Geoffrey Woodman at Vanderbilt University will examine the long-standing theoretical issue of how working memory and attention interact to determine what subset of information we process from complex visual scenes. A growing number of theories propose that during scene processing attention networks are guided to process task-relevant information through top-down control exerted by template representations in visual working memory. In this way, holding a representation in visual working memory is sufficient to guide perceptual attention mechanisms to select incoming information. In addition, some theories propose that representations of attended items need to pass through visual working memory for them to be categorized. But, the evidence for this direct linkage between visual working memory storage and attentional guidance is mixed, perhaps because the theories do not always consider the additional influence of long-term memory on mechanisms of attention. This project will use behavioral and electrophysiological methods to systematically explore the relative contributions of working memory and long-term memory to the top-down control of attention. By addressing these important questions regarding the interactions between memory systems and perceptual attention, the findings of the experiments in this project, will serve to constrain and shape the next generation of models of attention and control of information processing.
Our complex environment overloads our visual systems with a multitude of objects and surfaces, only some of which are relevant for our task at hand. For example, when driving we need to be attending to the cars that surround us and potential hazards entering the roadway, while the trees and houses lining the road are irrelevant for our task and should not be the focus of attention. Models of attention accounting for such complex processing tasks propose that we optimize information processing for task-relevant objects, like other cars on the roadway, by maintaining a representation of these cars in working memory. A better understanding of how our attention is controlled will allow us to predict when the performance of critical tasks, such as driving or controlling air traffic, will face potentially hazardous processing limitations. In addition, a number of cognitive disorders appear to be due to a compromised ability to control attention. Attention deficit disorder, reading disorders, and schizophrenia all appear to have underlying causes in an impaired ability to selectively process inputs and temporarily remember relevant information. The research proposed in this project will allow us to understand how attention is normally controlled and the methods developed here will provide tools for better understanding the deficits that are due to abnormal attentional control and selective memory storage.
In this project we developed methods to noninvasively record brain activity that specifically measure short-term memory and long-term memory representations. Our experiments showed how we could simultaneously measure the nature of the memory representations as participants find objects in complex visual scenes. Our findings challenged models of attention that propose short-term memory representations control attention as our brain process information, and bring the attention literature into register with the long history of research on learning and skill acquisition. In some of our work we have shown how our measures of brain activity can be used to understand the effects that other manipulations have on information processing. In one recent paper (Reinhart & Woodman, 2014, Cerebral Cortex) we show that reward can improve the processing of complex scene stimuli by brining multiple memory representations to bear on controlling attention. It appears that the tools developed in this project can be applied to understand how cognitive processing is influence by manipulations of the environment or subject in a variety of situations. Most recently, our interest in learning sparked by this project lead us to combine brain stimulation with measurements of behavior and brain activity. We have shown that we can accelerate learning during simple tasks and during search for objects in complex scenes. These findings have provided critical information about the nature of the neural mechanisms of memory and attention, and our findings have indicated how our methods can be used to improve cognition in a variety of different subject populations (e.g., students, aging adults, patients with a variety of disorders, etc.). The support from this grant resulted in 28 publications in press or print.
