Gaze-based Metrics of Cognitive Load in a Conjunctive Visual Memory Task

Measurement of Cognitive Load (CL) is of considerable importance to Human-Computer Interaction (HCI) as it relates to ease of learn- and usability. Numerous methods have been used for this purpose, both subjective and objective. The former relies on perception of expended effort subject to confounding factors. Among objective measures, eye tracking has demonstrated its effectiveness as a precise and non-invasive alternative. Since eye-tracking indices have mainly been tested in tasks involving single visual properties (e.g., shape), this work aims to compare their sensitivity to conjunctive features. Specifically, this study evaluates the Low/High Index of Pupillary Activity (LHIPA) and microsaccade magnitude in the Color Visual Short-Term Memory (CVSTM) task versus the previously validated n-back task. Results show LHIPA and microsaccade magnitude are as effective in discerning baseline and task in CVSTM as in n-back, replicating earlier results and extending their reliability for evaluation of CL in tasks involving multivalent conjunctive features.