Why finding a likely target and ignoring a likely distractor are separate adjustments
In a typical visual search task, you look for a target among other items while a more colorful or otherwise salient distractor tries to grab your attention. If the target appears more often on the left, search gets faster on the left. If the distractor appears more often on the right, the distraction caused by items on the right gets smaller. These effects are often described as implicit statistical learning because participants are not told the probabilities and do not need to try to learn them [2][3].
The two effects can come apart. In one set of experiments, younger and older adults both showed a strong, persistent advantage for targets at frequently attended locations. But older adults did not benefit from learning where distractors were likely; their distraction remained throughout the experiment regardless of the distractor-location contingencies [2]. That pattern suggests target selection and distractor suppression rely on at least partly different mechanisms, and that the suppression side may be more fragile with age [2].
A separate study found that object structure affected target search but not distractor inhibition. When a high-probability distractor location was embedded in one of three objects, target selection was slower at an equidistant location within that same object than at a location in a different object. Yet attention capture by the singleton distractor itself was hardly modulated by objects [1]. In other words, the object grouping shaped how attention moved toward targets, not how strongly the distractor captured attention.
Distractor suppression tracks the local history of a specific region
How much a distractor is suppressed depends on how often it has appeared in that particular region, not on some global pool of inhibition spread across the display. When researchers varied the ratio of distractors in a frequent versus rare region while holding the overall distractor frequency fixed, the suppression effect grew as the ratio became more extreme. Crucially, changing the distractor frequency in one region did not significantly change interference from distractors in another region [3]. That is consistent with a local learning or habituation-like response at each location rather than a shared, limited inhibition resource [3].
The neural evidence points in a similar direction but adds a timing distinction. EEG recordings showed that both distractor feature regularities and distractor location regularities reduced distractor costs and produced an earlier distractor-evoked Pd component, a brain signal associated with suppression. However, control analyses indicated that much of the observed effect reflected intertrial repetition, meaning the distractor on one trial was similar to the distractor on the immediately preceding trial. The authors suggest that early and late Pd components may reflect different processes: early sensory suppression tied to intertrial priming, and later suppression sensitive to expectations built over longer time scales [4].
Expecting a target and expecting a distractor engage attention differently
When people are told in advance that a target will probably appear in a certain location, that expectation and spatial attention interact: the benefit of attention depends on whether the location was expected. When the expectation is instead about where a distractor will probably appear, expectation and attention affect performance additively, each contributing separately rather than amplifying or weakening each other [5]. This difference held across several experiments and did not change when the task was made harder by adding more items [5].
The authors interpret this as evidence that promoting a target and inhibiting a distractor are handled by different neural mechanisms, so expectation anchored to one does not combine with attention in the same way as expectation anchored to the other [5]. That interpretation fits the broader pattern in which target-related learning is robust and persistent while distractor-related learning is more local and more easily disrupted [2][3].
What the short-term memory part of the question leaves open
The studies here focus on statistical learning over many trials, not on implicit short-term memory as a separate construct. The evidence supports the idea that recent repeated experience, including intertrial repetition, contributes to distractor suppression [4], and that local frequency in a region drives how much suppression develops there [3]. But these designs do not isolate a short-term memory store from longer-term probability learning, so the precise role of implicit short-term memory in target activation and distractor inhibition remains an open question rather than a settled finding.
Sources used in this answer
Object-based suppression in target search but not in distractor inhibition
Object grouping slowed target selection at a location within the same object as a high-probability distractor location, but did not modulate attention capture by the singleton distractor itself.
Age-related differences in the statistical learning of target selection and distractor suppression.
Older adults showed preserved and persistent target-location learning but did not benefit from implicit learning of distractor locations, unlike younger adults.
Learning to suppress likely distractor locations in visual search is driven by the local distractor frequency.
Distractor suppression increased with the local ratio of distractors in a region and was not significantly affected by distractor frequency in another region, consistent with local learning.
Neural mechanisms underlying distractor inhibition on the basis of feature and/or spatial expectations
Distractor feature and location regularities reduced distractor costs and produced an earlier Pd ERP component, with much of the effect reflecting intertrial repetition and early versus late Pd components possibly reflecting different processes.
The effects of expectation on attention are dependent on whether expectation is on the target or on the distractor?
Expectation about a target location interacted with spatial attention, whereas expectation about a distractor location affected performance additively, and this pattern did not change with task difficulty.