Workload assessment for mental arithmetic tasks using the task-evoked pupillary response

Abstract

Pupillometry is a promising method for assessing mental workload and could
be helpful in the optimization of systems that involve human–computer
interaction. The present study focuses on replicating the studies by Ahern
(1978) and Klingner (2010), which found that for three levels of difficulty
of mental multiplications, the more difficult multiplications yielded larger
dilations of the pupil. Using a remote eye tracker, our research expands upon
these two previous studies by statistically testing for each 1.5 s interval of the
calculation period (1) the mean absolute pupil diameter (MPD), (2) the mean
pupil diameter change (MPDC) with respect to the pupil diameter during the
pre-stimulus accommodation period, and (3) the mean pupil diameter change rate
(MPDCR). An additional novelty of our research is that we compared the pupil
diameter measures with a self-report measure of workload, the NASA Task Load
Index (NASA-TLX), and with the mean blink rate (MBR). The results showed that
the findings of Ahern and Klingner were replicated, and that the MPD and MPDC
discriminated just as well between the lowest and highest difficulty levels as did
the NASA-TLX. The MBR, on the other hand, did not differentiate between the
difficulty levels. Moderate to strong correlations were found between the MPDC
and the proportion of incorrect responses, indicating that the MPDC was higher for
participants with a poorer performance. For practical applications, validity could be
improved by combining pupillometry with other physiological techniques.