Related Experiment Video
Updated: Jul 3, 2026
09:06
Whole-brain Segmentation and Change-point Analysis of Anatomical Brain MRI—Application in Premanifest Huntington's Disease
Published on: June 9, 2018
Development and validation of a machine learning model to detect psychiatric symptoms in Huntington's disease using
Quang Tuan Rémy Nguyen1,2,3,4,5,6, Hadrien Titeux3,5,6,7, Rachid Riad6,7
1Département d'Études Cognitives, École normale supérieure, PSL University, Paris, France.
Plos One
|July 1, 2026
Summary
Speech analysis can identify psychiatric symptoms like obsessive-compulsive behavior, depression, and apathy in Huntington
Area of Science:
- Neurology
- Psychiatry
- Computational Linguistics
Background:
- Huntington's disease (HD) is a progressive neurodegenerative disorder characterized by motor, cognitive, and psychiatric symptoms.
- Current machine learning (ML) speech analysis effectively detects motor and cognitive symptoms in HD but struggles with psychiatric manifestations.
- Accurate detection of psychiatric symptoms is crucial for comprehensive HD management and treatment.
Purpose of the Study:
- To investigate the efficacy of ML-based speech analysis in detecting psychiatric symptoms in individuals with Huntington's disease.
- To determine which speech feature types (linguistic, LASER, acoustic) are most effective for identifying specific psychiatric conditions in HD patients.
- To explore the potential of speech analysis as a non-invasive tool for psychiatric characterization in HD.
Main Methods:
- Audio recordings of six narrative speech tasks were collected from 89 genetically confirmed HD participants.
- Speech samples were analyzed for linguistic, LASER, and acoustic features by blinded speech therapists.
- ML classifier models were trained and tested to detect psychiatric symptoms, assessed via the Problem Behaviors Assessment Short version (PBA-s).
Main Results:
- Linguistic features successfully detected obsessive-compulsive behavior (OCB) and depression, with varying task performance.
- LASER features demonstrated effectiveness in identifying OCB, depression, and apathy, particularly from the 'red-riding hood' task.
- Acoustic features also detected depression and OCB, indicating a complementary role alongside linguistic and LASER features. Irritability detection was not successful across all feature types.
Conclusions:
- Speech analysis, particularly using linguistic and LASER features, can effectively detect specific psychiatric symptoms such as OCB, depression, and apathy in individuals with Huntington's disease.
- Acoustic features offer complementary insights, enhancing the psychiatric characterization of HD.
- These findings support the development of speech-based biomarkers for non-invasive psychiatric assessment in HD management.
