Clinical assessment and interpretation of dysarthria in ALS using attention based deep learning AI models.

Journal: NPJ digital medicine
Published Date:

Abstract

Speech dysarthria is a key symptom of neurological conditions like ALS, yet existing AI models designed to analyze it from audio signal rely on handcrafted features with limited inference performance. Deep learning approaches improve accuracy but lack interpretability. We propose an attention-based deep learning AI model to assess dysarthria severity based on listener effort ratings. Using 2,102 recordings from 125 participants, rated by three speech-language pathologists on a 100-point scale, we trained models directly from recordings collected remotely. Our best model achieved R of 0.92 and RMSE of 6.78. Attention-based interpretability identified key phonemes, such as vowel sounds influenced by 'r' (e.g., "car," "more"), and isolated inspiration sounds as markers of speech deterioration. This model enhances precision in dysarthria assessment while maintaining clinical interpretability. By improving sensitivity to subtle speech changes, it offers a valuable tool for research and patient care in ALS and other neurological disorders.

Authors

  • Michele Merler
    IBM Research, Yorktown Heights, NY, USA.
  • Carla Agurto
    IBM Research, Yorktown Heights, NY, USA.
  • Julian Peller
    EverythingALS, Peter Cohen Foundation, Los Altos, CA, USA.
  • Esteban Roitberg
    EverythingALS, Peter Cohen Foundation, Los Altos, CA, USA.
  • Alan Taitz
    SRI International, Menlo Park, CA, USA.
  • Marcos A Trevisan
    Universidad de Buenos Aires, Facultad de Ciencias Exactas y Naturales, Departamento de Física - CONICET - Instituto de Física Interdisciplinaria y Aplicada (INFINA), Buenos Aires, Argentina.
  • Indu Navar
    EverythingALS, Peter Cohen Foundation, Los Altos, CA, USA.
  • James D Berry
    MGH Institute of Health Professions, Boston, MA, USA.
  • Ernest Fraenkel
    Massachusetts Institute of Technology, Cambridge, MA, USA.
  • Lyle W Ostrow
    Lewis Katz School of Medicine, Temple University, Philadelphia, PA, USA.
  • Guillermo A Cecchi
    IBM Research, Yorktown Heights, NY, USA.
  • Raquel Norel
    IBM Research, Yorktown Heights, NY, USA. rnorel@us.ibm.com.

Keywords

No keywords available for this article.