Ocular Indicators of Mental Workload: A Comparison of Scanpath Entropy and Fixations Clustering

Among the different eye-tracking metrics, both the entropy-based analysis of the scanpath and the spatial distribution of fixations points have been suggested as indices of mental workload. However, they have never been directly compared so far. In this study, eye movements were recorded from fourteen subjects while performing a visuo-spatial task (the “spot the differences” puzzle game). Subjective workload ratings were also collected throughout the session along with performance data. Entropy Rate and Nearest Neighbor Index show matching patterns both indicating high mental workload after two minutes of visual scanning. As predicted elsewhere, the distribution of fixations becomes more clustered under increases of the visuo-spatial demand and showing the same pattern of the Entropy rate that becomes more stereotyped. Temporal demand, instead, has been reported to produce spreading of the fixation pattern. Interestingly, both ocular indices seem to anticipate the drop in performance that is visible after the sixth minute of visual scanning.