Attentional modulation of the carry over of eye-movements between tasks

Abstract

Task demands that influence scanning behaviour in one task can cause that behaviour to persist to a second unrelated task (carry over). This can also affect performance on a second task (e.g., hazard perception ratings), and has been attributed to a process of attentional bias that is modulated by top-down influences (Thompson & Crundall, 2011). In a series of experiments we explored how these top-down influences impact upon carry over. In all experiments, participants searched letters that were presented horizontally, vertically, or in a random array. They were then presented with a driving scene and rated the hazardousness of the scene. Carry over of eye-movements from the letter search to the scene was observed in all experiments. Furthermore, it was demonstrated that this carry over effect influenced hazard perception accuracy. The magnitude of carry over was correlated with task switching abilities, attentional conflicting, and attentional orienting (Experiment 1), and was affected by predictability of the primary task (Experiment 2). Furthermore, direct current stimulation of the left dorsolateral prefrontal cortex and parietal areas affected the magnitude of the effect (Experiment 3). These results indicate that carry over is modulated by the specific ability to orient attention and disengage from this orientation. Over orienting leads to increased carry over and insufficient task switching is detrimental to task performance. As a result the current experiments provide evidence that the carry over effect is strongly influenced by attentional processes, namely orienting, inhibition, and task switching.

Detrimental effects of carryover of eye movement behaviour on hazard perception accuracy: Effects of driver experience, difficulty of task, and hazardousness of road

“…Immediately following this, participants saw a road scene (or video clip) and were asked to memorise it (Experiment 1), rate it for hazardousness (Experiment 2), or respond to the onset of the hazard (Experiment 3). Even though the time spent completing the letter search was minimal, the orientation of letters in this task influenced the allocation of attention and eye movements when viewing the road scene, with increased vertical search following the vertically orientated letter-search task (see also Hills, Thompson, Jones, Piech, Painter, & Pake, 2016;Thompson, Howting, & Hills, 2015). In addition, responses to the hazards in Experiment 3 were made significantly quicker following letters presented horizontally compared to letters presented randomly or vertically This finding may have some important implications for real world driving scenarios.…”

“…They also display more horizontal scanning for more hazardous roads than novice drivers. Given that the carryover of eye movements interferes with hazard detection latency (Thompson & Crundall, 2011) and accuracy (Hills et al, 2016), potentially by adding to the task difficulty, it might be that it occurs more for hazardous road scenes than less hazardous ones. This is an empirical question rather than one based on theory.…”

Carryover of scanning behaviour affects upright face recognition differently to inverted face recognition

“…The direct cause of this carryover is under debate. It may be based on a failure to inhibit eye movement behaviour or attentional distribution from one task to another (Hills, et al, 2016). It may also be caused by the persistence of attentional weights to various regions of the scene (based on the demands of a preceding task) (Thompson & Crundall, 2011;Thompson et al, 2015).…”

“…Any trials where the eyes were not fixated on the screen where the face would appear were removed from the subsequent analysis (this occurred in less than 1% of trials). In half the trials a further two letter searches oriented in the same way as the first were presented, in order to enhance the unpredictability of the timing of when the pictures were presented (see Thompson & Crundall, 2011;Hills et al, 2016). The different numbers of letter strings prior to the main task prevents participants developing anticipatory strategies that inhibit the carryover effect (see Thompson, Howting, & Hills, 2015 for a more extensive discussion on this).…”

“…Letter strings were either arranged horizontally, vertically, or randomly across the screen and have been shown to elicit eye movement changes (Hills, Thompson, Jones, Piech, Painter, & Pake, 2016). Given that during upright face processing, there is a distinct scanpath consisting of proportionally more horizontal saccades (moving between the eyes) than vertical saccades (e.g., Rizzo, Hurtig, & Damasio, 1987), we would expect that the carryover from the horizontal letter string will aid processing of a face by speeding up the recognition of the face, whereas carryover from a vertical letter string may be detrimental to face processing.…”