Speech Acoustic Markers Can Detect Mild Cognitive Impairment in Parkinson's Disease

  • 0Department of Neurology, Boston Medical Center, Boston, MA 02118 USA.
IEEE Journal of Selected Topics in Signal Processing +

|

|

Summary

This summary is machine-generated.

Researchers identified distinct speech acoustic features linked to mild cognitive impairment in Parkinson's disease (PD-MCI). These biomarkers, particularly from picture descriptions, show promise for developing new tools to detect and monitor PD-MCI.

Area Of Science

  • Neurology
  • Speech Science
  • Biomarker Discovery

Background

  • Speech biomarkers are established for Parkinson's disease (PD) motor symptoms.
  • Biomarkers for mild cognitive impairment in PD (PD-MCI) remain understudied.

Purpose Of The Study

  • Identify speech acoustic features associated with PD-MCI.
  • Develop and evaluate a model to differentiate PD-MCI from cognitively normal PD (PD-NC) individuals.

Main Methods

  • Analyzed speech samples from 42 PD participants classified as PD-MCI or PD-NC.
  • Extracted acoustic features from reading and picture description tasks.
  • Utilized Gaussian mixture models (GMMs) for classification.

Main Results

  • Picture description tasks yielded more PD-MCI-associated acoustic features than reading tasks.
  • A fused GMM model achieved an AUC of 0.82 for discriminating PD-MCI from PD-NC.
  • Associated acoustic features involved multiple speech subsystems.

Conclusions

  • PD-MCI exhibits a unique speech acoustic signature.
  • These findings support the development of novel tools for PD-MCI detection and monitoring.

Related Concept Videos