| The first major project in the lab investigated the effect of training on spatial working memory performance, in collaboration with Yuhong Jiang. For the basic version of the task, subjects had to memorize the locations of 6-12 squares presented briefly on the computer screen. After a delay, they had to produce the location of the missing squares when all but one reappeared. Throughout the experiment, half of the displays were repeated once every 12 trials, while half of the displays were randomly generated each time. In a non-associate learning condition, different squares were probed at each occurrence of a repeated display, and performance did not improve. In the associative learning condition, the same square was probed each time, and performance was enhanced. At the end of the experiment, we tested long term memory for the repeated displays, and all were recognizable, even in the non-associative learning condition. |
|
| Another major project of ours investigates how much control we have over the contents of our visual working memory. In each experiment, subjets are instructed to remember a set of items that are outlined or underlined, and to ignore any other items. After a delay, a single item appears on the screen, and subjects must indicate if this was one of the target items from the previous display. In sequential presentations of stimuli, we discovered that lure (to-be-ignored) objects harm performance on this task, because they are weakly encoded and consequently mistaken as target (to-be-remembered) objects at retreival. We have replicated this effect with different delay intervals, different types of cues, and with shapes, faces, and letters. However, in simultaneous presentation of stimuli, lure items do not intrude into the visual working memory store. Thus, we believe that visual transients cause unselected and automatic entry into the visual working memory store. | 
|
| In two separate patient populations (from Philadelphia and Boston), we investigated the role of the medial temporal lobe in visual short term memory by testing patients with damage to this area of the brain on a variety of different tasks, including color memory, face memory, spatial memory, and object/location binding memory. Our findings are that damage to the MTL causes visual short term memory impairment, and that the area seems to play a particular role in binding. | 
|
| In a collaborative project with Geoffrey Aguirre, we used perfusion imaging to investigate slow learning of a motor task. Because of its sensitivity to slow neural changes over time, perfusion imaging is a good method to studying learning tasks. Our experiment found brian/behavior learning correlations, thus demonstrating that perfusion fMRI is a good tool for measuring cognitive function, and that its applications should be expanded beyond the clinical realm. | 
|
| A recent project in the lab concerns preparatory processing. Specifically, we are interested in seeing how the brain changes in preparation to make a response or a judgment, based on how much information is given before the task. In our experiment, subjects view a 4x4 grid with 1, 4, 8, or 0 filled in locations. After a delay period, subjects perform a visual search task for faces in another 4x4 grid. The filled locations indicate possible locations for the target item in the search task. Thus, when one location is filled subjects have the most information; they have no preparatory information when none of the locations are filled. 4 and 8 provide intermediate amounts of information. As predicted, subjects are faster to find the target the more preparatory information they have. |  |