutional neural networks effectively identify Parkinson's disease from spectrogram images of voice samples?

Conclusion

The study found that a convolutional neural network can effectively identify Parkinson's disease from voice samples, showing better performance than traditional methods.

Supporting Evidence

  • The approach was tested on a larger dataset recorded using smartphones.
  • The study reported differences in important features due to the limited bandwidth of telephonic lines.
  • Mel-scale spectrograms showed a small but statistically significant gain in classification performance.

Takeaway

Researchers used a computer program to listen to people's voices and found a way to tell if they have Parkinson's disease.

Methodology

The study used convolutional neural networks with transfer learning to analyze spectrogram images of voice recordings.

Statistical Information

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.21203/rs.3.rs-5348708

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication