Procedural trial generation for cognitive assessment: An executive function planning task

Area of Science:

• Cognitive Science
• Computational Psychology
• Human-Computer Interaction

Background:

• Cognitive assessments require diverse trials for robust evaluation.
• Traditional methods for trial generation can be time-consuming and lack scalability.
• Procedural content generation (PCG) offers a potential solution for automated content creation.

Purpose of the Study:

• To introduce a computational approach for generating new cognitive assessment trials.
• To demonstrate this methodology using a novel executive functioning planning task.
• To provide guidelines for digital cognitive task development and procedural trial generation.

Main Methods:

• Developed a scalable and adaptable method for generating experimental trials, inspired by PCG in gaming.
• Created an executive functioning planning task involving block matching, assessing planning abilities.
• Employed a generate-and-test methodology: pseudo-random puzzle creation followed by a solving algorithm to ensure solvability.

Main Results:

• The procedural generation technique successfully produced a large number of diverse and solvable trials for the planning task.
• The developed task is accessible online and has undergone usability and functional testing.
• The approach demonstrates the feasibility of using PCG for creating varied cognitive assessment content.

Conclusions:

• PCG is a promising technique for the scalable and automated generation of high-quality cognitive assessment trials.
• Future work should explore advanced algorithms for optimizing trial generation and psychometric validation.
• Guidelines are provided to facilitate the development of digital cognitive tasks and PCG for various research applications.