Speech analysis software for diagnosis of dementias

This technology is an automated screening algorithm that analyzes patterns in verbal communication for early stage diagnosis of Alzheimer’s disease and related cognitive impairments.

Unmet Need: Affordable, noninvasive diagnostic assay to detect cognitive impairments

Mild cognitive impairment (MCI), Alzheimer’s disease (AD), and related early-stage dementias remain chronically underdiagnosed and untreated due to various factors such as the inability of patients to recognize early symptoms, limited availability of biomarkers, and clinicians’ having insufficient time and data to make a final diagnostic assessment. The currently available diagnostic assays remain invasive, costly and are unable to detect early stages of disease.

The Technology: Speech analysis algorithm to detect cognitive impairment

MCIADscreen utilizes a unique pipeline to model the three primary elements of speech: phonetic motor planning, semantic and syntactic language organization, and psycholinguistic features. This software processes audio recordings of patients’ verbal descriptions during a picture description test and uses various machine learning models to analyze the speech trace. By integrating acoustic, linguistic, and semantic features of speech, MCIADscreen provides an accurate and unbiased prediction for the likelihood of a patient developing symptoms of cognitive impairment.

This technology has been validated using the Dementia Bank English Pitt Corpus dataset, specifically through the “Cookie-Theft” picture description test.

Applications:

• Diagnostic assays for mild cognitive impairment, Alzheimer’s disease, and related dementias
• Screening platform for other cognitive disorders, such as stroke and traumatic brain injury
• Research tool for testing of neurological therapeutics
• Monitoring mechanism for longitudinal clinical trials

Advantages:

• Cost-effective
• Non-invasive
• Automated and unbiased detection algorithm
• Rapid detection of cognitive status
• Facilitates early intervention opportunities
• Generalizable to other neurological assessment tests and clinical settings

Lead Inventor:

Maryam Zolnoori, Ph.D.

Patent Information:

Patent Pending

Related Publications:

• Zolnoori M, Zolnour A, Topaz M. ADscreen. “A speech processing-based screening system for automatic identification of patients with Alzheimer’s disease and related dementia” Artificial Intelligence in Medicine. 2023 Sept; 143: 102624

Tech Ventures Reference:

• IR CU23077

• Licensing Contact: Sara Gusik