Development and Piloting of Co.Ge.: A Web-Based Digital Platform for Generative and Clinical Cognitive Assessment

Area of Science:

• Digital health technologies
• Cognitive science
• Psychiatry

Background:

• Current cognitive testing methods can be labor-intensive and may not capture high-resolution behavioral data.
• There is a need for advanced digital platforms to facilitate cognitive assessment and data analysis.

Purpose of the Study:

• To introduce Co.Ge., a Cognitive Generative digital platform for cognitive testing.
• To describe the architecture of Co.Ge. and report on a pilot study evaluating its usability and data collection capabilities.

Main Methods:

• Co.Ge. is a modular, web-based platform (Laravel-PHP, MySQL) with features like study management, APIs, and data encryption.
• A pilot study administered the Auditory Verbal Learning Test (AVLT) to clinical and non-clinical participants.
• Data analyzed included accuracy, word recall, reaction times (RTs), and user experience ratings using Frequentist and Bayesian Generalized Linear Mixed Models (GLMMs).

Main Results:

• User experience ratings averaged above 4/5, indicating high acceptability among participants (n=30).
• Co.Ge. successfully provided standardized clinical ratings, accuracy, and RTs from the AVLT pilot data (n=123).
• Bayesian GLMMs with a Gamma distribution provided the best fit for RT data, revealing associations (e.g., education) missed by simpler analyses.

Conclusions:

• The Co.Ge. prototype is technically sound and clinically accurate, capable of extracting high-resolution behavioral data.
• Co.Ge. supports traditional cognitive outcomes and advanced generative models for exploring individual cognitive mechanisms.
• The platform is poised to advance personalized profiling and digital phenotyping for precision psychiatry and rehabilitation.