Abstract
Many studies have shown that task performance improves with practice. Most such improvements are the result of stimulus over-exposure leading to automatic processing, or memorized stimulus-response mappings. The question of whether or not practice can lead to improved performance through an increased ability to suppress distracting information remains open. Subjects reported whether there were more red or green dots in a randomly generated 5×5 array. On half of all trials, a distracting item (a flashing red and green disk) appeared in either the upper left or lower right corners of the display, just prior to the array onset. Early in learning, this distractor captured the subjects' attention, resulting in slower discrimination response times (RTs). As learning proceeded, this RT cost declined, and eventually it was eliminated completely. This indicates that practice resulted in improved filtering of the distracting items. For a second group of subjects, the position of the distacting items was switched to the opposite corners of the display after several blocks of learning. This almost completely abolished the practice benefit; the benefit returned, however, by the end of the session. For a third group of subjects, the distractor disks were replaced with faces after several blocks. Again, this almost completely abolished the practice benefit. These data indicate that, although improvements in distractor filtering can develop rapidly, these improvements are specific to the identity and location of the distracting items, and do not transfer to new types of information.