Improving Performance in Real-World Visual Searches

There has been a great deal of research on visual search. Although much of that work was focused on understanding TSAbagbasic attentional processes, there is some hope that the knowledge gained from that work would be applicable to important real-world search scenarios such as baggage screening or cancer screening/radiology. However, there are significant differences between the methods used in typical lab-based search tasks and many real-world scenarios. Most notably, most lab-based searches have used very high prevalence rates (targets are usually present on 50% or more of the search trials) and have presented trial-by-trail performance based feedback. By contrast, targets are extremely rare in most real-world scenarios, and there is no ability to provide trial-by-trail feedback. Recently, researchers have begun to appreciate that visual search processes are highly impacted by the prevalence rate of targets. Specifically, under low prevalence search, the rate of successful target detection is dramatically reduced, an effect known as the low-prevalence effect.

This area of research has three foci: To determine the underlying search and decision mechanisms responsible for the low-prevalence effect, to investigate novel ways of presenting the search task to avoid the low-prevalence effect, and to take an individual differences approach to identifying people who cognitive and personality characteristics that would allow them to perform well at low prevalence search tasks.

Volitional Attentional Control – Mechanisms and Limitations

One’s goals can have a profound influence on the allocation of attention.  In search tasks, knowledge of the target allows one to constrain search to only those objects which share features with the target  (Egeth, Virzi, & Garbart, 1984; Williams, 1967) and the capture of attention by distractors is contingent on the distractor matching the goal-SearchDisplayrelevant features for one’s task (e.g., Folk, Leber, & Egeth, 2002; Folk, Remington, & Johnston, 1992).  These findings suggest that attentional allocation can be influenced by the application of an attentional control setting (ACS) that defines task-relevant stimulus features.  Most of the research supporting these conclusions involved tasks in which a single ACS was relevant at a time.  However, recent claims (e.g., Soto, Humphreys, & Blanco, 2005) suggest that activating an ACS involves representing the target in working memory (WM).  The fact that WM can simultaneously represent multiple items (e.g., Zhang & Luck, 2008) begs the question of whether multiple ACSs can be active simultaneously.  Recently, we have performed a number of experiments to examine whether two ACSs can simultaneously drive attentional allocation and to determine the flexibility and limits of doing so.   In addition to defining these limits, the nature of these limits will inform our understanding of the mechanisms involved in the volitional control of attention. 

Within this framework, we have recently addressed the following questions:  Can two independent ACSs be used simultaneously to constrain search?  If so, can each ACS be constrained to a specific region of space?  Can two simultaneously held ACSs increase the gain of low-level units tuned to the features of each ACSs?  Does the ability to simultaneously activate two ACSs indicate that one can also simultaneously consolidate two distinct features into WM?


The bandwidth of Working Memory Consolidation

In collaboration with Dr. Taosheng Liu, we have been investigating the bandwidth of the process of converting perceptual information into visual working memory (VWM) representations. In this work we have briefly presented stimuli either simultaneously or sequentially, to investigate the extent to which multiple stimuli can be transferred toERPData VWM simultaneously. Interestingly we find different results for different basic visual features. For instance, it appears that at least two colors can be simultaneously encoded into VWM without any cost. However, two orientations appear to require a strictly serial encoding. Recently we also performed an ERP version of the experiment which seemed to confirm these results. We believe this bandwidth represents a basic limit in visual cognition, believe we have an effective method for investigating these limitations, and continue to examine how and why different features have different outcomes. See my publications for a number of recent papers on this topic.

The Design and Evaluation of Better Product Labeling

Dr. Laura Bix in the MSU department of packaging and I have been collaborating on a series of experiments that apply basic research and techniques from visual attention to the design and evaluation of more effective product labels. Labels offer a cost-effective method of attempting to provide information that can empower people to make better/healthy decisions, but only if they are effective. From an information processing approach, a label must go NutritionLabelthrough a number of stages before it can influence behavior. One of those critical stages is that the label must garner attention; if a label is not attended its processing is derailed early in processing. Little work has investigated how different labeling techniques influence attention. In this work we apply basic research on visual attention to design labels that might be more effective and then use a combination of visual search, eye-tracking, change detection, and speeded sorting tasks to evaluate the effectiveness of those labels. We have investigated prescription and over-the-counter drug labels, medical device packaging, and front-of-pack nutritional labels. For more information about these projects see our publications or visit Dr. Bix's lab site at https://www.msu.edu/~bixlaura/index.htm.


Emotion and Attention

Emotional stimuli can impact the attentional allocation system in a number of ways. Our lab has investigated a number of questions concerning the impact that emotional stimuli can have on attentional processing. These include: An EmotionalStimevaluation of whether the presence of negative stimuli act as a simple alarm system, whether the congruency between one's mood and the emotional valence of stimuli impact the likelihood of attending to the stimuli, the extent to which anxiety interferes with top-down/volitional control of attention, whether associating a visual feature with punishment leads to a suppression of attention capture by that feature, and the extent to which threat relevant stimuli alter visual processing. Please see my papers for answers to these questions. In sum, across these studies it is clear that emotional stimuli can have diverse and nuanced effects on visual processing. I continue to investigate this issue with a particular emphasis on how threat may impact visual processing and whether people can be trained to avoid non-important visual features.

Media Multitasking and Attention

With my former graduate student, Reem Alzahabi, we investigated the impacts of frequent media multitasking on cognitive performance, mood and anxiety. One line of this work investigated the relationship between frequent media multitasking and the ability to rapidly switch between tasks and/or perform two tasks simultaneously. The results suggest that frequent media multitaskers are unable to perform two tasks simultaneously any better than non-multitasking controls (both show similar psychological refractory effects), but they are able to rapidly switch between tasks more efficiently. In follow-up work, we performed an individual differences experiment to identify which aspects of task switching are related with media multitasking. Finally, while doing this work we also found that people that self-report higher rates of media multitasking also report higher depressive and anxiety symptoms. Most of my papers with Alzahabi as an author investigate this issue.