Abstract
The current study sought to extend upon the findings of DeLucia et al. (2018), who observed a vigilance decrement when observers were asked to detect the presence of an elementary perceptual feature. A second, harder condition was introduced that asked participants to search for the absence of an elementary perceptual feature. Results indicated that correct detections and response time were worse in the feature-absence condition than in the feature-presence condition, and both conditions elicited progressively slower response times as the task progressed (i.e., a vigilance decrement). Eye-tracking data suggested that searching for feature-presence was more efficient than searching for feature-absence. These results echo those of DeLucia et al., indicating that elementary feature detection is not automatic and cannot protect fully against the vigilance decrement. Our results also suggest that display designs that task users with detecting elementary features can augment overall, average vigilance performance.
Vigilance is an individual’s ability to sustain attentional focus over time for the detection of rare events (Davies & Parasuraman, 1982). Decades of vigilance research have shown that human performance is unreliable in vigilance tasks; the speed and accuracy of detections declines over time on task (Warm et al., 2008). This vigilance decrement is problematic because vigilance tasks remain prevalent in many settings such as healthcare and aviation (Warm et al., 2008). Consequently, many studies of vigilance have been conducted with the aim of developing a better understanding of the factors that affect vigilance performance so that vigilance performance can be improved (e.g., through improved display design; Greenlee et al., 2022). The current study continued these efforts by examining one potential means of improving vigilance performance or mitigating the vigilance decrement: elementary perceptual features.
Elementary features are simple perceptual characteristics of visual stimuli (e.g., color, orientation) that can be processed rapidly with minimal attentional effort. Research on visual search tasks suggests that individuals can search a distractor- filled display for an elementary feature in parallel, meaning that search and target detection are rapid and unaffected by the number of distractors in the display. These findings led to a proposal that the processing of elementary features was pre-attentive and automatic (Treisman & Gormican, 1988). Accordingly, searching for an elementary feature may be easier and less attentionally burdensome than searching for other types of stimuli, making elementary features a potential means of mitigating the vigilance decrement.
The primary aim of the current study was to evaluate this possibility. Toward that end, we replicated the task used in a previous study from DeLucia et al. (2018), that required detection of a target that possessed an elementary feature among fields of distractors that lacked that feature, but we also added a comparison condition that did the opposite (i.e., it required detection of a target that lacked an elementary feature among a field of distractors that all possessed that feature. Visual search research suggests that detecting elementary feature absence is more demanding than detecting feature presence (Treisman & Gormican, 1988). Thus, performance should be poorer when detecting elementary feature absence compared to detecting elementary feature presence.
The current study also included eye-tracking measures of task demand and search efficiency: average dwell time (operationalized as fixation duration) and the nearest neighbor index (NNI). Dwell times tend to decrease as visual task demands increase, suggesting that individuals must rapidly scan the display in the shortest time possible to detect a target (Dillard et al., 2014; Di Nocera et al., 2007). NNI measures the spatial distribution of fixations such that a wider distribution of fixations (i.e., higher NNI) indicates higher visual task demands while a clustered distribution of fixations (i.e., lower NNI) suggests lower visual task demands (Dillard et al., 2014).
Forty-nine individuals (33 females, 16 males) were randomly assigned to one of two display conditions: feature-present or feature-absent. In the feature-present condition, they searched for a slanted line (critical signal) among vertical distractor stimuli (neutral stimuli). In the feature-absent condition, they searched for a vertical line (critical signal) among slanted distractor stimuli (neutral stimuli).
Results revealed that those in the feature-present condition had overall greater correct detections and faster response times than those in the feature-absent condition. These findings echo past research indicating that elementary feature detection is more rapid and easier than searching for other types of stimuli (Treisman & Gormican, 1988). The performance results are further supported by eye-tracking measures that showed longer dwell times and lower NNI scores (i.e., more clustered fixations) in feature-presence than in feature-absence. This indicates that visual search was more efficient in feature-presence (i.e., little visual scanning was needed). Taken together, the current findings suggest that displays that code targets as elementary features can effectively augment mean vigilance performance.
However, elementary feature detection may not be beneficial for mitigating the vigilance decrement. Results found that response times slowed over time on task in both conditions and correct detections declined selectively in the feature-present condition. This suggests that searching for elementary feature presence does require some degree of attention (i.e., elementary feature detection is not automatic).
Practically, display designers should integrate feature-present critical signals into vigilance displays to augment mean vigilance performance, while keeping in mind that elementary features may not prevent the vigilance decrement.
