Michael L. Paavola¹, Cathleen M. Moore¹; ¹University of Iowa

There are situations in the real-world where people continuously monitor dynamic scenes for rare target events, such as in lifeguard surveillance. Standard visual search studies, in which participants search within static displays for targets defined by features, and which use discrete-trial designs with each ~2sec trial either containing a target or not, have shown that people tend to respond more slowly and make more errors with low-prevalence (LP) targets than higher-prevalence (HP) targets (Low-Prevalence Effect, LPE). We asked whether these findings generalize to conditions in which the participants continuously monitor dynamic displays over an extended period for target events, which requires the detection of spatiotemporal patterns. In addition, inspired by a practice in the aquatics community, we asked whether adding a probe-detection task would mitigate the LPE. Displays consisted of 8 to 20 gray disks continuously cycling between high and near-threshold contrast at random phases. Participants monitored for circles that remained at near-threshold contrast for longer than the standard time. They pressed the spacebar when a target event was detected and indicated its location using the mouse. For different groups of participants, target events occurred with a probability of .25 (LP) or .75 (HP). Additionally, a red circle occasionally flashed somewhere within the display. Half the participants were to respond to these by pressing the space bar and indicating its location. The other half were told nothing about them. Results confirmed an LPE. The probe task reduced the LPE, but did so by increasing the miss rate for HP targets rather than reducing it for LP targets. Finally, the addition of the probe task also reduced the standard effect of set size on response time. These results provide initial insight into how dynamic continuous monitoring for event targets is similar to and different from static visual search in discrete-trial designs.