Biomarkers

Area of Science:

• Neurology
• Artificial Intelligence in Healthcare
• Digital Health

Background:

• Early dementia detection is crucial for timely treatment and support.
• Current diagnostic pathways involve lengthy waits and require specialized staff.
• Remote, smart technologies are needed to improve dementia diagnosis efficiency.

Purpose of the Study:

• To evaluate CognoSpeak, a novel speech and video analysis tool, for detecting cognitive impairment.
• To assess the feasibility and accuracy of CognoSpeak in distinguishing between dementia, mild cognitive impairment (MCI), and healthy controls (HC).

Main Methods:

• CognoSpeak utilizes a web-based app with a virtual agent to collect speech and video recordings.
• Subjects respond to questions designed to assess cognitive status.
• Data from over 1000 healthy volunteers and over 200 individuals with MCI and dementia were analyzed.

Main Results:

• Initial results from 126 subjects (8 dementia, 39 MCI, 79 HC) show promising classification accuracy.
• CognoSpeak is a low-cost, repeatable, and non-invasive tool.
• The system can be administered quickly and remotely, and also tracks mood and anxiety.

Conclusions:

• CognoSpeak demonstrates potential as a valuable tool for clinical practice and trials.
• The technology facilitates remote, at-home cognitive assessments.
• Further analysis on a larger dataset will provide more insights into its predictive capabilities for cognition, mood, and anxiety.