COGNITIVE MODELING FOR HUMAN FACTORS TESTING: DEVELOPING A STRUCTURED TASK ANALYSIS FRAMEWORK FOR DESIGNERS

Abstract

As part of system testing and verification (STV) for devices, designers must empiricallydemonstrate that their products can be used safely before they reach the market. Regulatory agencies such as the FDA require evaluation of human performance, both psychomotor and cognitive activities in the use of medical devices. Traditional human factors evaluations emphasize psychomotor activities, limiting the ability to identify cognitive root causes of use error. Existing Cognitive Task Analysis (CTA) methods offer structured approaches for modeling cognition, but these methods were designed for experts with a background in cognitive sciences. These methods also lack standardized syntax for describing the tasks, a discrete list of tasks to use, and were not designed for integration with product documentation. These limitations will be addressed by developing and validating a structured, designer-accessible CTA method that embeds cognitive and perceptual tasks within psychomotor task flows. A five-phase methodology was used to construct and validate this method. Phase 1 involved selecting, refining, and embedding conceptual models of cognitive and psychomotor task lists. Phase 2 involved combining the unified model with an information processing model, establishing information sequencing rules. Phase 3 called for analyzing Instructions for Use (IFUs) to derive Task Categories, creating a bridge between human–machine interactions (HMI) and the unified model. Phase 4 involved the development of rule-based cognitive sentence syntax and CTA templates to standardize task decomposition. Phase 5 empirically validated the latent constructs of the unified model through a study with 145 subjects using exploratory structural modeling (ESEM) to ensure the model was designer accessible and that it aligned with its theoretical underpinnings. ESEM factor loadings demonstrated theoretical coherence, supporting the validity of the standardized structure. The outcome of this research is an empirically validated method for use in CTA creation that reduces bias in task decomposition, improves repeatability, and combines cognition with product design in a form accessible to designers. The method strengthens risk analysis and STV by providing design teams with a standardized way to represent cognition alongside psychomotor tasks, improving the traceability of use error pathways and supporting safer, more reliable products.