Much of what we know about visual working memory has been uncovered using simple stimuli such as colored squares, but recent studies have suggested that alternative conclusions can be drawn using more complex stimuli, such as real-world objects. A possible explanation as to why these differences exist is that objects are redundantly coded broadly across several brain regions, whereas simple stimuli like colored squares are encoded in a more limited set of regions. We hypothesize that spatial position is enhanced automatically by this redundant encoding. In our experiment, participants (N=30) were given a delayed spatial recall task in which participants remembered arrays of either real world objects or colored squares. Participants maintained the precise spatial position of 1, 2, or 6 visual stimuli over a brief 1.5s delay period. We found that participants recalled spatial positions of objects with higher precision than the colored squares. Moreover, in separate trials where participants discriminated which of two stimuli appeared at a probed location, performance was identical between stimulus conditions. Altogether, this result supports the notion that incidental features of real-world objects, such as their spatial position, can be remembered with greater precision than those of simplistic stimuli typically used in laboratory tests.

Attention is a cognitive mechanism used to select important information. Here, importance is defined as one of three factors: image salience, or something that stands out based on local feature contrast; goal relevance, or items that are related to accomplishing a task; and selection history, or implicitly learned stimulus properties. Features of an item, like color and motion, are prioritized for attentional selection, but the exact mechanisms that allow for feature-specific attentional selection are unclear. Across three studies, evidence is presented on how information is attended based on feature-specific image salience, goal relevance, and selection history. Chapters 2 and 3 describe two fMRI studies investigating how feature-selective regions of visual cortex are recruited to track the location of color and motion stimuli throughout the visual field. Stimuli were either made salient because they were a different feature value from the rest of the display (Chapter 2), or relevant because participants were instructed to attend a certain feature dimension (Chapter 3). Inverted encoding models were computed to reconstruct spatial maps using neural activation patterns from retinotopic color (hV4/VO1/VO2) and motion (TO1/TO2) regions. Neural responses to the stimuli were strongest when the preferred feature of each region was important through manipulations of salience or relevance. These findings implicate color and motion areas as ‘feature dimension maps’—regions that are critical for indexing important color and motion information for attention. Chapter 4 describes a study concretely demonstrating that feature values learned over the course of an experiment can modulate what is prioritized. Specifically, the color of a regularly presented distracting stimulus was learned, and eventually ignored to improve target detection. Suppression was potent enough to interfere with ongoing goals, as participants took longer to select the target when it was shown in the learned distractor color. This rules out alternative accounts suggesting that only spatial information can be suppressed through selection history. Together, this body of work highlights how stimulus features are leveraged to successfully navigate our environment.

Though our experience of the world often appears rich and detailed, less information is immediately accessible than our intuition would suggest. When viewing a complex scene – such as a crowded city street – information about irrelevant features of the environment is lost due to noisy neural processing, while information about relevant features (those selected by visual attention) is spared. Similarly, behavioral experiments demonstrate that when even modest amounts of information must be held briefly in mind (in visual working memory), the amount of available information about each item is diminished, and this available information decreases with increasing information load. In what manner do visual representations across large-scale neural activity patterns support these behavioral information processing limits? In three studies, we examined the fidelity with which human cortical neural activation patterns measured with functional magnetic resonance imaging represent visual information. To this end, we developed a novel analysis technique whereby we reconstruct images of visual stimuli using neural activation patterns measured over entire brain regions. Using this technique, we established that the neural representation of a relevant visual stimulus is enhanced in its amplitude over a noisy baseline in several visual and parietal cortical regions, suggestive of an increase in the representation’s information content. Subsequently, we demonstrated that under conditions where no information is available in a display, the maintenance of a larger number of items in visual working memory is accompanied by a degradation in each item’s representation amplitude, indicative of lower population-level information content. Finally, we evaluated the relationship between these two findings by directing participants to attend to one of several items held in visual working memory. Surprisingly, we discovered that degraded representations can recover with visual attention, and the degree of recovery was related to behavioral task performance. Such recovery of degraded information suggests that additional information must be available to the system but invisible to our measurements before attention is allocated. Together, these results demonstrate that behavioral limits on information processing are related to the fidelity with which visual information is represented in large-scale neural codes.