PsycAssist: A Web-Based Artificial Intelligence System Designed for Adaptive Neuropsychological Assessment and...

Area of Science:

• Neuropsychology
• Artificial Intelligence
• Cognitive Assessment

Background:

• Assessing executive functions in clinical populations is challenging due to limitations with conventional tests.
• Adaptive assessments offer a personalized approach for evaluating individuals with specific needs.

Purpose of the Study:

• Introduce PsycAssist, a web-based AI system for adaptive neuropsychological assessment and training.
• Present the architecture and adaptive engine of PsycAssist.
• Introduce two available tests: Adap-ToL for planning and MatriKS for fluid intelligence.

Main Methods:

• Developed PsycAssist based on procedural knowledge space theory.
• Implemented an adaptive assessment engine within PsycAssist.
• Created Adap-ToL (Tower of London-like) and MatriKS (Raven-like) tests.

Main Results:

• Evaluators found Adap-ToL and MatriKS appropriate for their intended neuropsychological assessment purposes.
• Test-takers reported a positive experience with the adaptive assessments.
• PsycAssist demonstrated flexibility and scalability for various test types.

Conclusions:

• PsycAssist is an innovative tool for tailored neuropsychological evaluation and training.
• The system is well-suited for clinical practice, adapting to individual characteristics.
• Adaptive assessments like Adap-ToL and MatriKS show promise in improving assessment accuracy and user experience.