Measuring Cognitive Load with Eye Tracking to Guide the Design of Mixed Reality Data Visualisation Systems

This PhD research investigates how Mixed Reality (MR) technology can be leveraged to enhance learning and performance in scientific data analysis. In an era where vast datasets shape our understanding of the world, extracting insights from increasingly complex data and utilising them meaningfully presents a significant challenge. MR technology offers new ways to visualise and explore multifaceted information providing a flexible, three-dimensional space to support multiformat and multidimensional scientific data analysis. However, the components that constitute visualisations in MR and the cognitive aspects involved in processing it are often conflated in its evaluation. Cognitive load is a concept that refers to the burden on the limited memory capacity required to process information and perform tasks. Measuring cognitive load could help understand how specific visualisation elements impact information processing in analytical operations and lead to positive user outcomes. This research explores the relationships between MR visualisation techniques, cognitive factors, and eye movement metrics through three studies. The pilot study investigates eye movement metrics as a cognitive load measure while comparing 2D and 3D representations. The main study further applies this to examine how the use of MR’s extended space affects cognitive processing. Finally, the interview study evaluates the applicability of the design recommendations elicited from the main study. As a result, clearer links are established between fixation metrics and cognitive demands placed on the memory system. The research also demonstrates that stereoscopic 3D representation can be perceived as more detailed than 2D, and that extended screen space can help reduce the cognitive demand on working memory. The proposed cognitive model accommodates the extended spatial characteristics of MR and helps identify design attributes that reduce unnecessary cognitive burden. The practical design recommendations offer guidance for the development of MR systems that support cognitively more efficient learning and performance in scientific data analysis.