[Automated speech analysis can distinguish frontotemporal dementia from depression: a pilot study]

Area of Science:

• Neurology
• Computational Linguistics
• Psychiatry

Context:

• Differentiating behavioral variant frontotemporal dementia (bvFTD) from depression is diagnostically challenging.
• Automated speech analysis is an emerging tool for neurological and psychiatric diagnostics.

Purpose:

• To evaluate the efficacy of automated speech analysis in distinguishing bvFTD from depressive disorders.
• To compare the diagnostic value of acoustic and semantic speech features.

Summary:

• Acoustic speech features demonstrated 80% accuracy, while semantic features showed 70.8% accuracy in differentiating bvFTD from depression.
• Machine learning classified acoustic and semantic values extracted from transcribed interviews of patients with bvFTD, depression, and healthy controls.

Impact:

• Automated speech analysis, particularly acoustic features, shows promise as a non-invasive tool for diagnosing bvFTD.
• Future research should explore larger populations and account for confounders like age, sex, and ethnicity in speech analysis for bvFTD diagnosis.