Insights into cognitive pupillometry: Evaluation of the utility of pupillary metrics for assessing cognitive load in normative and clinical samples.

Abstract

Despite increasing use of pupillometry to understand cognitive deficits in clinical populations, there is no consensus on what pupillary metrics are most useful. In this study, we compare the reliability, sensitivity to mild traumatic brain injury (TBI) and cognitive load differences, and the likelihood of replication between various pupillary metrics. Raw pupil diameter was recorded from 15 patients with chronic mild TBI and 23 uninjured controls while they performed a cognitive task with three levels of cognitive load. Several pupillary metrics were quantified from the pupillary data and submitted to tests of internal consistency, group ∗ cognitive load repeated measures ANOVAs, and bootstrapping analyses of effect size. Most pupillary metrics demonstrated acceptable, good, or excellent reliability. Metrics differed in sensitivity to group, cognitive load, and their interaction. Bootstrapping analyses revealed that peak-based metrics are more likely to replicate than means- or ratio-based metrics. Several pupillary metrics were determined to have great utility for measuring cognitive load in clinical or normative samples. This study directly addresses a known methodological gap in the cognitive pupillometry literature.