Eye-Tracking Analysis for Cognitive Load Estimation in Wearable Mixed Reality

Previous research has indicated that cognitive load significantly affects performance and user experience in mixed reality applications. This paper analyses eye-tracking data to investigate their relation- ship with cognitive load in wearable mixed reality (on HoloLens2 headset). The aim is to determine whether fixation and saccadic features, together with age, visual condition and previous experience with mixed reality systems, can be used to infer cognitive load levels in a specific application scenario. The analysis is done on a user study with 17 individuals performing a discovery task in a wearable mixed reality application designed for building occupancy monitoring. Results reveal that participants can be grouped into two distinct groups that separate individuals by experience and may match two different cognitive load levels, confirming that longer fixation frequency and saccade duration appear to be significant predictors of lower cognitive load in wearable mixed reality.