Investigating the links between sustained attention, gaze fixation patterns, and face identity discrimination performance

Background. Sustained attention ability is often only considered during cognitively demanding task completion; it is rarely considered when tasks are less demanding or primarily perceptual. Should variability in sustained attention be shown to impact performance on such tasks, researchers may benefit from considering this in future experiments. Methods. 19 adult participants (19-25 years) completed a two-alternative forced-choice face identity discrimination task with eye tracking. Face pairs with three different viewing conditions were used to manipulate access to local information cues: (i) both forward-facing (F-F), (ii) forward- vs. side-facing (F-S), and (iii), both side-facing (S-S). In each trial, one face identity was either identical to the other face or changed by either 4% or 8% toward another identity. Sustained attention was defined via a novel metric derived from the Connors CPT. Finally, performance was compared across two identical sequentially presented blocks of trials. Results. A logistic mixed model revealed no attention-related (p=.26) or block-related (p=.12) differences in identity discrimination. A mixed model demonstrated a trend for the novel attention measure to significantly predict the number of fixations made during trials; participants with lower sustained attention made fewer fixations before responding (p=.09). All participants further reduced their number of fixations in the second testing block (p < .001). Additionally, they simultaneously made more errors (p < .001) and made fewer fixations (p = .01) in the F-S condition with restricted access to local information. Conclusions. Manipulating access to local information cues affected the number of fixations made and increased error rate during task completion. While lower sustained attention ability may result in fewer fixations per trial, this behaviour did not negatively impact discrimination accuracy. Results are interpreted as more liberal response bias in participants with lower sustained attention ability. Additional analyses will examine visual areas of interest fixated during the task.