How does associative learning shape visual selective attention and distractor suppression?

Experience quietly retunes what grabs your attention

When a physically salient but task-irrelevant object keeps showing up in the same place, it becomes less distracting there. In a visual search task where a color-singleton distractor appeared more often in one hemifield, participants' brains showed reduced neural excitability in early visual cortex at the retinotopic location tied to the higher distractor probability, and this reduction occurred in the prestimulus interval, before the search array appeared [1]. That timing matters: it suggests the visual system is preparing in advance, not just reacting more slowly after the distractor arrives.

The learning is not limited to location. Distractors defined by shape also lose some of their attentional pull when a particular shape is a high-probability distractor, even though shape is represented later in the visual pathway than color or orientation [2]. That extends the reach of statistical learning beyond the earliest feature stages, consistent with the idea that learned priority can be attached to more complex object representations.

Knowing what to ignore can be used deliberately, not just absorbed implicitly

Statistical learning is usually described as implicit: participants are not told the probabilities and often cannot report them. But explicit foreknowledge of a distractor's location or color can also improve selective control. When people were told about distractor characteristics in advance, they showed reduced electrophysiological signs of covert attentional selection of those objects and made fewer eye movements toward them, and this control was established before the distractor appeared [4]. The strength of cue-elicited alpha-band activity over visual cortex predicted how well participants suppressed the distractor, linking a measurable brain rhythm to the implementation of strategic suppression [4].

This complicates a simple story in which only implicit learning matters. The same behavioral outcome, less interference from a known distractor, can arise from statistical learning or from deliberate strategy, and the two may engage overlapping but not identical mechanisms. Notably, the implicit statistical learning study found no evidence that alpha-band activity carried the expectation-driven suppression, suggesting that prestimulus excitability changes and alpha oscillations may reflect different attentional mechanisms [1].

Some of the “suppression” is really the rare distractor hitting harder

A frequent distractor in one location is usually compared with the same distractor in other, low-probability locations. That design cannot tell whether interference dropped at the high-probability location or rose at the low-probability ones. When high-probability, low-probability, and equal-probability locations were compared directly, interference was indeed reduced at the high-probability location relative to equal-probability locations, consistent with genuine attentional suppression [3]. But interference was also increased at low-probability locations relative to equal-probability locations, meaning rarity itself boosts distraction [3].

So the net benefit of a predictable distractor location is partly a suppression effect and partly a rarity penalty paid elsewhere. This does not erase statistical learning of distractor suppression, but it changes how large and how clean the effect is in typical experiments.

Suppression learned from experience fades faster than facilitation

Learning to ignore a location does not appear to stick as durably as learning to attend to one. In eye-movement experiments where some locations were associated only with target selection and others only with distractor suppression, selection history effects were long lasting, while suppression history traces were relatively short lived in both long-term and short-term tests [5]. The authors propose that selection may depend on lasting changes in stored representations of visual space, whereas suppression may rely on a more transient, constantly updated functional plasticity supporting ongoing oculomotor control [5].

That asymmetry is a useful caution: a distractor-suppression effect measured minutes after learning may not reflect a permanent change in how the visual system prioritizes that object. It may instead reflect a flexible, moment-to-moment adjustment that decays when the statistics change.

Sources used in this answer

Related research questions